Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Microbe to microbe: monthly microbial community dynamics and interactions at the San Pedro Ocean Time-series
(USC Thesis Other)
Microbe to microbe: monthly microbial community dynamics and interactions at the San Pedro Ocean Time-series
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
MICROBE TO MICROBE: MONTHLY MICROBIAL COMMUNITY DYNAMICS AND INTERACTIONS AT THE SAN PEDRO OCEAN TIME-SERIES by Cheryl-Emiliane T. Chow A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (BIOLOGY) August 2012 Copyright 2012 Cheryl-Emiliane T. Chow ii Dedication To my family for their unwavering love, support and encouragement in this adventure, to my sister for keeping me on track and to my grandfather for the words of wisdom and the frequent postcards of Albert Einstein. iii Acknowledgements I would like to begin by thanking my advisor, Jed Fuhrman, for his guidance through this, at times wandering, adventure and my committee Dave Caron, John Heidelberg, Eric Webb and Richard Roberts for their mentoring and advice throughout this research. I am indebted to the entire Microbial Observatory team as this work would not have been possible without them, especially Troy Gunderson, Diane Kim, Rohan Sachdeva, Jacob Cram, Joshua Steele, David Needham, Anand Patel, Tu My To, Alma Parada, Elizabeth Teel, Mike Beman, Ian Hewson, Mike Schwalbach, Sheila O’Brien, Mahira Kakajiwala, Pete Countway, Adriane Jones, and the captain and crews of the R/V Seawatch and R/V Yellowfin. I would also like to thank additional collaborators: André Comeau for the g23 primer sequences; Eric Wommack, Shawn Polson, and Daniel Nasko for their assistance and access to VIROME; Steven Hatosy, Adam Martiny, Matthew Haynes, Forest Rohwer, Matthew Sullivan and Bonnie Poulos for helpful discussions, technical assistance, and advice. This research was supported by NSF Microbial Observatories, Biological Oceanography and Dimensions in Biodiversity programs (grant numbers: 0703159, 1031743, and 1136818), USC Wrigley Institute for Environmental Studies, and by NSF Graduate Research, USC Rose Hills and Wrigley fellowships. Sequencing of the viral metagenomes was provided by the Broad Institute through the Marine Phage Sequencing project under the Marine Microbiology Initiative at the Gordon and Betty Moore Foundation. iv Table of Contents Dedication .................................................................................................................................. ii Acknowledgements ................................................................................................................... iii List of Tables ............................................................................................................................ vi List of Figures .......................................................................................................................... vii Abstract ..................................................................................................................................... ix Chapter One: Overview of Current Perspectives on the Ecological Roles of Marine Viruses and the Methods Employed to Investigate Them ..................................................... 1 Dissertation Research Aims .................................................................................. 5 Study Site: San Pedro Ocean Time-Series Station ................................................ 6 Overview of Methodological Approaches ............................................................. 7 Chapter One References ...................................................................................... 14 Chapter Two: Temporal Variability and Coherence of Bacteria Communities from a Ten-Year Time-Series in the Euphotic Zone of the Southern California Bight ............. 19 Abstract ............................................................................................................... 19 Introduction ......................................................................................................... 20 Materials and Methods ........................................................................................ 21 Results and Discussion ........................................................................................ 26 Conclusions ......................................................................................................... 49 Chapter Two References ..................................................................................... 50 Chapter Three: Seasonality and Monthly Dynamics of Marine Myovirus Communities ............................................................................................................................ 55 Abstract ............................................................................................................... 55 Introduction ......................................................................................................... 56 Experimental Procedures ..................................................................................... 59 Results ................................................................................................................. 63 Discussion ........................................................................................................... 75 Conclusions ......................................................................................................... 79 Chapter Three References ................................................................................... 81 Chapter Four: Identifying Controls on Bacterial Community Structure: Evaluating Connections to Viral And Protistan Communities .............................................................. 86 Abstract ............................................................................................................... 86 Introduction ......................................................................................................... 86 Materials and Methods ........................................................................................ 90 Results ................................................................................................................. 93 Discussion ......................................................................................................... 107 Conclusions ....................................................................................................... 112 Chapter Four References ................................................................................... 114 v Chapter Five: Metagenomic Depth Profile of Viral Communities Highlights Depth- dependent Variation .............................................................................................................. 121 Abstract ............................................................................................................. 121 Introduction ....................................................................................................... 122 Materials and Methods ...................................................................................... 124 Results ............................................................................................................... 127 Discussion ......................................................................................................... 136 Conclusions ....................................................................................................... 138 Chapter Five References ................................................................................... 140 Chapter Six: Synthesis .......................................................................................................... 144 Concluding Remarks and Future Directions ..................................................... 150 Chapter Six References ..................................................................................... 152 Bibliography (Alphabetized) ................................................................................................ 156 vi List of Tables Table 2-S1. Correlation of the relative abundance of the top 5 OTUs between 0-5m and DCM.. ................................................................................................................................ 30 Table 2-S2. Network Statistics for depth-specific networks and description of unique features within ................................................................................................................... 42 Table 3-1. Overview of g23-TRFLP OTUs ............................................................................. 70 Table 4-1. Correlation of Bray-Curtis similarities between microbial communities. ............... 94 Table 4-2. Distribution of significant LSA correlations (edges) between all microbial OTUs and environmental parameters. .............................................................................. 98 Table 4-3. Description of LS correlations for exemplar networks in Figure 2. ........................ 98 Table 5-1. Overview of Metagenome Sample Libraries ......................................................... 128 Table 5-2. Phylogenetic Assignment by Shotgun UniFrac ..................................................... 135 vii List of Figures Figure 1-1. Viral shunt redirects nutrient cycling ....................................................................... 2 Figure 1-2. Map of San Pedro Ocean Time-series location. ....................................................... 6 Figure 1-3. Description of molecular approaches to characterize viral communities ................. 9 Figure 1-4. Methodological pipeline for studying aquatic microbial diversity. ....................... 10 Figure 1-5. Overview association network of local similarity correlations from three-year time-series of bacterial, T4-like viral, and protistan communities. ................................... 11 Figure 2-1. Environmental variability over ten years in the euphotic zone at SPOT ............... 28 Figure 2-2. Inter-annual and monthly variability of top 5 ARISA OTUs over ten-years. ........ 31 Figure 2-S1. Monthly variability in cyanobacteria relative abundance at SPOT. .................... 32 Figure 2-S2. Monthly and inter-annual variability of SAR11 ecotypes in (a) the surface ocean and (b) DCM ........................................................................................................... 33 Figure 2-S3. Time-series and discriminant function analyses from 50 highest contributing OTUs ................................................................................................................................. 37 Figure 2-3. Seasonal and inter-annual patterns in Bray-Curtis community similarity. ............ 38 Figure 2-S4. Seasonal and inter-annual patterns in Bray-Curtis community similarity at 0- 5m for (a) the 50 most important OTUs and (b) those OTUs occurring in more than 50 months .......................................................................................................................... 38 Figure 2-4. Bray-Curtis similarity between 0-5m and DCM (y-axis) is negatively correlated with temperature difference (y-axis) ................................................................ 41 Figure 2-5. Two highly interconnected sub-modules linked within a defined module in the surface ocean association network .............................................................................. 43 Figure 2-6. Connections that occur in both 0-5m and DCM networks between the five most abundant OTUs and their nearest neighbors. ........................................................... 43 Figure 2-7. Defining the core microbiome of the euphotic zone at SPOT. .............................. 48 Figure 3-1. Venn diagram of percent of total OTUs shared over five months. ........................ 65 Figure 3-S1. Distribution of OTUS from a) sequencing clone libraries, b) in silico digestion of clone libraries, c) environmental TRFLP or d) “reduced” environmental TRFLP ............................................................................................................................... 66 Figure 3-S2. Environmental variability at SPOT 0-5m from March 2008 to January 2011 ..... 67 viii Figure 3-S3. Schematic of 3’-HincII TRFLP fingerprints illustrates changes in relative abundances of unique fragment length over time. ............................................................ 68 Figure 3-2. Virus OTUs detected more frequently have higher average and maximum contributions to the T4-like community ............................................................................ 70 Figure 3-3. Seasonality by Discriminant Function and Bray-Curtis Similarity Analyses ........ 71 Figure 3-S4. Discriminant function and time series analyses demonstrate re-occurring viral communities from all four g23-TRFLP assays ......................................................... 72 Figure 3-4. Individual 3’-HincII TRFLP OTUs exhibiting three distinct seasonal patterns .... 74 Figure 4-1. Month – to – month shifts in Bray-Curtis Similarity within microbial communities ...................................................................................................................... 95 Figure 4-S1. Distribution of LS values, rounded to the nearest tenth ....................................... 97 Figure 4-2. Three mini networks and the relative abundance of each OTU over time for LS correlations observed in the overall T4-like virus, protist, and bacteria microbial association network ........................................................................................................... 99 Figure 4-3. Broad overview of interactions between (A) protists and bacteria and (B) T4- like viruses and bacteria .................................................................................................. 101 Figure 4-S2. Bacteria-Protist correlations, including secondary protist-protist and bacteria-bacteria correlations .......................................................................................... 102 Figure 4-S3. Bacteria and 3H-Virus interactions, including connections between nodes of the same type ................................................................................................................... 103 Figure 4-4. Network of Top 5 bacterial OTUs and primary LS correlations to their first neighbors ......................................................................................................................... 104 Figure 4-S4. Top 5 OTUs and their first neighbors, including all edges between all nodes shown .............................................................................................................................. 105 Figure 4-5. Cyanobacteria OTUs and their first neighbors ..................................................... 106 Figure 5-1. Environmental characteristics of the water column at SPOT ............................... 129 Figure 5-2. Taxonomic identification of ORFs by domain per library ................................... 131 Figure 5-3. Taxonomic distribution of phage ORFs differs with depth .................................. 132 Figure 5-4. SEED Subsystem Functional Gene Distribution by metagenomic library ........... 134 ix Abstract Our ability to define and quantify individual microbe-microbe interactions with traditional community ecology principles is still in development yet these relationships are key to understanding microbial roles in the ocean. Application of molecular methods, such as community fingerprinting and metagenomics have added significant insight into marine viral and microbial ecology in general over the last decade. Great progress has been made in identifying what organisms or taxonomic groups are present, their abundances and their genomic context. Metagenomic and genomic studies have highlighted the phylogenetic and functional diversity of viruses in the ocean. However, aside from work on the limited number of cultured organisms, our understanding of microbe-microbe relationships is often broad and does not typically lead to systematic identification of multiple relationships from the environment at the same time. The ease and feasibility of these molecular tools are increasingly an essential part of the microbial ecologists’ toolbox. Continuing research with microbial association networks can lead to a better understanding of how individual members of a diverse microbial community relate to one another and will ultimately facilitate predictions on how resilient or susceptible the microbial community may be to their future ocean climate. The microbial community at the San Pedro Ocean Time-series was characterized in great detail – with a specific focus on viral and bacterial communities and inter-microbe relationships as part of the USC Microbial Observatory program. First, ten-years of observations of bacterial communities in the euphotic zone revealed seasonal and annual trends in the surface ocean (Chapter 2). There were only a few dissimilarities between the surface and deep chlorophyll maximum (DCM) depths in terms of the microbial constituents; thus, a core euphotic zone microbiome at this location was described based on taxonomic identity and persistence of x organisms. Differences between depths were not in the taxonomic composition of the bacterial communities, but rather were observed in depth-specific association networks. More specifically, it was the relative abundances and correlations between the OTUs that differed, suggesting that microbial interactions differ between depths. Second, a new molecular fingerprinting assay was developed to characterize the T4-like myovirus family – an abundant and diverse group of viruses in the ocean (Chapter 3). Seasonal trends were also observed for this virus family over a three- year period. A persistent subset of all OTUs was also identified, some of which demonstrated seasonality while others were moderately abundant and steady year-round. Third, interrogation of individual relationships between members of viral, bacterial and protistan communities in the surface ocean by statistical and network-based analyses led to identification of individual virus- bacteria and protist-bacteria correlations (Chapter 4). In general, T4-like viruses and bacteria were more tightly coupled than protists to bacteria based on network analysis and shifts in community composition. Lastly, differences in the water column were observed by viral metagenomics and found that mesopelagic viruses differed from the upper water column (Chapter 5). Photosynthesis genes, from cyanophage most likely, were primarily seen in the euphotic zone. Siphoviruses were a larger proportion of the community below the euphotic zone. Collectively, this dissertation detailed temporal variability in microbial communities by identifying several hundred viral and bacterial OTUs over time and the underlying connections. These results will provide new insights into the individual links that sustain community-level relationships within the microbial loop and our understanding of microbial roles in the ocean. 1 Chapter One: Overview of Current Perspectives on the Ecological Roles of Marine Viruses and the Methods Employed to Investigate Them “There is an intrinsic simplicity of nature and the ultimate contribution of science resides in the discovery of unifying and simplifying generalizations, rather than in the description of isolated situations – in the visualization of simple, overall patterns rather than in the analysis of patchworks.” – Salvador Luria, General Virology, 1953 Microbes (bacteria, archaea, protists) and viruses are fundamental players in the microbial loop (Azam et al. 1983; Fuhrman 1999; Wilhelm & Suttle 1999). Even though marine viruses were hardly known a few decades ago (Bergh et al. 1989; Proctor & Fuhrman 1990), we now know that such viruses are not only tremendously abundant (approximately one million in one drop of seawater) but are also key to global marine food webs. Changes in viral diversity over space and time is thought to have significant influence on microbial processes, yet we have little information on how viruses and their microbial hosts vary and interact in natural systems. Experimental field studies generally report that viral-based bacterial mortality leads to or maintains bacterial diversity but do not always show a clear result (Wommack et al. 1999; Schwalbach et al. 2004; Winter et al. 2004; Bouvier & del Giorgio 2007; Sandaa et al. 2009). Culture-independent molecular methods have facilitated an expansion into describing whole communities in greater detail than was previously available (Thurber 2009; Zinger et al. 2012). While we are starting to understand the diversity and complexity of these abundant organisms, it remains difficult to assess how the community functions as an integrated unit and also how these individual “species” units relate to one another. The intricate and dynamic relationship between microbes remains a complex issue. 2 Figure 1-1. Viral shunt redirects nutrient cycling. (A) Simple diagram of the marine microbial food web, with the viral shunt highlighted in red. Grazing shuttles carbon up the traditional food chain into higher trophic levels. By contrast, the viral shunt produces dissolved organic matter that can be consumed by other prokaryotes, to be recycled by marine microbes to stimulate nutrient and energy cycling. Figure and modified caption from Breitbart 2012. (B) The microbial loop as originally described in Azam et al. (1983) (solid arrows) and with later additions (dashed arrows). DOC refers to dissolved organic matter. Figure and modified caption from Fenchel 2008. A B 3 The Microbial Loop The “microbial loop” is a widely accepted view of how microbes influence global biogeochemical cycles. In this model, heterotrophic bacteria are major consumers of dissolved organic matter produced by phytoplankton in the ocean (Azam et al. 1983; Pomeroy et al. 2007; Fenchel 2008). Both heterotrophic bacteria and phytoplankton are then consumed by heterotrophic protists. Ducklow further proposed a multistep process for the degradation of dissolved organic matter (DOM) and emphasized the role of microbes in the regeneration of macronutrients, specifically N and P (Ducklow 1983). In the microbial loop, a significant amount of organic carbon produced by phytoplankton is recycled through several microbes before being remineralized to inorganic carbon dioxide, allowing for continued production in nutrient limited systems (Sherr & Sherr 1988; Pomeroy et al. 2007). Secondly, consumption of bacterial biomass by heterotrophic bacteria ultimately limits nutrient transport into higher trophic levels. This semi- closed system whereby the bacteria consume biomass produced from other bacteria following natural death, sloppy grazing by protists, or viral lysis inevitably links the life strategies of these organisms and creates a complex microbial community. Marine Viruses and Their Ecological Importance in the Ocean Through their diversity and abundance, viruses play an important ecological role in nutrient cycling on Earth and in shaping the marine community through the “viral shunt” within the microbial loop -” (Fuhrman 1999; Suttle 2005; 2007; Brussaard et al. 2008; Breitbart 2012). Previous estimates suggest that viral lysis of heterotrophic bacteria transfers 6-26% of carbon fixed by primary producers into the DOC pool (Fuhrman 1999). Lysis of phytoplankton blooms has also been shown to increase ambient DOC by 29%, demonstrating another transfer mechanism of organic carbon to heterotrophic bacteria (Wilhelm & Suttle 1999). Traditionally, 4 the “viral shunt” describes how viral lysis prevents transfer of nutrients from the bacteria to higher trophic levels by regeneration of DOM and POM pools for use by bacteria, as illustrated in Figure 1-1. Virus-bacteria relationships are a type of top-down control, akin to a predator-prey relationship. The “kill the winner” model was developed to describe the temporal dynamics of this relationship (Fuhrman & Suttle 1993; Bratbak et al. 1994; Fuhrman 1999; Thingstad 2000; Suttle 2007) - an interaction based on the Lotka-Volterra predator-prey relationship, where the virus preferentially infects the most abundant or dominant bacterial host. The model suggests that as a specific host becomes a dominant “winner” due to any competitive advantage, any of its viruses will subsequently replicate as a function of an increased host encounter rate. This species-specific infection leads to an increase in viral lysis of that host and effectively “kill(s) the winner.” Following increased viral lysis, the original winner would decline and open a new niche - allowing for the evolution of a new winner and the cycle to begin again (Wommack & Colwell 2000). The model, as proposed, suggests that the bacteria and the virus would maintain a constant cycle of fluctuation as new hosts became more competitive and dominant. In terms of viral abundance itself, the Bank Model suggests that there is constant production of rare “non-winning” viruses at a reduced level or a re-introduction into the system from neighboring water masses that would not be dependent upon a host’s dominance (Breitbart & Rohwer 2005). Models of viral genome abundance, based on sequencing, suggested that the most abundant viral genome is only up 2-3% of the community (Breitbart 2002; Angly et al. 2005). This predicted 2-3% abundant genotype would correlate to ~5x10 10 viral particles within the 200-liter sample (with ~2x10 12 viruses total) and would need to lyse 2x10 9 bacteria to maintain the observed abundance, assuming an average burst size of 24 - the number of viruses released by a host cell upon lysis (Wommack & Colwell 2000; Breitbart & Rohwer 2005). It 5 follows that an abundant phage must infect ~1% of the bacterial population at any one point in time in order to maintain its position in the community. This model also estimated a power-law dominated system of distribution and extremely high phage diversity, which has been consistent across later metagenomic surveys (Breitbart 2002; Angly et al. 2005; 2006; Hoffmann et al. 2007; Rodriguez-Brito et al. 2010). Despite their relatively small size, their incredible abundance and lytic activity allows the viral community to significantly impact nutrient biogeochemical cycles. Global models of nutrient allocation and microbial abundance must then also consider the perhaps differential, top- down effects of grazers and viruses on structuring bacterial populations and vice versa. In order to do so, it is essential to first understand how exactly these communities interact as a whole and on an individual or population level. Dissertation Research Aims The main objective of this research was to describe temporal variation in microbial communities alongside individual relationships within contemporaneous T4-like myoviral, bacterial and eukaryotic communities. This research was conducted at the USC Microbial Observatory at the San Pedro Ocean Time Series station (SPOT), located in the Southern California Bight. Guiding Questions 1. How do the bacterial and viral communities change in relation to the environment and each other over time and depth at the San Pedro Ocean Time-series station? 2. What are the individual connections that form ecological links within and between microbial groups, specifically for T4-like viruses, bacteria, and protists? 6 Specific Research Goals 1. Investigate shifts in viral and bacterial community structure over time (Chapters 2, 3 and 4). 2. Develop novel TRFLP-based methods to determine community structure of a common viral family, the T4-like myoviruses (Chapter 3). 3. Define connections within microbial association networks and their ecological implications (Chapters 2 and 4). 4. Characterize depth-dependent differences in bacterial communities by community fingerprinting (Chapter 2) and in viral communities by metagenomics (Chapter 5). Study Site: San Pedro Ocean Time-Series Station Figure 1-2. Map of the San Pedro Ocean Time-series location. The San Pedro Ocean Time-series station (SPOT) is located midway between Los Angeles, CA and Santa Catalina Island at 33˚33’N, 118˚24’W (Figure 1-2). Significant results to 7 date at this location include the identification of seasonality among the surface ocean bacterial and protistan communities, strong correlations among individual OTUs, and seasonal differences in different depths of the water column (Countway & Caron 2006; Fuhrman et al. 2006; Hewson & Fuhrman 2006; Beman et al. 2010; Countway et al. 2010; Beman et al. 2011). Research at SPOT has also focused on interrogating ecological relationships between the picoplankton (bacterial and eukaryotic) and the environment in order to understand what causes the observed seasonal changes in community structure (Beman et al. 2011; Steele et al. 2011). Overview of Methodological Approaches Community Fingerprinting of Microbial Communities Descriptions of microbial communities now often rely heavily on molecular characterization to describe species in the ocean (Figures 1-3, 1-4). Due to limitations to accurately describe a species by current taxonomic requirements, microbial communities are characterized into operational taxonomic units (OTUs). These OTUs, in essence, suggest that organisms differ genetically, but the lack of a cultured isolate for every microbe in the ocean prevents classical taxonomic identification at the species level. OTUs can be defined by sequence divergence or from bands on a gel. Methods range in resolution and specificity: from the more coarse Pulsed-Field Gel Electrophoresis (PFGE) for whole viral communities to Randomly Amplified Polymorphic DNA (RAPD) for random amplification and assessment of whole viral communities to conserved gene markers of specific taxonomic groups – whether a viral family or microbial domain, to sequencing everything by metagenomics (Wommack & Colwell 2000; Thurber 2009; Zinger et al. 2012). PCR-based surveys of single marker genes combined with sequencing, denaturing gradient gel electrophoresis (DGGE) and Terminal Restriction Fragment Length Polymorphism (TRFLP) have previously been widely adopted. 8 Marker gene methods rely on the presence of specific functional or structural genes that are considered phylogenetically representative. In general, molecular profiling of a community is consistent and reproducible, allowing for rapid characterization of the community structure and overall diversity for a large quantity of samples, particularly when there is a focus on temporal patterns, while sequencing can provide additional information on potential function (metagenomics) and the identification of organisms in a mixed community. All of the fingerprinting tools are efficient and cost-effective methods to surveying many samples and provide a look at the alpha- or beta- diversity of the community at large (Figure 1-3). Alpha-diversity, simply put, is the assessment of diversity within one space or sample, while beta-diversity allows for the comparison of diversity across sites or samples (Zinger et al. 2012). Measures of diversity include richness (the number of species), evenness (relative dominance or not of species). Using similarity indices, like the Bray-Curtis Similarity, the number and type of species (or OTUs) present and their relative abundances, can be compared across samples. Both alpha- and beta-diversity assessments provide insight into the ecology of the marine microbial realm. Bacteriophage diversity of one of the three main dsDNA bacterial virus families, T4-like myoviruses, has been studied using the major capsid protein (gp23) and portal protein (gp20) (Fuller et al. 1998; Wilson et al. 2000; Filee 2005; McDaniel & Delarosa 2006; Comeau & Krisch 2008; Sullivan et al. 2008; Clokie et al. 2010). G20 and g23 encode proteins required to form the capsid, the virion’s protective shell that encloses its nucleic acids. The capsid structure itself and size has traditionally been a distinguishing feature for electron microscopy (Suttle 2005). A need to look at the broader class of T4-like viruses led to the development of the core gene marker, g23 to replace the cyanophage-specific marker of g20. Validation of g23 as a good proxy for T4 phage diversity was based on an initial comparison of g23-based phylogeny to one from whole 9 genome sequences of 16 T4-like phages (Filee 2005). More recently, the addition of thousands of viral sequences from the Sorcerer II Global Ocean Sampling expedition and an expanded sequencing effort confirmed that the T4-like family is diverse and widespread (Yooseph et al. 2007; Comeau & Krisch 2008), suggesting that this virus family would be an excellent candidate for method development and target for interrogating viral ecology in the ocean. The development of this new fingerprinting method for the conserved core gene g23 for T4-like myoviruses will facilitate future studies on viral community structure by removing past limitations of specificity and feasibility to analyze a large number of samples. Figure 1-3. Description of molecular approaches to characterize viral communities. Table from Thurber 2009. Phylogenetic analysis of genes specific to different viral groups It is increasingly common for researchers to use only molecular techniques to describe phage consortia dynamics.Yet,phagegenomesdonotshareanyonegene, restricting molecular taxonomic analysis. Instead, inves- tigatorsusesingleormultiplephylogeneticmarkergenes such as the viral capsid structural genes, RNA and DNA polymerases, and the photosynthetic gene psbD to inves- tigatetheinterrelatednessofparticularphagegroups[21– 25]. One such study used the structural gp20 gene to investigate temporal variations in the abundance and composition of cyanomyophage assemblages at two locations over three years [26]. Genetic diversity was found to be low in winter and high in summer and, although not definitive, phylotype presence or absence trended toward a seasonal cycle. This study has been continued, and now more than 10 years of cyanomyoph- agedataawaitpublication.Theselongertermstudiesare a critical component to the evaluation of phage biogeo- graphy and should be extended to other systems. Virome shotgun sequencing The above technique requires at least some a priori knowledge of phage gene sequences. However, isolation and analysis of randomly selected (shotgun) genome sequence from biome specific viral assemblages requires almost no prior information about the targeted viral con- sortia.Thismethod,oftencalledviralmetagenomics,has been used to investigate viral consortia from a variety of sampletypes[27–30].Inagreementwiththemicroscopy, fingerprinting,andphylogeneticbasedmethodsmanyof theseearlymeta-viromestudiesfoundthatcertainphage sequences were ubiquitous or cosmopolitan [23,31–34]. These findings were previously reviewed and therefore will not be discussed here [1]. However, newer studies have found data to both agree and conflict with these previous virome findings. For example, microbial meta- genomes collected during the Global Ocean Survey expedition [35] from Nova Scotia to French Polynesia contained about 3% phage protein sequences [36 !! ]. Using protein sequence fragment recruitment to known viral genomes, the distribution of three different double stranded DNA phage families (Myoviridae, Siphoviridae, and Podoviridae) was determined. Although myoviruses (e.g. T4-like) were found to be cosmopolitan in the survey,therewereindicationsthatmyovirusesweremore abundant in the nutrient poor waters of the tropics. Podophage and siphophage sequences were compara- tively more insular. Podoviruses, which include the T7- likephages,werefoundlargelyintemperateareassuchas along the eastern coasts of Canada and the US. The siphoviruses(e.g.l-likephages),whileparticularlydomi- nant at single sites, were found haphazardly across the transect and showed no obvious biogeographical pattern or association with any abiotic parameter [36 !! ]. 584 Genomics Table 1 Different approaches for measuring patterns in phage biogeography Approach Analysis method Benefits Drawbacks Publications of interest Morphology and relative abundance Electron and epifluorescence microscopy Large number of samples for comparison Relatively cheap and easy Cannot estimate total viral diversity No direct functional information provided [49] [6] [8] [50] Fingerprinting PFGE RAPD Large number of samples for comparison Relatively cheap and easy genomes can be evaluated later No PCR bias for PFGE Novel virus discovery Cannot estimate total viral diversity No direct functional information provided [15] [19] Target gene sequencing PCR of target sequence phylogenetic analysis of sequence Large number of samples for comparison Relatively cheap and easy High taxonomic resolution Novel virus discovery Cannot estimate total viral diversity No direct functional information provided [22] [25] Viromics Shotgun sequencing Large amount of sequence data Diversity analysis possible Functional analysis possible Novel virus discovery Sequence analysis difficult Contamination issues Relatively expensive Large number of unknowns [28] [51] Genomics Culture of host and viral isolation shotgun sequencing Small genomes amendable to assembly Functional analysis possible Future empirical experimentation possible Novel virus discovery Few hosts and viruses can be cultivated Sequence analysis difficult Large number of unknowns [52] [53] Current Opinion in Microbiology 2009, 12:582–587 www.sciencedirect.com 10 Figure 1-4. Methodological pipeline for studying aquatic microbial diversity. Grey boxes indicate the type of diversity that can be assessed with the different methods. Modified caption and figure from Zinger et al (2012). Similar methods apply for bacterial and protistan community analysis – although typically focused on the universal ribosomal RNA gene (rRNA). Automated Ribosomal Intergenic Spacer Analysis (ARISA) and TRFLP are among the most common methods to assess genetic variation, but next generation sequencing has become increasingly popular. These methods similarly result in: 1) a list of present operational taxonomic units (OTUs) and 2) the relative abundance of each OTU by sequence counts or peak area for each sample. The detection limit of ARISA is approximately 1/1000 and results in OTUs that approximate species-level identification. This research focused on the use of molecular fingerprints to identify relevant patterns from beta-diversity, such as shifts in community composition between samples. Coupling of molecular analysis with environmental data has facilitated an improved understanding of the ecological implications of spatial and temporal variation in microbial community structure (Martiny et al. 2006; Pommier et al. 2007; Fuhrman et al. 2008; Zinger et al. 2011; Giovannoni & Vergin 2012). 11 Exploring microbial relationships by association networks Figure 1-5. Overview association network of local similarity correlations from three-year time-series of bacterial, T4-like viral, and protistan communities. Node colors represent (clockwise from top left): physical oceanographic parameters (n=6), yellow; biological and chemical oceanographic parameters (n=21), green; protists (n=60), blue; T4-like viruses (n=168), grey; bacteria (n=227), red. Each line represents a statistically significant LS correlation. Ecological networks have a long tradition in terrestrial ecology to study trophic interactions, such as plant-pollinators or other food web interactions (Montoya et al. 2006; Allesina & Pascual 2008; Ings et al. 2009; Olesen et al. 2011). Positive interactions were inferred to represent mutualism or predation while negative interactions suggested competition (Zhang et al. 2011). An overabundance of positive correlations suggested stable co-existence of organisms (Allesina & Pascual 2008; Zhang et al, 2011). Recent studies have begun to apply network theory to the microbial realm (Chaffron et al. 2010; Steele et al. 2011; Eiler et al. 2012; Gilbert et al. 2012). Collectively, these studies have identified complex relationships within the microbial communities and tribes (clusters) of bacteria that tend to co-occur. These diverse microbial 12 communities typically result in complex networks with thousands of interactions between species or OTUs (Figure 1-5). Application of network theory still has much to reveal in order to fully deconstruct the intricate ecological relationships between viruses, bacteria, eukaryotes, and their environment. These results could be used to highlight unexpected, but important relationships that seed further exploratory research (e.g. find the proverbial needle in the haystack). Capturing the Details with Viral Metagenomics Metagenomics, eco-genomics, or environmental genomics are all terms that recognize the study of all genomes of an environmental population versus an individual. As metagenomics is not dependent on culturing, it allows a look into the vast uncultured fragment of the population and provides information on potential biological activity and taxonomic identification. Through metagenomics, we can question who is present in the marine viral community as well as their roles in the environment. Breitbart et al (2002) published the first marine viral metagenomes from two coastal ocean locations. All major dsDNA phage families were represented in both coastal marine metagenomes, which primarily matched to DNA and RNA polymerases, helicases, DNA maturation proteins, packaging genes, terminases and structural proteins. Unfortunately those annotations were a small fraction of the overall sequence data as over 65% of the sequences in both of these libraries were not significantly similar to any other reported sequences in GenBank. Re-analysis in 2005 and 2010 still found that most sequences still had no significant similarity to known sequences despite the doubling of the GenBank database (Edwards & Rohwer 2005; Kristensen et al. 2010). Through the advance of genomic and metagenomic methods, phages are also thought to directly and influence nutrient cycling in the ocean. Viral metagenomes from around the world have yielded an improved understanding of the diversity and abundance of marine viral families, 13 and identification of genes initially thought to be exclusive to cellular hosts (Angly et al. 2006; Culley & Lang 2006; Kristensen et al. 2010; Goldsmith et al. 2011; Sharon et al. 2011; Breitbart 2012) Phages are now known to encode host-derived metabolic genes that facilitate photosynthesis and other processes in host cells under infection; these genes have been observed repeatedly across viral families (Clokie et al. 2006; Sullivan et al. 2006; Lindell et al. 2007; Goldsmith et al. 2011; Monier et al. 2011). The continued addition of new phage libraries and sequences to GenBank will facilitate future identification and characterization of viral communities. This dissertation research, thus, sought to not only characterize viruses from a new environment, but also to complete the first, to our knowledge, depth profile of viral metagenomes from the marine environment. 14 Chapter One References Allesina, S. & Pascual, M. (2008). Network structure, predator–prey modules, and stability in large food webs. Theor Ecol, 1, 55–64. Angly, F.E., Felts, B., Breitbart, M., Salamon, P., Edwards, R.A., Carlson, C.A., et al. (2006). The marine viromes of four oceanic regions. Plos Biol, 4, e368. Angly, F.E., Rodriguez-Brito, B., Bangor, D., McNairnie, P., Breitbart, M., Salamon, P., et al. (2005). PHACCS, an online tool for estimating the structure and diversity of uncultured viral communities using metagenomic information. BMC Bioinformatics, 6, 41. Azam, F., Fenchel, T., Field, J., Gray, J., Meyer-Reil, L. & Thingstad, F. (1983). The ecological role of water-column microbes in the sea. Mar. Ecol. Prog. Ser., 10, 257–263. Beman, J.M., Sachdeva, R. & Fuhrman, J.A. (2010). Population ecology of nitrifying Archaea and Bacteria in the Southern California Bight. Environ Microbiol, 12, 1282–1292. Beman, J.M., Steele, J.A. & Fuhrman, J.A. (2011). Co-occurrence patterns for abundant marine archaeal and bacterial lineages in the deep chlorophyll maximum of coastal California. ISME J, 5, 1077–1085. Bergh, Ø., Børsheim, K.Y., Bratbak, G. & Heldal, M. (1989). High abundance of viruses found in aquatic environments. Nature, 340, 467–468. Bouvier, T. & del Giorgio, P.A. (2007). Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol, 9, 287–297. Bratbak, G., Thingstad, F. & Heldal, M. (1994). Viruses and the microbial loop. Microb Ecol, 28, 209–221. Breitbart, M. (2002). Genomic analysis of uncultured marine viral communities. Proc. Natl. Acad. Sci. U.S.A., 99, 14250–14255. Breitbart, M. (2012). Marine Viruses: Truth or Dare. Annu. Rev. Marine. Sci., 4, 425–448. Breitbart, M. & Rohwer, F. (2005). Here a virus, there a virus, everywhere the same virus? Trends Microbiol., 13, 278–284. Brussaard, C.P.D., Wilhelm, S.W., Thingstad, F., Weinbauer, M.G., Bratbak, G., Heldal, M., et al. (2008). Global-scale processes with a nanoscale drive: the role of marine viruses. ISME J, 2, 575–578. Chaffron, S., Rehrauer, H., Pernthaler, J. & Mering, von, C. (2010). A global network of coexisting microbes from environmental and whole-genome sequence data. Genome Research, 20, 947–959. Clokie, M.R.J., Millard, A.D. & Mann, N.H. (2010). T4 genes in the marine ecosystem: studies of the T4-like cyanophages and their role in marine ecology. Virol J, 7, 291. 15 Clokie, M.R.J., Shan, J., Bailey, S., Jia, Y., Krisch, H.M., West, S., et al. (2006). Transcription of a “photosynthetic” T4-type phage during infection of a marine cyanobacterium. Environ Microbiol, 8, 827–835. Comeau, A.M. & Krisch, H.M. (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, 1321–1332. Countway, P.D. & Caron, D.A. (2006). Abundance and distribution of Ostreococcus sp. in the San Pedro Channel, California, as revealed by quantitative PCR. Appl. Environ. Microbiol., 72, 2496–2506. Countway, P.D., Vigil, P.D., Schnetzer, A., Moorthi, S.D. & Caron, D.A. (2010). Seasonal analysis of protistan community structure and diversity at the USC Microbial Observatory (San Pedro Channel, North Pacific Ocean). Limnol Oceanogr., 55, 2381–2396. Culley, A. & Lang, A. (2006). Metagenomic Analysis of Coastal RNA Virus Communities. Science. Ducklow, H. (1983). Production and Fate of Bacteria in the Oceans. Bioscience. Edwards, R.A. & Rohwer, F. (2005). Viral metagenomics. Nat Rev Micro, 3, 504–510. Eiler, A., Heinrich, F. & Bertilsson, S. (2012). Coherent dynamics and association networks among lake bacterioplankton taxa. ISME J, 6, 330–342. Fenchel, T. (2008). The microbial loop – 25 years later. J. Exp. Mar. Biol. Ecol., 366, 99–103. Filee, J. (2005). Marine T4-type bacteriophages, a ubiquitous component of the dark matter of the biosphere. Proc. Natl. Acad. Sci. U.S.A., 102, 12471–12476. Fuhrman, J.A. (1999). Marine viruses and their biogeochemical and ecological effects. Nature, 399, 541–548. Fuhrman, J.A. & Suttle, C.A. (1993). Viruses in marine planktonic systems. Oceanography, 6, 51–63. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Fuhrman, J.A., Steele, J.A., Hewson, I., Schwalbach, M.S., Brown, M.V., Green, J.L., et al. (2008). A latitudinal diversity gradient in planktonic marine bacteria. Proc. Natl. Acad. Sci. U.S.A., 105, 7774–7778. Fuller, N.J., Wilson, W.H., Joint, I.R. & Mann, N.H. (1998). Occurrence of a sequence in marine cyanophages similar to that of T4 g20 and its application to PCR-based detection and quantification techniques. Appl. Environ. Microbiol., 64, 2051–2060. Gilbert, J.A., Steele, J.A., Caporaso, J.G., Steinbrück, L., Reeder, J., Temperton, B., et al. (2012). Defining seasonal marine microbial community dynamics. ISME J, 6, 298–308. 16 Giovannoni, S.J. & Vergin, K.L. (2012). Seasonality in Ocean Microbial Communities. Science, 335, 671–676. Goldsmith, D.B., Crosti, G., Dwivedi, B., McDaniel, L.D., Varsani, A., Suttle, C.A., et al. (2011). Development of phoH as a Novel Signature Gene for Assessing Marine Phage Diversity. Appl. Environ. Microbiol., 77, 7730–7739. Hewson, I. & Fuhrman, J.A. (2006). Viral impacts upon marine bacterioplankton assemblage structure. J. Mar. Biol. Ass., 86, 577–589. Hoffmann, K.H., Rodriguez-Brito, B., Breitbart, M., Bangor, D., Angly, F.E., Felts, B., et al. (2007). Power law rank abundance models for marine phage communities. FEMS Microbiol Lett, 273, 224–228. Ings, T.C., Montoya, J.M., Bascompte, J., Blüthgen, N., Brown, L., Dormann, C.F., et al. (2009). Review: Ecological networks - beyond food webs. J Anim Ecology, 78, 253–269. Kristensen, D.M., Mushegian, A.R., Dolja, V.V. & Koonin, E.V. (2010). New dimensions of the virus world discovered through metagenomics. Trends Microbiol., 18, 11–19. Lindell, D., Jaffe, J.D., Coleman, M.L., Futschik, M.E., Axmann, I.M., Rector, T., et al. (2007). Genome-wide expression dynamics of a marine virus and host reveal features of co- evolution. Nature, 449, 83–86. Martiny, J.B.H., Bohannan, B.J.M., Brown, J.H., Colwell, R.K., Fuhrman, J.A., Green, J.L., et al. (2006). Microbial biogeography: putting microorganisms on the map. Nat Rev Micro, 4, 102–112. McDaniel, L. & Delarosa, M. (2006). Cambridge Journals Online - Abstract - Temperate and lytic cyanophages from the Gulf of Mexico. J. Mar. Biol. Ass. Monier, A., Welsh, R.M., Gentemann, C., Weinstock, G., Sodergren, E., Armbrust, E.V., et al. (2011). Phosphate transporters in marine phytoplankton and their viruses: cross-domain commonalities in viral-host gene exchanges. Environ Microbiol, 14, 162–176. Montoya, J.M., Pimm, S.L. & Solé, R.V. (2006). Ecological networks and their fragility. Nature, 442, 259–264. Olesen, J.M., Stefanescu, C. & Traveset, A. (2011). Strong, Long-Term Temporal Dynamics of an Ecological Network. PLoS ONE, 6, e26455. Pomeroy, L.R., Williams, P.I., Azam, F. & Hobbie, J.E. (2007). The Microbial Loop. Oceanography, 20, 28–33. Pommier, T., Canbäck, B., Riemann, L., Boström, K.H., Simu, K., Lundberg, P., et al. (2007). Global patterns of diversity and community structure in marine bacterioplankton. Mol. Ecol., 16, 867–880. Proctor, L.M. & Fuhrman, J.A. (1990). Viral mortality of marine bacteria and cyanobacteria. Nature, 343, 60–62. 17 Rodriguez-Brito, B., Li, L., Wegley, L., Furlan, M., Angly, F.E., Breitbart, M., et al. (2010). Viral and microbial community dynamics in four aquatic environments. ISME J, 4, 739–751. Sandaa, R.-A., Gómez-Consarnau, L., Pinhassi, J., Riemann, L., Malits, A., Weinbauer, M.G., et al. (2009). Viral control of bacterial biodiversity - evidence from a nutrient-enriched marine mesocosm experiment. Environ Microbiol, 11, 2585–2597. Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2004). Viral effects on bacterial community composition in marine plankton microcosms. Aquat. Microb. Ecol., 34, 117–127. Sharon, I., Battchikova, N., Aro, E.-M., Giglione, C., Meinnel, T., Glaser, F., et al. (2011). Comparative metagenomics of microbial traits within oceanic viral communities. ISME J, 5, 1178–1190. Sherr, E. & Sherr, B.F. (1988). Role of microbes in pelagic food webs: a revised concept. Limnol Oceanogr., 33, 1225–1227. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. Sullivan, M.B., Coleman, M.L., Quinlivan, V., Rosenkrantz, J.E., DeFrancesco, A.S., Tan, G., et al. (2008). Portal protein diversity and phage ecology. Environ Microbiol, 10, 2810–2823. Sullivan, M.B., Lindell, D., Lee, J.A., Thompson, L.R., Bielawski, J.P. & Chisholm, S.W. (2006). Prevalence and Evolution of Core Photosystem II Genes in Marine Cyanobacterial Viruses and Their Hosts. Plos Biol, 4, e234. Suttle, C.A. (2005). Viruses in the sea. Nature, 437, 356–361. Suttle, C.A. (2007). Marine viruses — major players in the global ecosystem. Nat Rev Micro, 5, 801–812. Thingstad, T.F. (2000). Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol Oceanogr., 1320–1328. Thurber, R.V. (2009). Current insights into phage biodiversity and biogeography. Curr. Opin. Microbiol, 12, 582–587. Wilhelm, S.W. & Suttle, C.A. (1999). Viruses and nutrient cycles in the sea. Bioscience, 49, 781– 788. Wilson, W.H., Fuller, N.J., Joint, I.R. & Mann, N. (2000). Analysis of cyanophage diversity in the marine environment using denaturing gradient gel electrophoresis. Microbial Biosystems: New Frontiers Proceedings of the 8th International Symposium on Microbial Ecology. Atlantic Canada Society for Microbial Ecology, 565–570. 18 Winter, C., Smit, A., Herndl, G.J. & Weinbauer, M.G. (2004). Impact of virioplankton on archaeal and bacterial community richness as assessed in seawater batch cultures. Appl. Environ. Microbiol., 70, 804–813. Wommack, K.E. & Colwell, R.R. (2000). Virioplankton: viruses in aquatic ecosystems. Microbiol. Mol. Biol. Rev., 64, 69–114. Wommack, K.E., Ravel, J., Hill, R., Chun, J. & Colwell, R.R. (1999). Population dynamics of chesapeake bay virioplankton: total-community analysis by pulsed-field gel electrophoresis. Appl. Environ. Microbiol., 65, 231–240. Yooseph, S., Sutton, G., Rusch, D.B., Halpern, A.L., Williamson, S.J., Remington, K., et al. (2007). The Sorcerer II Global Ocean Sampling Expedition: Expanding the Universe of Protein Families. Plos Biol, 5, e16. Zhang, W.J., Aldrich, P.R., Tacutu, R., Budovsky, A., Yanai, H., Wolfson, M., et al. (2011). Constructing ecological interaction networks by correlation analysis: hints from community sampling. Network Biology, 1, 81–98. Zinger, L., Amaral-Zettler, L.A., Fuhrman, J.A., Horner-Devine, M.C., Huse, S.M., Welch, D.B.M., et al. (2011). Global Patterns of Bacterial Beta-Diversity in Seafloor and Seawater Ecosystems. PLoS ONE, 6, e24570. 19 Chapter Two: Temporal variability and coherence of bacteria communities from a ten-year time-series in the euphotic zone of the Southern California Bight Abstract Time-series have become extremely important as biological surveys and for our understanding of the ocean ecosystem. To characterize the natural variability of bacterial communities in the euphotic zone, bacterial populations were investigated for ten-years at the USC Microbial Observatory at the San Pedro Ocean Time-series (SPOT) off the coast of Southern California. Community structure was determined by Automated Ribosomal Intergenic Spacer Analysis (ARISA) and coupled with measurements of oceanographic parameters for two discrete depths (surface, 0-5m and deep chlorophyll maximum (DCM), average depth 30m). Although both depths displayed predictable seasonality by discriminant function analysis, the surface exhibited more pronounced seasonal variation than the DCM by Bray-Curtis community similarity. Inter-depth Bray-Curtis similarity repeatedly decreased as the water column stratified each summer. SAR11 and cyanobacterial ecotypes each comprised up to one-third or more of the measured community and were temporally variable, similar to observations at other long-term time-series sites. Association networks based on local similarity correlations revealed a complex structure of interactions between bacterial OTUs and environmental parameters from which we identified potential niches and analysis of shared LS interactions identified consistent correlations between depths. Persistent OTUs tended to also be the most abundant by relative contributions, yet ephemeral OTUs outnumbered the moderately frequent and persistent OTUs. We defined a potential core microbiome for the euphotic zone at SPOT based on persistence and average relative abundance of several OTUs that included several taxonomic groups. 20 Introduction Investigations into temporal dynamics of marine microbial communities have revealed some remarkable similarities and dissimilarities between ocean basins and insight into the complex ecology of the microbial realm (Ducklow et al. 2009; Fuhrman 2009; Giovannoni & Vergin 2012). A holistic understanding of microbes in the ocean requires knowing not just who is present, but also when they occur and how they contribute to the community. Knowledge of each of these parameters will improve our models of the microbial loop and understanding of microbial roles in the sea. Seasonal and monthly patterns of variation have been observed using molecular methods for whole microbial communities at many aquatic time-series sites around the world. Many, but not all, have shown seasonal and/or recurring annual communities (Acinas et al. 1997; Li 1998; Morris et al. 2005; Fuhrman et al. 2006; Alonso Sáez et al. 2007; Kan et al. 2007; Treusch et al. 2009; Campbell et al. 2011; Caporaso et al. 2011; Eiler et al. 2011; Gilbert et al. 2012; Robidart et al. 2012). Temporal surveys of specific bacterial groups suggest an environmentally selected response to changing conditions. Seasonality in Synechococcus ecotypes was observed in the Southern California Bight, for clades I and IV (Tai & Palenik 2009) and in Chesapeake Bay (Cai et al. 2010). Prochlorococcus ecotypes at the Hawaii Ocean Time-series (HOT) and Bermuda Atlantic Time-series Study (BATS) exhibited similar depth distribution patterns at each site, aside from the annual deep water column mixing events at BATS; however, annual cycling patterns for each Prochloroccus ecotype was less apparent at HOT than at BATS (Treusch et al. 2009; Malmstrom et al. 2010). SAR11 ecotypes were dominant members of microbial communities and similarly variable over time (Morris et al. 2002; Carlson et al. 2009; Eiler et al. 2009; Campbell et al. 2011; Giovannoni & Vergin 2012). Functional genes for ammonia-oxidizing archaea and bacteria also co-varied and re-occurred in tandem with specific environmental parameters such as 21 nitrite or ammonium concentrations (Beman et al. 2010; Bouskill et al. 2010). The application of high-throughput “next-generation” sequencing has increased the resolution with which we can characterize communities (compared to prior fingerprinting surveys) – although many of our earlier observations on seasonality remained essentially the same. In this study, we examined bacterial community structure in light of traditional ecological metrics for community membership, composition, phylogeny, persistence, and connectivity (reviewed in Shade & Handelsman 2012). Research at the San Pedro Ocean Time-series site (SPOT) has focused on determining the temporal variability of bacterial, archaeal, and protistan taxa and the development of ecological networks that link individual members of these microbial communities (Countway & Caron 2006; Fuhrman et al. 2006; Beman et al. 2010; Countway et al. 2010; Beman et al. 2011; Steele et al. 2011). Here, we assessed the inter-annual, seasonal and monthly variability in overall bacterial community composition and specific taxonomic groups in the surface water and deep chlorophyll maximum (DCM), as identified by Automated Ribosomal Intergenic Spacer Analysis (ARISA), as part of the USC Microbial Observatory program at SPOT. Investigation into the prevalence of seasonal trends for the bacterial community at large and specific taxonomic groups revealed complex patterns. Seasonality and annual recurrence of bacterial communities was significant in the surface ocean, but seasonality was less apparent in the DCM. We identified OTUs that were a) common and abundant, b) moderately frequent and rarely to moderately abundant or c) infrequent and rarely abundant. Microbial association networks exhibited significant interconnectedness and ‘small-world properties’ based on observed network topology. An improved understanding of the community composition and which constituents are persistent, intermittent, or ephemeral (as well as seasonal) will help us to understand processes controlling the distribution of organisms and to predict future community response to changes in the ocean. 22 Materials and Methods Sample Collection Seawater was collected approximately monthly from August 2000 to January 2011 at 0- 5m and the deep chlorophyll maximum depth, DCM (average 28m, range 7-45m), as determined from in situ fluorescence, as part of the University of Southern California Microbial Observatory at the San Pedro Ocean Time-series (33° 33’ N, 118° 24’ W). No data were available for a few months due to inclement weather, equipment malfunction, or occasional issues with ARISA amplification; in total, 103 months were sampled at 0-5m and 89 months at DCM over a 126- month span. Seawater was collected (~10L) and serially filtered through 142mm Type A/E glass-fiber filter (Pall Life Sciences; Ann Arbor, MI) and a 0.22 µm Durapore (Whatman, GVWP) by positive pressure (10-15psi) to collect prokaryotic cells; filters were stored at -80ºC. Bacterial production by tritiated thymidine and leucine incorporation, bacterial and viral abundance by SYBR Green I epifluorescence microscopy and other environmental parameters (nutrients, chlorophyll a) were determined as previously described (Brown et al. 2005; Fuhrman et al. 2006; Patel et al. 2007; Beman et al. 2010; 2011; Steele et al. 2011). DNA extraction, Bacterial Community Fingerprinting, and Taxonomic Assignment DNA extraction and amplification. Briefly, DNA was extracted by hot SDS lysis followed by phenol-chloroform extraction and then stored in TE buffer or dry at -80ºC. DNA was measured with QuantIt Picogreen (Invitrogen; Grand Island, NY) on a Stratagene MX3000. PCR amplification for ARISA (Fisher & Triplett 1999; Brown et al. 2005) consisted of the following final concentrations per reaction (50µl): 2ng DNA, 1X Buffer, 2.5mM MgCl 2 , 0.2mM each DNTP, 0.2mg/ml BSA, 0.4µM each primer, and 0.5 unit AmpliTaq Gold (Applied Biosystems). 23 Thermocycling was preceded by 5:00 at 95ºC, followed by 30 cycles of 40 seconds each at 95ºC, 56ºC and 90 seconds at 72ºC, with a final 7 minute extension at 72ºC. DNA samples were re- amplified for this study and successful amplification was verified by agarose gel electrophoresis. PCR products were concentrated to 10ul with the Zymo Research Clean & Concentrator TM -5, per manufacturer instructions (Zymo Research; Irvine, CA) and quantified by Picogreen. Fragment Analysis. Each PCR product was analyzed in duplicate on non-adjacent lanes by slab gel acrylamide electrophoresis (ABI 377, Applied Biosystems). 2.5µl of HiDi formamide (Lonza), 0.5µl of loading dye, and 1µl of internal size standards were added to 1µl at 10ng/µl of PCR product. Peak data were analyzed in DAx and standards were interpolated by the local southern method (Chow & Fuhrman 2012). Replicates were dynamically binned with a maximum bin size of 1bp (390-450bp), 2bp (450-650bp), 3bp (650-900bp) and 5bp (900- 1200bp), as previously described (Fuhrman et al. 2006; Ruan et al. 2006b; Steele et al. 2011). Bins were manually curated to merge any bin <0.1bp with the nearest neighbor. Relative abundances reflect the proportional area of an individual peak divided by total measured peak area in each fingerprint. ARISA Peak Identification. Each ARISA bin was identified where possible based on 16S-ITS sequences from one of several sources: 1) observed ARISA length of SPOT clones from surface water, 2) published cyanobacterial ITS sequences, 3) in silico amplification of ARISA products from whole genome sequences (curated from marine habitat and euphotic zone or surface waters), 4) observed ARISA lengths of 16S-ITS clones constructed from sites in the (a) central Pacific and (b) tropical Atlantic Oceans, 5) observed ARISA lengths from 16S-ITS clones isolated from the Indian Ocean, 6) in silico amplification of ARISA products from assembled rRNA-containing scaffolds from the Global Ocean Survey (GOS), 7) observed ARISA length of SPOT clones from a) 150m and b) 890m, 8) in silico amplification of ARISA products from whole genome 24 sequences for isolates originating from below the euphotic zone, and lastly 9) in silico amplification of ARISA products from GOS rRNA-containing scaffolds from below the surface ocean. Sequences in this analysis included previously published sequences from SPOT (Brown & Fuhrman 2005; Brown et al. 2005), marine cyanobacteria (Rocap et al. 2002) and bacterial genomes (Kottmann et al. 2009). SPOT, Atlantic Ocean, Pacific Ocean, and Indian Ocean 16S- ITS clones were amplified and analyzed for ARISA, similar to environmental samples, in order to determine a measured ARISA length. Calculated ARISA lengths were determined for all sequences where no measured lengths were available; calculation was based on a linear interpolation of all observed lengths versus counted empirical lengths in base pairs. Two interpolations were determined: one for lengths from 400-800bp and the second for lengths from 800-1200bp due to decreased sizing accuracy after 800bp. All sequences were searched by BLAST against Greengenes (McDonald et al. 2011), Ribosomal Database Project, RDP, (Cole et al. 2009), and the SILVA 108 SSU Full and 16S databases (Pruesse et al. 2007). A hybrid taxonomy for each clone or sequence was constructed for the top hit (by e-value) consisting of: 1) explicit taxonomic ranks from Greengenes, 2) a SILVA identifier from the lowest taxonomic level identified, excluding ‘uncultured’ or ‘unidentified’, 3) SAR11 clade designations from RDP release 10, and 4) cyanobacteria ecotype based on previous publications for ITS sequences and known isolates (Rocap et al. 2002; Brown & Fuhrman 2005) or phylogenetic placement of 16S rDNA sequences. All sequences from SPOT were aligned with the SINA WebAligner and then added by parsimony to SILVA 108 SSU Ref tree (Pruesse et al. 2007). Sequence alignment was manually refined to allow for identification of cyanobacterial ecotypes. An ARISA bin was assigned to any sequence whose measured or calculated lengths were within the lower and upper bounds (bins are numbered by their upper bound). Sequences were matched, in order of the sources listed above, with a preference towards surface water as the original isolation source. 25 Once a bin received an identity, any matches from lower priority sources were ignored. If multiple sequences within one source were assigned to a bin, the most abundant taxonomy was kept; taxonomies were manually merged if equally abundant (i.e. two different whole genomes had nearly identical ARISA lengths). Data Analysis Statistics. Community composition data were normalized by total peak area within each month for each depth individually, prior to any analysis; relative abundances (also referred to as percent contributions) are presented here. Both Bray-Curtis similarity and SIMPER (determines total contribution of each OTU to overall Bray-Curtis similarities) were calculated in PRIMER-E v6 (Plymouth, UK: Clarke 1993; Clarke & Gorley 2006). Discriminant function analyses (DFA) and time-series plots were calculated in Systat 11 using the 50 most important OTUs, as determined in PRIMER based on the OTU’s relative abundance and occurrence. Global correlations between Bray-Curtis similarities and other parameters were determined in Sigmaplot 11. Network Analysis. We used Local Similarity Analysis (eLSA) to define local intervals of correlation between OTUs and environmental parameters with 2000 permutations and linear interpolation of missing values (Ruan et al. 2006a; Xia et al. 2011). Permutation tests allow for randomization of the order of months to check for spurious correlations and provide significance values. We excluded any parameter that occurred in 4 or less months and any eLSA correlations with p-value >0.01 or q-value >0.05 to minimize spurious correlations. Q-values provide a false discovery rate, based on the distribution of p-values (Storey 2002). Significant eLSA correlations were visualized in Cytoscape (v2.8.2); a network was calculated for each sampling depth independently. We determined network statistics for undirected networks with Network Analyzer (Assenov et al. 2008) and used ‘Advanced Network Merge’ to determine similarities and 26 differences between the 0-5m and DCM association networks. Random undirected networks of equal nodes and edges to our observed networks were constructed by the Erdős–Rényi model using default settings in the Random Network plugin for Cytoscape; statistics for these random undirected networks were calculated as before (Steele et al. 2011). Interconnected modules were identified with the AllegroMCODE plugin using the default settings (Yoon and Jung, 2011). Results and Discussion Annual environmental variability in the surface ocean and deep chlorophyll maximum Environmental parameters suggest repeated seasonal stratification and influx of nutrients from nearby upwelled waters (Figure 2-1). Seasonal stratification was demonstrated by increased sea surface temperature in late summer up to 20ºC and a deepening of the average DCM depth in June (32m) and September (33.5m). Sea surface temperature consistently peaked in August and September at about 20ºC, with average DCM temperatures at 14ºC. Average DCM depth was 28.2m and ranged from 7 - 45m during the months sampled. The DCM was below the mixed layer depth from May through November (data not shown). During winter mixing from December to March, mixed layer depth was equal to or slightly below the DCM and did not rise above the DCM until April. Both bacterial abundance and production increased in April although higher nutrient concentrations were not observed in the DCM until May or later (Figure 2-1). Nutrient concentrations began a gradual increase in February in most years and peaked in May-June before declining by October. Bacterial production rates in both depths peaked around a million cells/ml/day in April. These rates were double to ten-fold higher than was measured across the rest of the year (between 1.3x10 5 to 5.5x10 5 cells/ml/day), where the lowest rates occurred in winter. Bacterial abundance ranged from 1.2 - 3x10 6 cells/ml, with a significant increase from 27 March to April; abundance was similar between depths, except for July. Viral abundance varied little from month to month, within the same order of magnitude (2x10 7 – 4x10 7 particles/ml). This observation is in contrast with the repeated seasonal trends seen in viral abundance at BATS (Parsons et al. 2012). Regional upwelling has been previously linked to local productivity and particle export at SPOT from a 4 year study (January 2004-December 2007) that lies within our sampling period (Collins et al. 2011). Onset of seasonal stratification typically occurred in late spring and coincided with the period of highest bacterial productivity and abundance – activity likely due to water column stabilization following the increased nutrient availability from winter-spring mixing events. Although the mixed layer depth at SPOT is much shallower than either HOT or BAT, this is due in part to regional hydrography and location closer to shore. The seasonal position of the mixed layer depth relative to the DCM at SPOT lies between the trends observed at two other long-term time-series sites (Giovannoni & Vergin 2012): mixing is not as pronounced as at BATS nor is stratification as pronounced as at HOT. 28 Figure 2-1. Environmental variability over ten years in the euphotic zone at SPOT. Monthly averages are shown and error bars indicate standard error of the mean for: a) bacterial production by leucine (Leu) and thymidine (Tdr) incorporation, b) bacterial and viral abundance, c) DCM sampling depth, d) temperature, e) salinity, f) phosphate, g) nitrite, and h) nitrate, and h) phosphate. Filled symbols, 0-5m; open symbols, DCM. 29 Monthly variability of individual bacterial OTUs Over ten years, ~400 unique OTUs were observed in each depth (0-5m: 407; DCM: 396). An average of 106 ± 2 OTUs were detected each month (range: 54-174, 0-5m; 57-162, DCM). It is important to note that although ARISA has inherent biases as a PCR-based method, we primarily assessed changes in the relative contributions of each peak to the overall fingerprint. The relative contribution for each peak to the measured bacterial community fingerprint was determined, with a detection limit of 0.1% of the community. Any PCR biases are reflected in the percentages themselves; with that said, changes in relative abundance of ARISA OTUs identified as Prochloroccus did parallel changes in cellular abundance assessed by flow cytometry (Brown et al. 2005). Subsequent analyses on overall changes in community composition were based on pattern detection and were not affected by these inherent methodological limitations. In this study, we utilized ARISA to detect OTUs (peaks) in a semi-quantitative manner to describe temporal shifts in community composition and correlations between OTUs. Only a few OTUs consistently dominated the ARISA fingerprint. The five most abundant ARISA OTUs in the euphotic zone are: Actino_435.5 (Class: Actinobacteria), SAR11_S1 666.4, SAR11 662.0, SAR11_S1 686.9, and SAR11_S1 670.5, in descending order of average relative abundance. All five were detected in over 75% of the months observed (n > 91, 0-5m; 79, DCM). These five OTUs on average comprised one-third of the measured bacterial community and up to a maximum of 63.8% (0-5m) or 76.7% (DCM) in a single month, as indicated by relative contribution to the ARISA community fingerprint (Figure 2). Relative contributions of each OTU were significantly correlated between surface and DCM although the coefficients ranged from 0.3-0.7 (p<0.05, Table 2-S1). These five most abundant OTUs characterize over 50% of the observed similarity within each depth (52.1%, 0-5m; 55.4%, DCM), and as such are key members of the euphotic zone microbial communities 30 Table 2-S1. Correlation of the relative abundance of the top 5 OTUs between 0-5m and DCM. S1 indicates Surface Clade 1 for the SAR11 ecotype. OTU Pearson Spearman OCS155_435.5 Correlation 0.312 0.362 p-value 0.00812 0.00203 n 71 71 SAR11_662 Correlation 0.486 0.507 p-value 0.00002 9.13E-06 n 70 70 SAR11_S1_666.4 Correlation 0.721 0.498 p-value 2.03E-13 5.87E-06 n 76 76 SAR11_S1_670.5 Correlation 0.3 0.334 p-value 0.00809 0.00308 n 77 77 SAR11_S1_686.9 Correlation 0.593 0.508 p-value 5.15E-08 7.56E-06 n 71 71 31 Figure 2-2. Inter-annual and monthly variability of top 5 ARISA OTUs over ten-years. Proportion (y- axis) is denoted for each OTU over time (x-axis), so that each column reflects the cumulative relative abundance of all five OTUs within the measured bacterial community each month. Each OTU shown is represented by a unique color. Letters indicate month (x-axis): September (S), December (D), March (M), and June (J). Gaps indicate months where ARISA data were unavailable (i.e. for 2007 in DCM). 32 Figure 2-S1. Monthly variability in cyanobacteria relative abundance at SPOT. (A) Cumulative relative abundance of all cyanobacterial OTUs for 0-5m (dark teal) and DCM (light green). Individual monthly contributions of each OTU as a proportion of the measured cyanobacterial contributions (i.e. relative abundance of OTU 828.8 divided by total cyanobacterial contribution that month) is shown for (b) 0-5m and (c) DCM. Each OTU is identified by a unique color, as shown; Prochlorococcus (Pro) OTUs are green or yellow, while Synechococcus OTUs are red, pink, or brown. Letters indicate month (x-axis): September (S), December (D), March (M), and June (J). Gaps indicate months where ARISA data were unavailable. 33 Figure 2-S2. Monthly and inter-annual variability of SAR11 ecotypes in (a) the surface ocean and (b) DCM. Each SAR11 OTU is marked with a unique color; text label indicates SAR11_clade _fragment length, where available. The y-axis denotes the relative abundance of each ecotype, and it is cumulative across OTUs so that total bar height reflects overall proportion of SAR11 in the measured community. Letters indicate month (x-axis): September (S), December (D), March (M), and June (J). 34 Cyanobacteria collectively were less than a third of the measured community, with a monthly average of only 4.7% (0-5m) or 2.0% (DCM). The maximum total contribution of cyanobacteria observed was 34.2% (5m) and 19.1% (DCM) of the community in any given month (Figure 2-S1). Increased cyanobacterial contributions to the community in 0-5m tended to occur in the latter half of the year as the DCM depth deepened. Although individual OTU contributions varied monthly in the DCM, seasonal trends were less apparent. Cyanobacterial OTUs were assigned to different ecotype clades, such as presented in Rocap et al (2002) and Brown & Fuhrman (2005), based on phylogenetic assignment of the clone sequences that identified each ARISA bin. High-light Prochlorococcus OTUs were the largest contributors overall, and one high-light Prochloroccus ecotype from clade I, OTU 828.8, dominated the cyanobacterial community in most months in the surface ocean, and shared dominance in the DCM with another high-light clade I ecotype, OTU 831.8. A low-light Prochlorococcus OTU (912.5) was a sporadically high proportion of the DCM communities in the latter half of 2003- 2009. Although present year-round, Synechococcus peaked typically between March – June; these episodes were concurrent with periods of high productivity. Synechococcus also had episodic peaks in abundance in the fall of 2000 and 2007 in 0-5m and was exceptionally abundant at 4.5% of the community in April 2002 (DCM). Cumulative relative abundance of SAR11 OTUs was greater than cumulative abundance of all cyanobacterial OTUs at any given date – 35.8% (0-5m) and 31.9% (DCM) on average and up to 66.3% (0-5m) and 62.5% (DCM) of the community in a single month (Figure 2-S2). The dominant OTUs were from the Surface Clade 1 (666.4, 670.5, and 686.9) and SAR11 662. The remaining SAR11 Surface 1 OTUs (667.6 and 692.2) were observed in less than half of the months and relatively minor contributors to the overall community structure (<2%). SAR11 Surface 2 clades were observed infrequently and typically <5%. A consistently minor contributor 35 was SAR11_S4_703.7, which occurred in 86 months but at <1% on average (maximum contribution in any month, 7.3%) in 0-5m and in 72 months at <1% (maximum of 3.8%) in DCM. The overall relative contributions of these two key microbial groups (SAR11 and cyanobacteria) were consistent with temporal variation and abundances observed in microbial communities elsewhere, yet there was location-based variability (Morris et al. 2005; Kan et al. 2007; Carlson et al. 2009; Tai & Palenik 2009; Cai et al. 2010; Malmstrom et al. 2010; Gilbert et al. 2012). The SAR11, cyanobacteria, and Actinobacteria OTU 435.5 together comprised about one half of the measured bacterial community on average (50.8% (0-5m), 47.7% (DCM)), but up to 77% and 81.8% in a given month, respectively. Synechococcus OTUs were observed year- round, however, our seasonal spring increases in abundance following upwelling events differed from reported decreases in abundance after spring upwelling events in Monterey Bay (Paerl et al. 2012). SAR11 relative abundances, estimated by ARISA, were comparable to the mean contribution of 38% for SAR11 in the photic zone at HOT, as observed by quantitative PCR (Eiler et al. 2009). Peaks in cumulative SAR11 relative abundances were generally seen in late summer, similar to BATS, and opposite the winter peaks observed at HOT and the Western English Channel (Giovannoni & Vergin 2012). Ten-year seasonal and annual trends in overall community composition We previously identified predictable seasonal differences and annual recurrence of communities by discriminant function analysis (DFA) from 4.5 years in the surface ocean (Fuhrman et al. 2006). DFA relies on selecting individual OTUs from the whole community that are capable of distinguishing months. Here we found that prevailing seasonality was maintained in over ten years of community composition observations from both the surface and DCM depths (Figure 2-S3). Surface ocean and DCM communities were significantly positively auto-correlated at one month and negatively auto-correlated at 4-6 months, based on ten years of data. Positive 36 autocorrelation was also statistically significant at ten months in the surface ocean (samples are considered sequentially and the analysis ignores missing data, so the length of the periodicity was reduced). Although there is a sinusoidal trend in autocorrelation with a positively auto-correlated peak at 10-11 months in the DCM, the individual correlation coefficients at 10-11 months were not statistically significant. Missing data for several months interspersed throughout the time- series and the significant gap (10/2006-2/2008) were ignored in this analysis, which may have limited our ability to correctly interpret these long-term trends. Analysis by DFA for the DCM from 8/2000-9/2006 resulted in similar patterns: positive autocorrelation at one month and negative autocorrelation at 4-6 months lag. None of the datasets tested revealed statistically significant positive autocorrelation at one year in the DCM, although the sinusoidal trend was always apparent; datasets selected for DFA included: all OTUs, most abundant OTUs, and OTUs occurring in more than 70 months. An alternative conclusion is that the DCM communities were less variable across seasons when compared to the surface ocean, so that no OTU (or combination of) could adequately differentiate months from one another. In total, seasonality occurred in both depths, but only surface communities exhibited annual predictability that could be detected by DFA. 37 Figure 2-S3. Time-series and Discriminant Function Analyses from 50 highest contributing OTUs. Time-series plot of first discriminant function score where solid blue is raw score and red is moving average of 3 cases (left column: A, C, E). Autocorrelation of the first discriminant function based on the moving average (right column: B, D, F). Top row: 0-5m all available dates; middle row: DCM: all available dates; bottom row: DCM, 2000-2006 only. 38 Figure 2-3. Seasonal and inter-annual patterns in Bray-Curtis community similarity. Average pairwise community similarity was calculated from all OTUs for all months in 0-5m (upper) and DCM (lower). Time lag (x-axis) indicates the number of months between the communities compared. Figure 2-S4. Seasonal and inter-annual patterns in Bray-Curtis community similarity at 0-5m for (a) the 50 most important OTUs and (b) those OTUs occurring in more than 50 months. 0-5m 0 12 24 36 48 60 72 84 96 108 120 Avg. Bray-Curtis Similarity 34 36 38 40 42 44 46 48 50 52 DCM Time lag (months) 0 12 24 36 48 60 72 84 96 108 120 25 30 35 40 45 50 55 60 39 Community similarity by Bray-Curtis identified seasonal and annual patterns in bacterial community structure similar to those revealed by DFA (Figure 2-3, 2-S4). Bray-Curtis similarity is weighted by the relative abundance of each OTU and includes all OTUs unlike DFA. Comparison of technical replicates of environmental samples differ by up to 15% by Bray-Curtis similarity due to slight variations in PCR efficiency, electrophoresis, and peak detection, suggesting that our range of observation is from 15-85% similarity; for this reason, each sample was analyzed in duplicate and peaks were required to appear in both replicates to be included in further analysis. Bacterial communities were 40.2±0.2% (Average Bray-Curtis ±SEM) and 40.95±0.2% similar overall in the surface and DCM, respectively. The highest average Bray- Curtis similarity occurred between communities one month apart - 50.2% (0-5m) and 47.0% (DCM), consistent with the significant positive autocorrelation at one month by DFA. Annual recurrence was demonstrated by local maxima at yearly intervals (average 41.2%), while local minima for opposing seasons in 0-5m (6, 18, 30 months) indicated seasonal dissimilarity (average 37.9%). These trends were consistent for Bray-Curtis similarities calculated for the total observed community (Figure 2-3), the 50-most important OTUs, and OTUs that occurred in more than 50 months in 0-5m (Figure 2-S4). Bray-Curtis similarity in the DCM did not exhibit this sinusoidal trend and community similarity overall was often higher than at 0-5m across time lags; these results are consistent with the lack of yearly autocorrelation by DFA. DCM Bray-Curtis similarity also exhibited increased variance for communities ten years apart, which may be due to fewer observations or the El Niño Southern Oscillation event in 2010. Winter 2009-2010 was exceptionally warm, while winter 2010 – 2011 was exceptionally cool as compared to winter 2000-2001 (Oceanic Niño Index, NOAA). Despite these ENSO events, total community structure exhibited remarkable stability over ten years as evidenced by the average Bray-Curtis community similarity of 40% across all time lag comparisons. 40 This ten-year dataset confirms previously noted patterns on annual recurrence of bacterial communities in SPOT in the surface ocean (Fuhrman et al. 2006). Previous analysis at BATS and ALOHA identified similar seasonal differences in community structure within the euphotic zone (Treusch et al. 2009; Eiler et al. 2011) and annual recurrence was observed over six years in the Western English Channel (Gilbert et al. 2012). Seasonality and monthly variability of the bacterial community at 0-5m was consistent over ten years at SPOT by both DFA and Bray- Curtis similarities, resulting from changes in ARISA fingerprint patterns. DFA results from selected taxa in the DCM did reveal an underlying seasonal trend that was not seen by Bray- Curtis similarity but neither suggests an annually repeating trend. The surface ocean, thus, exhibited more pronounced and predictable seasonal patterns than the DCM, suggesting less variability in the DCM bacterial community structure. This distinction may be due to the relative isolation of the DCM from direct atmospheric forcing or our determination of the DCM based on biologically derived chlorophyll fluorescence. Coupling Community Structure and Environmental Changes We assessed coherence of the bacterial communities in the photic zone between depths by comparing contemporaneous communities, essentially asking to what extent the bacterial community in the surface ocean is similar to the DCM. From monthly averages of Bray-Curtis similarities, it appeared that the two depths were most similar between March-May and least similar in June, August and September (Figure 2-4). The between-depth Bray-Curtis similarity was not correlated to bacterial production, bacterial or viral abundance, or nutrient concentrations from either depth. However, we did observe a negative relationship with the between-depth temperature difference (Figure 2-4), This relationship suggests that during periods of winter mixing the two depths were relatively homogenized) and as the euphotic zone warms and 41 stratifies through the summer, bacterial communities diverged under different conditions, only to be mixed together again the following winter – a pattern that was also observed at BATS (Carlson et al. 2009; Treusch et al. 2009). Temperature is an established predictor of variability in community structure on spatial scales (Pommier et al. 2007; Fuhrman et al. 2008; Yilmaz et al. 2012), and our results suggests that it may be similarly predictive for spatiotemporal variation. Figure 2-4. Bray-Curtis similarity between 0-5m and DCM (y-axis) is negatively correlated with temperature difference (y-axis). Dotted lines indicate the 95% prediction levels. 42 Table 2-S2. Network Statistics for depth-specific networks and description of unique features within. ‘n.d.’= not determined. *Steele et al (2011). Depth-Based Networks 0-5m (10 years) DCM (4 years)* DCM (10 years) 0-5m + DCM Intersection Nodes 348 212 342 314 Edges 7277 1005 8806 1557 p-value cutoffs p<0.006, q<0.0482 p<0.01, q<0.067 p<0.0075, q<0.0478 Not applicable Diameter and Radius 3 (2) n.d. 3 (2) 4 (3) Average Clustering Coefficient (Cl) 0.369 0.27 0.43 0.264 Random Clustering Coefficient (Cl r ) 0.121 0.044 0.151 n.d. Ratio of Cl/Cl r 3.05 6.14 2.85 n.d. Characteristic Path length (L) 1.939 2.99 1.859 2.019 Random Characteristic Path length (L r ) 1.886 2.62 1.849 n.d. Log Response Ratio: Cl/Cl r 1.12 1.81 1.05 n.d. Log Response Ratio: L/L r 0.03 0.13 0.005 n.d. Between-Network Comparisons 0-5m DCM Unique Nodes 25 19 Description 1 Actinobacteria (OCS155); 2 Alphaproteobacteria (Defluviicoccus, Rhodobacteraceae): 3 Gammaproteobacteria (Vibrionaceae, Shewanella, SAR86); 2 Prochlorococcus; 17 Unidentified Day length; 4 Flavobacteria (Fluviivola, NS9, NS10); 1 Roseobacter; 1 Synechococcus Group A (III); Unidentified x 12 Unique Edges 5720 7249 Intersection Network Present in both 0-5m and DCM Shared nodes without shared edges 9 Description 2 Actinobacteria; 1 Prochloroccus HL (II)\; 1 SAR86; 5 Unidentified 43 Figure 2-5. Two highly interconnected sub-modules linked within a defined module in the surface ocean association network. Circles, ARISA OTUs; Squares, biotic; hexagon, abiotic. Node size reflects the commonness of an OTU – larger is more frequent. Solid lines, positive LS; dashed lines, negative LS; arrow, time-delayed LS correlations that point toward the lagging OTU. Figure 2-6. Connections that occur in both 0-5m and DCM networks between the five most abundant OTUs and their nearest neighbors. Size of nodes indicates relative frequency of OTU in 0-5m; all nodes represent bacteria. Only direction and delay is noted for LS correlations, as the values themselves differ between the two depths. Dashed lines, negative LS correlations; solid lines, positive LS correlations. Arrows indicate delayed correlation, with the arrow pointing towards the lagging OTU. 0.6407 0.7418 0.3027 0.4437 0.2661 0.5889 0.4399 0.7308 0.6106 ï 0.254 0.5569 0.4388 0.3022 0.441 0.3354 ï 0.2709 0.3216 0.8606 1.2662 1.2626 0.8618 0.5085 1.6768 0.8114 0.3585 1.4601 0.4547 0.9031 0.3051 0.3533 0.2551 0.3238 0.3293 0.3017 0.342 0.3281 0.3628 0.2315 0.2404 0.3041 0.25 ï 0.2373 0.2914 0.8473 0.4674 ï 0.1755 0.6636 0.3674 0.4922 0.4889 0.7566 0.2308 0.3101 0.6921 0.4262 0.3554 0.499 0.2867 0.2188 0.4594 0.3574 0.4237 0.4078 0.339 0.2574 0.9475 0.6346 1.2296 0.657 0.675 0.3702 0.9704 ï 0.21 0.2683 1.1288 1.4154 1.0764 0.7383 0.2209 0.684 0.42 0.7809 0.316 0.8417 ï 0.2719 1.2401 1.2799 0.7929 0.5527 0.4511 0.894 ï 0.2062 0.6407 0.246 0.481 0.4641 0.3408 0.6435 0.3135 0.307 0.4433 ï 0.2595 1.0517 1.0918 0.3778 0.4287 0.7623 0.5038 0.5823 0.4074 1.3499 0.6018 0.7687 0.4072 0.7317 1.253 0.2622 0.4488 0.9129 0.5507 0.2933 0.2698 0.2847 0.2529 0.2608 0.303 0.2167 0.2977 0.1971 0.2107 0.3123 0.2157 0.2681 0.3317 0.2598 ï 0.1652 ï 0.1637 ï 0.2447 ï 0.2196 ï 0.1805 ï 0.1839 ï 0.5885 ï 0.2488 ï 0.2097 0.2098 0.4766 0.7549 0.8002 0.3066 0.3387 0.5768 0.4206 0.4452 0.3222 0.4655 0.4475 0.5555 0.3353 0.536 0.8866 0.2492 0.3713 0.6616 0.4356 0.2954 0.2775 0.3187 0.4202 0.2345 0.348 0.2648 0.3191 0.3024 0.266 0.4742 0.2426 0.2912 0.3967 0.311 0.2904 0.4836 0.7326 0.3515 0.4456 0.1724 0.2857 0.3153 0.356 0.3912 0.6224 0.549 0.3764 0.2461 0.4759 0.458 0.2778 0.3543 0.1799 0.2777 0.2631 0.4409 0.3549 0.267 0.4472 0.2565 0.689 0.5146 0.3463 0.2908 0.2886 0.2319 0.3008 0.2646 0.6856 0.9875 0.3842 0.7042 0.6695 0.4041 0.7127 0.3563 0.4916 0.3223 0.2608 0.2034 0.3216 0.2435 0.2627 0.248 0.2337 0.2785 0.2682 0.3002 0.3045 0.4479 0.3471 0.501 0.5488 0.4231 0.44 0.316 0.6558 0.3767 0.5353 0.4002 0.2588 0.4407 0.2981 0.2752 0.2714 0.3897 0.2929 0.2228 0.2927 0.3573 0.4779 0.3095 0.3646 0.2884 0.2873 0.3349 0.3099 0.3146 0.4115 0.2532 0.3033 0.3268 0.2556 0.3171 0.2737 0.2819 0.3528 0.2542 0.1948 0.407 0.3462 0.2488 0.2368 0.2861 0.287 0.2252 0.3054 0.2728 0.2346 0.2007 0.2293 0.2547 0.2468 0.2584 0.3188 0.2242 0.216 0.2101 0.2441 0.3304 0.3664 0.2985 0.3075 0.2408 0.2575 0.2094 0.3959 0.3316 0.2757 0.3655 0.4441 0.2837 0.4893 0.2148 0.2928 0.3127 0.2251 0.2695 0.2293 0.2593 0.2097 0.2774 0.2326 0.2296 0.2755 0.2468 0.1889 0.2385 0.285 0.3855 0.3282 0.351 0.2926 OTU_863.9 OTU_1187.7 Owenwe_594.1 OTU_511 Sedimi/Punice_791.4 Mariba_800.9 Shewan_998.8 OTU_1145.6 OTU_808 OTU_874.7 NS2b_741.8 TurnLeu Flavo/Pro_874 Formos_770.5 OTU_545.8 SAR116_744.7 Leu Sphing_1011.5 Syn_B_1130.1 Pseudo/OM60_937.8 OTU_1191 Roseob_987.8 OTU_625.9 Shewan_602.5 SAR11_S2_721.2 SAR406_624.5 OTU_628.6 OCS116_852.1 SAR406_521.6 NS2b/SAR11_S2_709.4 Microb_634.7 OTU_547.8 SAR11_S1/3_674.2 NS4_729.4 NS9_683.9 NS9_732.2 NS2b_724.3 OTU_854.8 44 Ecological networks and the development of potential niches To further investigate these complex microbial communities and co-occurrence patterns at the OTU level, depth-specific association networks were constructed from local similarity (LS) correlations between all ARISA OTUs and environmental parameters. The resulting networks for both depths are complex and show a high degree of interconnectedness – more so than by random chance alone (Table 2-S3). As previously observed at SPOT and other locations, within-bacteria connections were more numerous than connections to environmental parameters (Steele et al. 2011; Eiler et al. 2012; Gilbert et al. 2012). Clustering coefficients were higher in both depth- based networks than observed in the prior 4-year DCM network (Steele et al. 2011) and random networks of equal size, giving more strength to the argument for small-world properties in microbial networks. The clustering coefficient ratio (Cl/Cl r ) is higher than before and still in alignment with previously observed ratios from food webs, pollinator-plant networks, and functional microbial networks, as previously summarized in Steele et al (2011) – suggesting that this longer dataset more fully captures the interactions within the bacterial community. The most interconnected cluster of nodes (or module) from the 0-5m network consisted of: 1) two highly-interconnected components that were linked primarily by time-delayed or negative correlations, 2) mostly positive LS correlations between bacterial nodes and 3) positive correlations to bacterial production measured by leucine (Leu) incorporation - and negative correlations to the calculated turnover rates (Figure 2-5). In this cluster, the OTUs in the left cluster appear to precede the right, following the direction of the time-delayed correlations (shown by arrows); Roseobacter_987.8, Shewan_998.8 and OTU_1191 also showed delayed correlations. Roseobacter sp. are commonly observed in areas or at times of high productivity (Buchan & Gonzalez 2005; Morris et al. 2012). Thus, this left cluster appears to include bacterial OTUs associated with heterotrophic activity and high productivity, which typically occurs in 45 March-April. The negative and delayed correlations suggest that each inter-connected component within the larger module may reflect separate bacterial niches within the dynamic surface ocean environment. The presence of multiple small modules supports the idea that the overall communities are stable, and each module may interact with the environment or other microbes in unique ways. Unique nodes and edges of each depth-based networks were determined by subtracting one from the other (Table 2-S2). Both value (positive or negative) and direction (undirected or directed) were considered when determining if an interaction was unique to a network. Because one caveat of any fingerprinting technique is that a peak could represent multiple organisms, it is possible that the same ARISA OTUs represented different taxa at different depths or times of the year. However, the consistency observed in community structure more likely suggests that each OTU (as best we can identify with ARISA) might interact differently with co-occurring microbes in its environments such that the niches are distinct between depths. In order to address coherence of microbial communities, we can define a “core” intersection network, based on LS correlations that are consistent to both sampling depths (again, by both value and direction). Only 10.7% of the original interactions were observed in both depths despite having retained most of the network nodes (Table S3). The five most abundant OTUs (overall) were present and still connected to many other bacteria (Figure 2-6). Within this network, most bacteria were negatively correlated to either SAR11_S1_666.4 or Actino_435.5 and these associated OTUs were not common (as indicated by the smaller node sizes). Few connections were positive, two of which connected SAR11_S1_666.4 and OTU_435.5 (Actino) directly to SAR11_S1_686.9. The other top two OTUs also clustered separately, perhaps implying that they occupy different niches. The interactions and nodes contained within this calculated ‘intersection’ network may reflect the “core microbiome” of the photic zone at SPOT. 46 Future comparisons in phylogeny and relative abundances of OTUs included in this network to other long-term time-series sites may aid definition of a truly global core microbiome that covers both time and space. Defining the microbial community by persistence and rarity The importance and role of the rare biosphere is of growing interest in microbial ecology, as new methods continue to push our limits of detection (Pedrós-Alió 2012). The bacterial community in the photic zone at SPOT was comprised of persistent, intermittent and ephemeral taxa. The persistent OTUs exhibited the highest average relative abundance within both depths (Figure 2-7), similar to the Western English Channel, Station ALOHA, and a freshwater lake (Gilbert et al. 2009; Caporaso et al. 2011; Eiler et al. 2011; 2012). In general, more OTUs were rare and infrequent than common: 60% (0-5m) and 58% (DCM) of OTUs were ephemeral (occurring in less than 25% of months observed), 33.4% (0-5m) and 35% (DCM) were intermittent (occurring in 25-75% of months observed) and the remaining 6-7% were persistent (occurring in >75% or months). Of the ephemeral and intermittent OTUs, their average abundance was no more than 1.2% as compared to >10% for the persistent OTUs. Of the OTUs in each depth with a maximum observed relative abundance >5%, most were intermittent or persistent. The taxonomic distribution between these three OTU types was consistent between depths, so the average contribution was shown for each class as named by the Greengenes taxonomy with explicit ranks (Figure 2-7D). We were least likely to identify the ephemeral OTUs than the persistent ones, with 60% of ephemeral OTUs as unknown and only 3.5% (average) unknown of the persistent OTUs (equates to 1 OTU in 0-5m and none in DCM). This pattern is expected because the clone libraries used for identification were more likely to include the 47 abundant organisms. Of the OTUs that we did identify (about half), the Alphaproteobacteria were a consistently large proportion. Persistent OTUs also included SAR406, Actinobacteria, Flavobacteria, Chloroplasts, Deltaproteobacteria, Gammaproteobacteria, and Synechococcophycideae. Betaproteobacteria, Oscillatoriophycideae, and Sphingobacteria were only observed in the intermittent OTUs. In the ephemeral OTUs, Verrucomicrobiae and Chlorobia were also observed in addition to the previous mentioned classes. All of these taxonomic groups are known key players of oceanic microbial communities over space and time (Morris et al. 2002; Treusch et al. 2009; Zinger et al. 2011; Gilbert et al. 2012; Morris et al. 2012; Yilmaz et al. 2012). 48 Figure 2-7. Defining the Microbiome of the euphotic zone at SPOT. The number of OTUs (A, y-axis) and average relative contribution of an OTU (B, y-axis) is shown relative to the OTU’s percent frequency (x-axis) at each depth. Percent frequency was determined by counting the numbers of months an OTU was observed and dividing by total number of sampling months (n = 103 (0-5m), 89 (DCM). White diamonds, 0-5m; black circles, DCM. (C) A generalized depiction of the number of OTUs observed within each category. Circles are drawn to scale according to their proportion of the community (average of 0-5m and DCM). (D) Taxonomic summary at Class level of all OTUs, within each percent frequency category (Persistent, Intermittent, Ephemeral). A) 0 5 10 15 20 25 0-5m 0 0.25 0.5 0.75 1 0 5 10 15 20 25 DCM B) DCM 0 0.25 0.50 0.75 1 0 0.03 0.06 0.09 0.12 0.15 0-5m 10% Persistent OTUs: 7% (0-5m), 6% (DCM) of all OTUs n = 28 (0-5m), 25 (DCM) Intermittent OTUs: 34% (0-5m), 35% (DCM) n = 136 (0-5m), n = 140 (DCM) Ephemeral OTUs: 60% (0-5m), 58% (DCM) of all OTUs n = 243 (0-5m), 231 (DCM) C) Unknown Verrucomicrobiae Synechococcophycideae Sphingobacteria Oscillatoriophycideae Gammaproteobacteria Flavobacteria Deltaproteobacteria Chloroplast Betaproteobacteria Alphaproteobacteria Actinobacteria SAR406 (AB16) Persistent Intermittent Ephemeral D) 49 Conclusions Underlying seasonal variation, a consistent core microbial community persists in both depths such that a minimum average ~40% similarity from month to month is maintained. Results from both Bray-Curtis Similarity and DFA suggest that the DCM comprises a more stable community year-round, but monthly variability is still apparent. We presume that it is the oscillation of the ephemeral or moderately abundant OTUs outside of a core community (both in its membership and the abundance of each) that was reflected in the seasonal variability. Although membership of the bacterial communities, as detectable with ARISA community fingerprints, was consistent within the surface and DCM, the interactions between their constituents within each depth differed. This observation may be due to the higher seasonal variation observed in 0-5m from changing environmental conditions and separation from the DCM by seasonal stratification of the water column, such that microbe-microbe interactions were mostly inconsistent between the two depths. For example, growth and activity of cyanobacteria has been shown to be dependent on light and nutrient conditions as well as the activity of phages and heterotrophic bacteria (Sher et al. 2011; Weinbauer et al. 2011). Further exploration of microbe-microbe interactions under varied environmental conditions and co-occurring communities would aid interpretation of ecological or association networks developed from fingerprinting and sequencing data. 50 Chapter Two References Acinas, S.G., Rodrı ́ guez Valera, F. & Pedrós-Alió, C. (1997). Spatial and temporal variation in marine bacterioplankton diversity as shown by RFLP fingerprinting of PCR amplified 16S rDNA. FEMS Microbiol Ecol, 24, 27–40. Alonso Sáez, L., Balagué, V., Sà, E.L., Sánchez, O., González, J.M., Pinhassi, J., et al. (2007). Seasonality in bacterial diversity in north‐west Mediterranean coastal waters: assessment through clone libraries, fingerprinting and FISH. FEMS Microbiol Ecol, 60, 98–112. Assenov, Y., Ramírez, F., Schelhorn, S., Lengauer, T. & Albrecht, M. (2008). Computing topological parameters of biological networks. Bioinformatics, 24, 282–284. Beman, J.M., Sachdeva, R. & Fuhrman, J.A. (2010). Population ecology of nitrifying Archaea and Bacteria in the Southern California Bight. Environ Microbiol, 12, 1282–1292. Beman, J.M., Steele, J.A. & Fuhrman, J.A. (2011). Co-occurrence patterns for abundant marine archaeal and bacterial lineages in the deep chlorophyll maximum of coastal California. ISME J, 5, 1077–1085. Bouskill, N.J., Eveillard, D., O'Mullan, G., Jackson, G.A. & Ward, B.B. (2010). Seasonal and annual reoccurrence in betaproteobacterial ammonia-oxidizing bacterial population structure. Environ Microbiol, 13, 872–886. Brown, M.V. & Fuhrman, J.A. (2005). Marine bacterial microdiversity as revealed by internal transcribed spacer analysis. Aquat. Microb. Ecol., 41, 15–23. Brown, M.V., Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2005). Coupling 16S-ITS rDNA clone libraries and automated ribosomal intergenic spacer analysis to show marine microbial diversity: development and application to a time series. Environ Microbiol, 7, 1466–1479. Buchan, A. & Gonzalez, J. (2005). Overview of the Marine Roseobacter Lineage. Appl. Environ. Microbiol. Cai, H., Wang, K., Huang, S., Jiao, N. & Chen, F. (2010). Distinct Patterns of Picocyanobacterial Communities in Winter and Summer in the Chesapeake Bay. Appl. Environ. Microbiol., 76, 2955–2960. Campbell, B.J., Yu, L., Heidelberg, J.F. & Kirchman, D.L. (2011). Activity of abundant and rare bacteria in a coastal ocean. Proc. Natl. Acad. Sci. U.S.A., 108, 12776. Caporaso, J.G., Paszkiewicz, K., Field, D., Knight, R. & Gilbert, J.A. (2011). The Western English Channel contains a persistent microbial seed bank. ISME J. Carlson, C.A., Morris, R., Parsons, R., Treusch, A.H., Giovannoni, S.J. & Vergin, K.L. (2009). Seasonal dynamics of SAR11 populations in the euphotic and mesopelagic zones of the northwestern Sargasso Sea. ISME J, 3, 283–295. Chow, C.-E.T. & Fuhrman, J.A. (2012). Seasonality and monthly dynamics of marine myovirus communities. Environ Microbiol. In press. 51 Clarke, K.R. (1993). Non‐parametric multivariate analyses of changes in community structure. Aust J Ecol, 18, 117–143. Clarke, K.R. & Gorley, R. (2006). PRIMER v6: User Manual/Tutorial. PRIMER-E, Plymouth, UK. Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., et al. (2009). The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res, 37, D141–D145. Collins, L.E., Berelson, W., Hammond, D.E., Knapp, A., Schwartz, R. & Capone, D. (2011). Particle fluxes in San Pedro Basin, California A four-year record of sedimentation and physical forcing. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 58, 898–914. Countway, P.D. & Caron, D.A. (2006). Abundance and distribution of Ostreococcus sp. in the San Pedro Channel, California, as revealed by quantitative PCR. Appl. Environ. Microbiol., 72, 2496–2506. Countway, P.D., Vigil, P.D., Schnetzer, A., Moorthi, S.D. & Caron, D.A. (2010). Seasonal analysis of protistan community structure and diversity at the USC Microbial Observatory (San Pedro Channel, North Pacific Ocean). Limnol Oceanogr., 55, 2381–2396. Ducklow, H.W., Doney, S.C. & Steinberg, D.K. (2009). Contributions of Long-Term Research and Time-Series Observations to Marine Ecology and Biogeochemistry. Annu. Rev. Marine. Sci., 1, 279–302. Eiler, A., Hayakawa, D.H. & Rappé, M.S. (2011). Non-random assembly of bacterioplankton communities in the subtropical north pacific ocean. Front. Microbio., 2, 1–10. Eiler, A., Hayakawa, D.H., Church, M.J., Karl, D.M. & Rappé, M.S. (2009). Dynamics of the SAR11 bacterioplankton lineage in relation to environmental conditions in the oligotrophic North Pacific subtropical gyre. Environ Microbiol, 11, 2291–2300. Eiler, A., Heinrich, F. & Bertilsson, S. (2012). Coherent dynamics and association networks among lake bacterioplankton taxa. ISME J, 6, 330–342. Fisher, M.M. & Triplett, E.W. (1999). Automated approach for ribosomal intergenic spacer analysis of microbial diversity and its application to freshwater bacterial communities. Appl. Environ. Microbiol., 65, 4630–4636. Fuhrman, J.A. (2009). Microbial community structure and its functional implications. Nature, 459, 193–199. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Fuhrman, J.A., Steele, J.A., Hewson, I., Schwalbach, M.S., Brown, M.V., Green, J.L., et al. (2008). A latitudinal diversity gradient in planktonic marine bacteria. Proc. Natl. Acad. Sci. U.S.A., 105, 7774–7778. 52 Gilbert, J.A., Field, D., Swift, P., Newbold, L., Oliver, A., Smyth, T., et al. (2009). The seasonal structure of microbial communities in the Western English Channel. Environ Microbiol, 11, 3132–3139. Gilbert, J.A., Steele, J.A., Caporaso, J.G., Steinbrück, L., Reeder, J., Temperton, B., et al. (2012). Defining seasonal marine microbial community dynamics. ISME J, 6, 298–308. Giovannoni, S.J. & Vergin, K.L. (2012). Seasonality in Ocean Microbial Communities. Science, 335, 671–676. Kan, J., Suzuki, M.T., Wang, K., Evans, S.E. & Chen, F. (2007). High Temporal but Low Spatial Heterogeneity of Bacterioplankton in the Chesapeake Bay. Appl. Environ. Microbiol., 73, 6776–6789. Kottmann, R., Kostadinov, I., Duhaime, M.B., Buttigieg, P.L., Yilmaz, P., Hankeln, W., et al. (2009). Megx.net: integrated database resource for marine ecological genomics. Nucleic Acids Res, 38, D391–D395. Li, W. (1998). Annual average abundance of heterotrophic bacteria and Synechococcus in surface ocean waters. Limnol Oceanogr., 1746–1753. Malmstrom, R.R., Coe, A., Kettler, G.C., Martiny, A.C., Frias-Lopez, J., Zinser, E.R., et al. (2010). Temporal dynamics of Prochlorococcus ecotypes in the Atlantic and Pacific oceans. ISME J, 4, 1252–1264. McDonald, D., Price, M.N., Goodrich, J., Nawrocki, E.P., DeSantis, T.Z., Probst, A., et al. (2011). An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J, 6, 610–618. Morris, R.M., Frazar, C.D. & Carlson, C.A. (2012). Basin-scale patterns in the abundance of SAR11 subclades, marine Actinobacteria (OM1), members of the Roseobacter clade and OCS116 in the South Atlantic. Environ Microbiol, In press. Morris, R.M., Rappé, M.S., Connon, S.A., Vergin, K.L., Siebold, W.A., Carlson, C.A., et al. (2002). SAR 11 clade dominates ocean surface bacterioplankton communities. Nature, 420, 806–810. Morris, R.M., Vergin, K.L., Cho, J.C., Rappé, M.S., Carlson, C.A. & Giovannoni, S.J. (2005). Temporal and spatial response of bacterioplankton lineages to annual convective overturn at the Bermuda Atlantic Time-series Study site. Limnol Oceanogr., 1687–1696. Paerl, R., Turk, K., Beinart, R., Chavez, F.P. & Zehr, J. (2012). Seasonal change in the abundance of Synechococcus and multiple distinct phylotypes in Monterey Bay determined by rbcL and narB quantitative PCR. Environ Microbiol, 14, 580–593. Parsons, R.J., Breitbart, M., Lomas, M.W. & Carlson, C.A. (2012). Ocean time-series reveals recurring seasonal patterns of virioplankton dynamics in the northwestern Sargasso Sea. ISME J, 6, 273–284. 53 Patel, A., Noble, R.T., Steele, J.A., Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2007). Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR Green I. Nat Protoc, 2, 269–276. Pedrós-Alió, C. (2012). The Rare Bacterial Biosphere. Annu. Rev. Marine. Sci., 4, 449–466. Pommier, T., Canbäck, B., Riemann, L., Boström, K.H., Simu, K., Lundberg, P., et al. (2007). Global patterns of diversity and community structure in marine bacterioplankton. Mol. Ecol., 16, 867–880. Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J., et al. (2007). SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res, 35, 7188–7196. Robidart, J.C., Preston, C.M., Paerl, R.W., Turk, K.A., Mosier, A.C., Francis, C.A., et al. (2012). Seasonal Synechococcus and Thaumarchaeal population dynamics examined with high resolution with remote in situ instrumentation. ISME J, 6, 513–523. Rocap, G., Distel, D.L., Waterbury, J.B. & Chisholm, S.W. (2002). Resolution of Prochlorococcus and Synechococcus ecotypes by using 16S-23S ribosomal DNA internal transcribed spacer sequences. Appl. Environ. Microbiol., 68, 1180–1191. Ruan, Q., Dutta, D., Schwalbach, M.S., Steele, J.A., Fuhrman, J.A. & Sun, F. (2006a). Local similarity analysis reveals unique associations among marine bacterioplankton species and environmental factors. Bioinformatics, 22, 2532–2538. Ruan, Q., Steele, J.A., Schwalbach, M.S., Fuhrman, J.A. & Sun, F. (2006b). A dynamic programming algorithm for binning microbial community profiles. Bioinformatics, 22, 1508– 1514. Shade, A. & Handelsman, J. (2012). Beyond the Venn diagram: the hunt for a core microbiome. Environ Microbiol, 14, 4–12. Sher, D., Thompson, J.W., Kashtan, N., Croal, L. & Chisholm, S.W. (2011). Response of Prochlorococcus ecotypes to co-culture with diverse marine bacteria. ISME J, 5, 1125–1132. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. Storey, J.D. (2002). A direct approach to false discovery rates. J Roy Stat Soc B Met, 64, 479– 498. Tai, V. & Palenik, B. (2009). Temporal variation of Synechococcus clades at a coastal Pacific Ocean monitoring site. ISME J, 3, 903–915. Treusch, A.H., Vergin, K.L., Finlay, L.A., Donatz, M.G., Burton, R.M., Carlson, C.A., et al. (2009). Seasonality and vertical structure of microbial communities in an ocean gyre. ISME J, 3, 1148–1163. 54 Weinbauer, M.G., Bonilla-Findji, O., Chan, A.M., Dolan, J.R., Short, S.M., Simek, K., et al. (2011). Synechococcus growth in the ocean may depend on the lysis of heterotrophic bacteria. J. Plankton Res., 33, 1465–1476. Xia, L.C., Steele, J.A., Cram, J.A., Cardon, Z.G., Simmons, S.L., Vallino, J.J., et al. (2011). Extended local similarity analysis (eLSA) of microbial community and other time series data with replicates. BMC Syst. Biol., 5, S15. Yilmaz, P., Iversen, M.H., Hankeln, W., Kottmann, R., Quast, C. & Glöckner, F.O. (2012). Ecological structuring of bacterial and archaeal taxa in surface ocean waters. FEMS Microbiol Ecol, n/a–n/a. Yoon, J.S. & Jung, W.H. (2011) "A GPU-accelerated bioinformatics application for large-scale protein interaction networks", APBC poster presentation. Zinger, L., Amaral-Zettler, L.A., Fuhrman, J.A., Horner-Devine, M.C., Huse, S.M., Welch, D.B.M., et al. (2011). Global Patterns of Bacterial Beta-Diversity in Seafloor and Seawater Ecosystems. PLoS ONE, 6, e24570. 55 Chapter Three: Seasonality and Monthly Dynamics of Marine Myovirus Communities Abstract Marine myoviruses (i.e. bacteriophages with a contractile tail sheath) are numerically abundant and genetically diverse. We developed a terminal restriction fragment length polymorphism assay (TRFLP) for g23, the conserved gene encoding the major capsid protein, to investigate T4-like myovirus communities at USC’s Microbial Observatory at the San Pedro Ocean Time-series (SPOT), where we previously reported bacterial seasonality. Between 71-154 Operational Taxonomic Units (OTUs) were observed monthly over three years. Roughly 25% of OTUs were detected in 31 or more months. T4-like myoviral community structure varied seasonally with some OTUs peaking repeatedly in spring-summer and others in fall-winter, while moderately abundant OTUs persisted year-round. Recurring community structure was demonstrated using discriminant function analysis (DFA, selecting taxa that best predict months) and average Bray-Curtis similarity. DFA showed communities from adjacent months or 12 months apart were positively auto-correlated, while communities 3-7 months apart were negatively auto-correlated, Bray-Curtis similarity was highest between adjacent months - with a local maximum at 12 month and local minima at 6 and 18-20 month lags. The T4-like virus community at SPOT exhibited seasonality, yet the somewhat unexpected persistence of moderately abundant OTUs and predictability of the community add new twists to existing conceptual models of marine viruses. 56 Introduction Changes in viral community structure may strongly influence shifts in the bacterial and eukaryotic communities through the interaction of the viruses with their potential hosts (Fuhrman 1999; Suttle 2007; Sandaa 2008; Rohwer & Thurber 2009; Breitbart 2012). Variability in viral community structure is thus a result of the net production and loss of viral populations, which is directly influenced by host abundance and susceptibility (Fuhrman 1999). Ecological investigations of viruses and their hosts have sought to quantify and describe these mechanisms of viral production and loss, and how viral infection and lysis shapes the co-existing host communities (Weinbauer & Rassoulzadegan 2003; Sandaa et al. 2009; Holmfeldt et al. 2010). Two complementary models to describe viral population dynamics are the Bank Model and “Kill the Winner.” The Bank Model suggests that individual viral types are typically rare, like a seed population, and proliferate only when hosts are available (Breitbart & Rohwer 2005). Abundant viral genotypes are estimated to be, at most, 5% of the community with most genotypes present at <0.01% of the community from metagenomics. Studies of prasinoviruses from Lake Ontario and Ostreococcus tauri viruses also demonstrated results consistent with this model (Bellec et al. 2010; Short et al. 2011). “Kill the Winner” focuses on describing host-virus interactions, where successive waves of competitively successfully bacteria are decimated by viral infection and then replaced by new winners, fostering a dynamic and diverse system (Fuhrman & Suttle 1993; Thingstad & Lignell 1997; Thingstad 2000; Breitbart 2012). For example, recurring patterns in viral abundance were observed at the Bermuda Atlantic Time- series Study (BATS) and showed strong correlations between viral abundance and SAR11, Prochlorococcus, and Rhodobacteraceae (Parsons et al. 2012). Temporal variation in viral genotypes following “kill the winner” could potentially occur at the strain level, while overall community composition remains stable at higher taxonomic levels (Rodriguez-Brito et al. 2010). 57 Because viruses lack a “universal” gene like rRNA used to study cellular life, our understanding of spatial and temporal variation within the marine virus community stems from research on isolated strains, morphology or coarse-resolution observation (e.g. total genome length variation) by pulsed-field gel electrophoresis (Børsheim 1993; Waterbury & Valois 1993; Wommack et al. 1999; Steward et al. 2000; Jiang et al. 2003; Larsen et al. 2004; Clokie et al. 2006; Sandaa & Larsen 2006; Sandaa et al. 2008). Many, but not all, studies have observed spatial and temporal variability of viral communities through analysis of selected viral genes (major capsid protein (g23), portal protein (g20), DNA polymerase pol, and DNA polymerase 1b) by sequence analysis, denaturing gradient gel electrophoresis or quantitative PCR; these studies together developed marine viral community fingerprinting as a relevant molecular tool and provide a framework for understanding viral dynamics in concert with environmental changes (Short & Suttle 2002; Zhong et al. 2002; Frederickson et al. 2003; Mühling et al. 2005; Sandaa & Larsen 2006; Labonte et al. 2009; Short & Short 2009; Jameson et al. 2011; Short et al. 2011) but very few of the referenced studies assessed individual viruses repeatedly over multiple years. Previous surveys in both marine and non-marine environment support the significance of T4-like myoviruses in the environment and the use of g23 as a marker for them (Filee 2005; Comeau & Krisch 2008; Fujii et al. 2008; Lopez-Bueno et al. 2009; Butina et al. 2010). Although genomic analysis has discovered other conserved genes within this phage family, g23 remains one of the key core single-copy genes (Millard et al. 2009; Sullivan et al. 2010). Tailed dsDNA phages, of which the T4-like phages are a subset, have been observed as a significant fraction of the viral community by both transmission electron microscopy (TEM) and metagenomic analysis (Wommack et al. 1992; Børsheim 1993; Waterbury & Valois 1993; Breitbart 2002; Breitbart et al. 2004; Angly et al. 2006; Bench et al. 2007; Williamson et al. 2008). The phage gene, g23, was frequently observed in the Global Ocean Survey (GOS) bacterial size-fraction metagenomes, 58 presumably from infected hosts rather than free viruses(Yooseph et al. 2007). Analysis of the GOS g23 sequences and known isolates determined that sequence variation within this gene broadly correlates with host phylogeny such that all known T4-like myoviruses infecting cyanobacteria clustered into a group, as did those infecting enteric bacteria (Comeau & Krisch 2008) – similar to observations of T4-like viruses using g20 (Short & Suttle 2005). We focused on g23 in this study and benefited from the availability of a new broadly-targeted primer set designed from an alignment of ~1500 environmental amino acid sequences (Comeau & Krisch 2008) to investigate viruses for whom hosts are largely unknown outside of cyanobacteria. Prior biodiversity research at the USC Microbial Observatory at the San Pedro Ocean Time-Series (SPOT) has focused on archaeal, bacterial and protistan communities. Elements of bacterial niches have been indicated for key bacteria in the photic zone based on co-occurrence of bacterial OTUs and environmental parameters, leading to a better understanding of microbial natural history and bacteria-bacteria or bacteria-protist associations (Fuhrman et al. 2006; Fuhrman & Steele 2008; Steele et al. 2011). Viruses in the Southern California Bight were shown to significantly affect bacterial community composition (Fuhrman & Schwalbach 2003; Schwalbach et al. 2004; Hewson & Fuhrman 2006; 2007), however changes within the viral community structure were not investigated. The goals of this study were to address key ecological questions: (1) how similar is the T4-like viral community from month to month? (2) to what extent are T4-like myovirus OTUs persistent? and (3) does the community exhibit repeating seasonal patterns, as has been observed in bacteria at the same location? In order to detect detailed community composition variation, a novel terminal restriction fragment length polymorphism assay (TRFLP) was developed for g23, a core gene marker, to focus specifically on the ecology of T4-like myoviruses. 59 Experimental Procedures Study Site and Sample Collection The USC Microbial Observatory at SPOT is located midway between Los Angeles and Santa Catalina Island (33˚33’N, 118˚24’W). 10L of seawater was collected monthly from March 2008 to January 2011 by rosette at 5m (or by bucket at 0m, when 5m samples were unobtainable, for December 2008, October 2009, June 2010 and September 2010) with no sample collected January 2009. Samples were collected routinely between 10:00am and 12:00p and processed the same day. The mixed layer depth is consistently below 5m at SPOT based on temperature and salinity profiles, so we expect these samples to be indistinguishable (data not shown). Both bacterial and viral abundances were determined by SYBR Green I staining and epifluorescence microscopy; 2ml of formalin-preserved seawater samples (50ml starting) were filtered onto a 25mm 0.02µm Anodisc filter, stained, and enumerated as described (Noble & Fuhrman 1998; Patel et al. 2007). Molecular Analysis DNA Collection and Extraction. Seawater samples were serially filtered by positive pressure at 10-15 psi through a 142mm glass-fiber filter (Type A/E; Pall) and 142mm 0.22µm Durapore GVWP (Millipore). The filtrate was collected and the viral fraction, from 30kDa through 0.2µm, was concentrated by tangential flow filtration (30kDa, Millipore spiral cartridge) to a volume of approximately 100ml. Viruses in the concentrates were precipitated by incubating with 10% w/v of polyethylene glycol (PEG) for 12 hours at 4ºC before pelleting (30-60 minutes at 13,000xg), resuspended in 500µl of TE buffer and archived at -80ºC (Colombet et al. 2007). Viral DNA was extracted using a CTAB-based method (Sambrook et al. 1989). Additional 0.5L – 1L seawater samples were collected beginning in June 2009 to provide a second source of viral DNA based on 60 newer protocols (Steward & Culley 2010); these samples were serially filtered through a 0.22 µm Sterivex (Millipore, with low binding Durapore filter) to remove bacteria and then onto a 25mm 0.02µm Anotop (Whatman) filter using a peristaltic pump. We applied gentle pressure with a sterile 60ml plastic syringe to clear the filters of any remaining seawater, and filters were stored dry, wrapped in parafilm, at -80ºC. DNA (presumably viral) was extracted from the Anotop filter using Epicentre Total DNA kit, with modified protocol as described in Steward and Culley (2010). DNA extracts were quantified using QuantIt PicoGreen (Invitrogen) with the quantitative plate-read mode on a Stratagene MX3000P real-time PCR machine, according to manufacturer’s instructions. PCR. g23 PCR primers were redesigned by André M. Comeau from MZIA1bis and MZIA6 (Filee 2005) to allow for more universal amplification of myoviruses based upon an alignment of over 1000 cultured and metagenomic g23 sequences originally presented in Comeau and Krisch (2008). Primer sequences are T4superF1, 5’-[TET] GAY HTI KSI GGI GTI CAR CCI ATG-3’ and T4superR1, 5’-[6FAM] GC IYK IAR RTC YTG IGC IAR YTC-3’, with degeneracies of 48 and 128, respectively. T4superF1 is located in the same site as the previous MZIA1bis, but is more degenerate; T4superR1 is located internally to the previous reverse primer, MZIA6. The resulting amplicon is approximately 100bp shorter than the MZIA1bis-MZIA6 amplicon and ranges in our experience from 400-500bp. Primers were successfully tested against 13 myovirus isolates, including cyanophages, T4, T6, and nt1 (A.M. Comeau, personal communication). Each PCR reaction (run in quadruplicate) for environmental samples contained 2ng DNA, 1x ThermoPol Buffer (includes 20mM MgSO 4 , NEB), additional 1.75mM MgCl 2 (final concentration is 3.75mM), 0.25mM each dNTP, 400nM each primer, 10µg BSA (Sigma #7030) and 5 units Taq DNA polymerase (NEB) with a final volume of 50µl. Cycling parameters were: 95ºC for 5 minutes, 35 cycles of 95ºC for 30 seconds, 54ºC for 30 seconds, and 72ºC for 30 61 seconds with a final extension at 72ºC for 9.5 minutes. 5µl of each PCR reaction was electrophoresed in 1% agarose in 1x TBE, pH 8.0 at 100V for 60 minutes to verify amplification. Cloning and Sequence analysis. Clone libraries were constructed for five months directly from PCR amplicons using TOPO TA for Sequencing kit (Invitrogen) according to the manufacturer’s instructions. 96 clones from each month were picked and grown overnight at 37ºC prior to Sanger sequencing (Genbank accession JN791695-JN792132). For each library, 85-91 quality sequences were obtained: May 2008 (n=86), August 2008 (n=91), November 2008 (n=85), December 2008 (n=89), and February 2009 (n=87). All sequences were translated and aligned by MUSCLE using the parameters as described in Comeau and Krisch (2008). SeqOTUs were defined using cluster.seqs in mothur v1.13 based on average-neighbor distance (Schloss et al. 2009). Each seqOTU was manually assigned to the Venn diagram and OTU table (Figures 3-1 and 3-S1). In-silico analysis. Predicted fragment lengths for the SPOT clone libraries and available myovirus genomes were calculated in silico using REPK (Collins & Rocap 2007). Two blunt- cutting enzymes (HincII and RsaI) were selected to maximize unique fragments and minimize generation of overlapping fragment lengths where possible. An OTU table based on fragment lengths and occurrence of each fragment length was constructed manually and each OTU was then assigned to the Venn diagram. Digestion. For environmental samples, quadruplicate PCR reactions were pooled and then split directly into two single equal volumes for reactions of ~100µl each for HincII and RsaI (NEB). Digestion reactions: 1X Buffer, 10units of restriction enzyme and 1X BSA (for HincII only) and digested for 8 hours at 37ºC followed by heat inactivation at 65ºC for 20 minutes. Digestion of amplified PCR products from selected clones from our libraries and an environmental sample were tested at 1, 2, and 8 hours. Complete digestion of the clones was only observed after 8 62 hours, although no difference in the TRFLP fingerprint from the environmental sample was observed visually between 2- and 8-hours (data not shown). Restriction digests were concentrated with Zymo Clean and Concentrator™ -5 to a final elution volume of 8µl. Concentrated samples were quantified by Picogreen and diluted to 20ng/µl before fragment analysis. Fragment Analysis. Restriction fragments were detected by automated slab gel electrophoresis in duplicates on non-adjacent lanes on an ABI 377 using 20ng/lane and a customized internal size standard (Bioventures) with standards every 25bp (50-900bp) or every 50bp (900-1400bp). Differences in G/C composition and fragment length results in fragments with apparent non- integer lengths, which allowed for discrimination of fragment sizes <1bp different. Peak data were analyzed using DAx (van Mierlo Inc., The Netherlands); peak finding was done using a peak half width of 3 and automatic slope peak detection with a threshold of 0. Non-topping peaks were detected using second-derivative analysis with a slant value of 1.3. The local southern method was used for standard calibration. Fragments <75bp length (too small to accurately size) and <25AU fluorescence (too small to accurately measure) were removed from further analysis; fragments lengths were rounded to the nearest 0.1bp before dynamic binning of peak data (Ruan et al. 2006), with a maximum bin size of 1bp. Dynamic bins were manually merged for any resulting bins that were <0.5bp wide, which resulted in a few final bins that were greater than 1bp. We analyzed all fragments between 75-500bp, including uncut fragments (>350bp) and fragments between ~75-350bp included both cut and uncut fragments based on our in silico analysis. Each distinct binned fragment length was considered a unique OTU. OTU contributions (or relative abundance) were determined by calculating peak area of each OTU and dividing by total peak area in each fingerprint, which yielded a normalized peak area (i.e. percentage of the community) to compare OTUs between samples. 63 Statistical analysis Each g23-TRFLP assay was analyzed independently and the patterns presented were consistent across all four assays unless otherwise noted. Bray-Curtis Similarity. Bray-Curtis Similarity was calculated for each g23-TRFLP assay independently using all standardized peak data, so that each TRFLP fingerprint represents 100% of the community, using PRIMER v6 (Clarke 1993; Clarke & Gorley 2006), Each sampling date was assigned a value for elapsed month, from 1 (March 2008) to 35 (January 2011). Time lags in months were determined by difference in elapsed months for each pairwise Bray-Curtis comparison (i.e. lag = elapsed month 2 – elapsed Month 1). Box-and-whisker plots were calculated in Sigmaplot 11 (Systat Software, Inc). Boxes were not calculated for time lags with less than three data points, nor were the 10 th and 90 th percentiles for time lags with fewer than nine data points. Discriminant Function Analysis. The “n-most important” OTUs were selected using PRIMER v6, where n = 25, for each TRFLP assay independently. Untransformed relative abundance data were used to determine the discriminant functions using Systat 11 (Systat Software, Inc.). The first discriminant function (score 1) was plotted over time (time-series plot) or auto-correlated according to consecutive order of samples (as cases), which was not adjusted for any missing months. Results Community Composition by g23 Sequence Discrimination We prepared five clone libraries to investigate myoviral community structure from March 2008 to February 2009 (Figures 3-1 and 3-S1) and compared the number and distribution of operational taxonomic units based on sequence identity (seqOTUs; Figure 3-1A), in silico 64 digestion with HincII or RsaI (in silico TRFLP OTUs, Figure 3-1B), and assessment of environmental samples by our g23-TRFLP assays (environmental TRFLP OTUs, Figure 3-1C). Use of labels at both 5’ and 3’ ends of the amplicon combined with two restriction enzymes resulted in four possible g23-TRFLP profiles for in silico and environmental DNA analyses. Values shown are the percent of OTUs shared between the indicated months: # OTUs shared/total # OTUs observed in library or fingerprint (Figures 3-1 and 3-S1). From our in silico results, TRFLP fragment lengths greater than 275bp potentially include undigested PCR products due to lack of the restriction sites. We completed our sequence comparisons in the context of known sequence diversity by analyzing 26 T4-like viral genomes (available in Genbank), which showed that the highest observed g23 similarity was 97% (T4 versus KC69 and TuIa, all of which are enteric viruses) at the amino acid level, followed by 95% similarity between coliphages T4 and T6. Cyanophage sequences, from the available genomes, were polymorphic and varied from 36- 89% similarity over the same region. Sequences from two cyanophages co-isolated from the same host and source water, P-HM1 and P-HM2, were 89% similar at the amino acid level in contrast to whole genome analysis which showed these two phages were 83% similar overall at the amino acid level for all shared proteins (Sullivan et al. 2010). We compared the distribution of seqOTUs clustered at three similarity levels based on the above results: 90%, 95%, and 97%. Less than 2% of all seqOTUs were shared between all five months yet 5-11.6% of in silico OTUs and 31.6% of environmental TRFLP OTUs were observed in all five months (Figure 3-1, Figure 3-S1). Seventy percent of all seqOTUs were observed only once in contrast to 19.3-26% of in silico TRFLP OTUs. Note that even at 90% amino acid similarity there are far fewer shared seqOTUs than in silico or environmental TRFLP OTUs. In order to make comparisons between TRFLP and clone library coverage more fair, we reduced the TRFLP dataset to approximate the coverage of our clone libraries. We removed 103 TRFLP OTUs with <0.5% relative abundance, 65 as these are most likely representative of rare OTUs unlikely to be observed in our clone libraries and observed a reduction in OTUs shared between all five months and an increase in OTUs unique to May 2008. Figure 3-1. Venn diagram of percent of total OTUs shared over five months. (A) sequence-derived OTUs at 95% amino acid similarity from clone libraries, (B) in silico digestion of clone libraries (i.e. simulated TRFLP) or (C) observed environmental TRFLP fingerprints. Values indicate 100 x (# shared OTUs / # total OTUs in all five months). 66 Figure 3-S1. Distribution of OTUS from a) sequencing clone libraries, b) in silico digestion of clone libraries, c) environmental TRFLP or d) “reduced” environmental TRFLP. Y-axis denotes the months where OTUs re-occurred (M=May 2008, A=August 2008, N=November 2008, D=December 2008, F=February 2009). Each sequence similarity or g23-TRFLP is color-coded: black = 97% similarity or 5’- HincII; medium grey = 95% similarity or 3’-HincII, dark grey = 90% similarity or 5’-RsaI; light gray = 3’- RsaI. 67 Figure 3-S2. Environmental variability at SPOT 0-5m from March 2008 to January 2011: a) salinity, b) temperature, c) bacterial (left axis) and viral (right-axis) abundance. 68 Figure 3-S3. Schematic of 3’-HincII TRFLP fingerprints illustrates changes in relative abundances of unique fragment length over time. Each column shows the profile from one month, in sequential order. Each bar represents an OTU with a resolution of 0.5-1.5bp, depending on fragment length; undigested fragments are >350bp. Numbers at top of figure indicate: year, month, DNA source (PEG concentrate or Anotop), DNA concentration (to 20ng/µl whenever possible), total number of OTUs observed each month. 69 Environmental Parameters Sea surface temperature ranged from 12.7-21.4ºC, with seasonal lows in February and March and highs in August and September, although 2010 peaked at only 18ªC (Figure 3-S2). Salinity changed relatively little, with slight increases during the spring and summer months (range: 33.1-33.6 psu, Figure 3-S2). Viral and bacterial abundances fluctuated within an order of magnitude and showed little seasonality aside from a seasonal low each October to November (bacteria: 3.7x10 5 – 7.5x10 6 cells/ml; viruses: 5.3x10 6 – 6.7x10 7 particles/ml, Figure 3-S2). Monthly and Seasonal Community and Individual OTU Dynamics Monthly viral community composition at SPOT was investigated by independently evaluating the 5’ or 3’ terminal fragments after digestion with RsaI or HincII (e.g. Figure 3-S3). In silico analysis of the clone libraries facilitated TRFLP enzyme selection. The resulting fragment lengths, which included both digested and undigested fragments, were consistent within a seqOTU and multiple seqOTUs were occasionally observed within a single fragment length (data not shown). We selected restriction enzymes that minimized this overlap as much as possible. Each enzyme-fragment combinations resulted in 180-225 total unique OTUs, with 71- 130 OTUs observed per month (Table 3-1). An OTU’s relative abundance (or contribution) was determined by dividing individual peak area by total peak area after removing peaks that did not meet minimum height (20AU) or area (>0.001) requirements. The average contribution of any OTU to the T4-like viral community ranged from 0.44-0.56±0.02% of the community, while the maximum contribution was 17.3-36.4%, depending on the enzyme and terminal fragment used. Over all four assays, less than 1.3% of OTUs were observed in only one month while >50% of OTUs were detected in more than 15 months. 70 Table 3-1. Overview of g23-TRFLP OTUs. Summary of TRFLP OTUS observed over all months based on four unique combinations of restriction enzyme and whether the forward or reverse fragments were fluorescently labeled. SEM is standard error of the mean. Figure 3-2. Virus OTUs detected more frequently have higher average and maximum contributions to the T4-like community. Each symbol represents an individual 3’-HincII TRFLP OTU, where symbol indicates the OTU’s occurrence. X-axis is the average proportion of an OTU peak, expressed as a percentage of the total amplified products (a measure of the OTU’s representation in the myoviral community). Y-axis is the maximum observed proportion of the OTU, and is also expressed as a percentage of the myoviral community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igure 3-3. Seasonality by Discriminant Function and Bray-Curtis Similarity Analyses. (A) Time- series plot shows the first discriminant function based on the 25 most important 3’-HincII TRFLP OTUs. X-axis shows time lag in months, but does not account for the few missing months. Raw score is dashed line; solid line is the 3-month moving average. (B) Auto-correlation of first discriminant function (solid curved line indicates p=0.05). (C-F) Box-and-whisker plots from all Bray-Curtis similarities. Lower and upper bounds of the boxes are 25th and 75th percentiles, respectively; internal line is the median, and error bars are the 10th and 90th percentiles. Circles show all outliers. (G) Average Bray-Curtis similarity for each TRFLP assay. 72 Figure 3-S4. Discriminant function and time series analyses demonstrate re-occurring viral communities from all four g23-TRFLP assays: a) 3'-HincII, b) 5'-HincII, c) 3'-RsaI, and d) 5'-RsaI. On the left column, the first discriminant function (score 1) is plotted over time (dashed line) and the smoothed moving average of 3 months (solid line). On the right column, autocorrelation of first discriminant function by lag (months) is shown, where the solid line represents p<0.05 significance level. 73 A positive relationship was observed between occurrence and the average or maximum observed contribution of an OTU (Figure 3-2). Although the most common contribution by any OTU to the community was low (<0.25%), common OTUs (i.e. appearing in 30 or more months) were major components of the community with high maximum (up to 25.2%) or average (up to 6.1%) contributions. Persistent OTUs (both cut and uncut fragments), which were present in all months, tended to have moderate to high contributions. OTUs that occurred infrequently were consistently minor contributors to the total community (i.e. represented no more than 1.5%). Discriminant function and time-series analyses revealed the presence of re-occurring and predictable viral communities. Because discriminant function analysis (DFA) specifically selects the first few dozen taxa that are strong predictors (of month, in this case) while ignoring the remaining community members, we focused our analysis on the “most important” OTUs (i.e. highest percent contributions to the community). The first discriminant function (a measure of community composition) for the 25 most important OTUs in the 3’-HincII assay shows a sinusoidal trend over time in both the raw score (dashed line) and the moving average (solid line) across 3 months (Figure 3-3, A). T4-like myovirus communities were positively auto-correlated at a 1-month lag and negatively auto-correlated for a 3-4 month lag (Figure 3-3, B). Auto- correlation (y-axis) is based on Pearson correlations, shifted by one or more observations (x-axis, lags); statistical significance is indicated by the correlation (bar) crossing the curved line (p<0.05). All enzyme-fragment combinations showed statistically significant positive autocorrelation between viral communities 1 month apart. Positive autocorrelation was observed for communities separated by 9-12 months, although it was not statistically significant (Figure 3- S4). Only 3’-HincII was statistically significant (p<0.05) for negative autocorrelation at 5-6 month lags, but the trend was observed in all four g23-TRFLP assays (Figure 3-S4). 74 Figure 3-4. Individual 3’-HincII TRFLP OTUs exhibiting three distinct seasonal patterns: 1) peaks in spring and summer (top three panels: 353.3, 288.5, 222.7); 2) mainly occurring or abundant in fall and winter (middle three panels: 316.4, 438.7, 261.4); or 3) dynamic but no apparent trend (lower three panels: 381.1, 378.6, 145.0). All patterns were observed over a range of relative abundances; note different y-axis for each panel. Y-axis is the normalized percent contribution of each OTU, e.g. proportion of total amplified OTUs. No data was available for October 2008 and January 2009; zero values indicate the OTU was not detected. 75 Seasonal patterns were observed by Bray-Curtis similarity (Figure 3-3, C-G). Bray-Curtis similarity differs from DFA because it analyzes variance of the whole community weighted by the relative abundance of individual OTUs. We observed the highest Bray-Curtis similarity between communities one month apart and the lowest when 6 or 19-20 months apart, as shown by the troughs in Figure 3-3,C-G. An additional local maximum occurred at 13 months (50% for 3’- RsaI to 55% for 3’-HincII), suggesting communities separated by one year were more similar to each other than those 6 or 19 months apart. The communities had an average Bray-Curtis similarity of 63% between adjacent months for the 5’-HincII (std. dev., ±9), 3’-HincII (±8) and 5’-RsaI (±8) fingerprinting assays with a lower average of 61±10% for the fourth, 3’-RsaI. Although each enzyme-fragment assay (HincII vs. RsaI and 5’ vs. 3’) yielded different similarity values, the overall trends were robust. Individual viral OTUs exhibited repeatable yet contrasting seasonal patterns at SPOT. Some OTUs had spring and summer highs, while others were greater in the fall or winter months (Figure 3-4). For example, 3’HincII OTU 222.7 repeatedly peaked each summer in contrast to 3’-HincII OTUs 316.4 and 438.7 peaked in late fall to early winter (Figure 3-4). Another distinctive pattern included OTUs that were moderately abundant year-round with little fluctuation in their relative abundance. Discussion We identified seasonal patterns in both the myovirus community structure and its individual constituents over time although we did not observe similar seasonal trends in viral or bacterial abundances in the surface ocean (0-5m) at SPOT for the three years shown here. Seasonal differences in viral community composition by g23-TRFLP suggest the presence of distinctive fall-winter and spring-summer communities. Many viral OTUs were observed year- 76 round with only modestly fluctuating seasonal transitions in abundance. The observed shifts in viral community structure are likely reflective of the changing environment (and concomitant host organisms), as previously suggested (Wommack et al. 1999; Short & Suttle 2003; Wang et al. 2011). Episodic peaks in abundance (a.k.a a boom-bust cycle), which would exhibit a high maximum abundance and low occurrence, were virtually absent (i.e. no open circles in the upper right quadrant of Figure 3-2). It seems very unlikely that the smooth monthly transitions of viral OTUs were observed by chance over so many months if in fact they varied randomly. We recognize that given our monthly sampling scheme, we could have easily missed any highly transient, opportunistic virus types (which could be addressed by more frequent sampling). Although the extent to which single gene sequence diversity can accurately represent strain-level or species-level differences in environmental viral genes is still under debate (Lavigne et al. 2009) the uncultivated myoviruses we were able to amplify were dynamic and community structure varied monthly. Across three sequence similarity levels, there was little variation in the number of seqOTUs within each month but the majority of seqOTUs were only observed once (Figure 3-1). Our modest clone libraries (85-91 sequences each) had relatively low coverage in contrast with the higher sensitivity of the g23-TRFLP as we can detect TRFLP OTUs that represent as little as 0.1% of the myoviral community. Both the in silico and environmental TRFLP fingerprints showed a much higher proportion of shared OTUs than the clone libraries (Figure 3-1, Figure 3-S1). It is possible that each OTU can represent more than one sequence, as is true of any TRFLP or other community fingerprinting method, and would result in a higher proportion of shared OTUs for the in silico analysis. When comparing our results to other single-family virus studies, g23-TRFLP identified more OTUs than other common methods; however, because TRFLP potentially groups different sequences together in a single peak, our analysis is a conservative estimate of viral diversity and 77 temporal variation. PFGE, DGGE, RFLP and RAPD typically only detect between 7-36 genotypes from a variety of environments, although these methods can encompass a wider range of families and have utilized different gene markers (Weinbauer & Rassoulzadegan 2003; Sandaa et al. 2010). This g23-TRFLP detected on average 111-130 OTUs per sample, two orders of magnitude higher than a prior TRFLP method for the T4-like portal protein (Wang & Chen 2004) – although inherent difference in within-sample diversity and enzyme selection may account for some of the discrepancy. Our detection of an increased number of OTUs is likely due to the adoption of TRFLP, use of redesigned degenerate primers, and improved fragment resolution and peak analysis. This g23-TRFLP allowed for repeatable detection of a few hundred OTUs and led to the identification of multiple instances of the same underlying patterns of seasonality (or persistence) in individual OTUs. Ecological studies based on the spatial range-abundance or the occupancy-abundance relationship suggests that species with a larger habitat range are more abundant (Brown 1984; Lawton 1993). We find this true when considering time range in lieu of spatial range, where more frequently observed or common viral OTUs were detected at a higher relative abundance than infrequently observed OTUs (Figure 3-2). While it is logical that rare OTUs are observed infrequently or simply undetectable by our method, it is of note that none of these rare OTUs displayed a high maximum or average relative abundance as might be expected from an episodic bloom-like event. Additionally, preliminary analysis of published bacterial data from SPOT for 2000-2004 (Fuhrman et al. 2006) showed a similar trend – where the maximum or average contribution of a bacterial OTU to the community tended to increase with its incidence (data not shown). Future time-series surveys will show if this pattern is a general ecological trend - where more frequently observed organisms are often dominant. 78 The observed myoviral community structure was in some ways consistent with the Bank Model, yet not completely so, suggesting other potential life strategies may occur. Observations consistent with the Bank Model were: (1) most g23-TRFLP OTUs at SPOT were observed at less than 1% of the total amplified T4-like viral community, regardless of how frequently an OTU was observed (Table 3-1, Figure 3-2, Figure 3-S3), (2) seasonally abundant genotypes comprised up to 26% of the myovirus community in a single month, (3) individual OTUs transitioned from undetectable to detectable contributors over time, some with a seasonal pattern, and (4) the identity of the most dominant OTUs changed throughout the year. However, in contrast with the expectation of the Bank Model, a subset of TRFLP OTUs seem to be following an alternative persistent and steady pattern at moderate abundance over all three years. These persistent OTUs are also inconsistent with the “kill the winner” model. Community-level time-series and discriminant function analyses allowed identification of non-random, repeatable patterns in virus community similarity. Time series analysis of the first discriminant function, as a measure of community composition, suggested that repeating temporal patterns exist within the observed myoviral community (Figure 3-3; Figure 3-S4). This measure of community composition is optimal for predicting the month of sampling from OTUs alone. Previous work with bacterial community structure in the photic zone at SPOT has shown repeating seasonal patterns defined by relative stability in community structure from month-to- month but dissimilar community structure when 5-6 months apart (Fuhrman et al. 2006). Viral communities were similarly stable from month to month and negatively correlated when 4-6 months apart, suggesting significant community turnover from season to season (Figure 3-3, Figure 3-S4). Although not all subsets of the viral OTUs showed a statistically significant negative correlation in opposite seasons by discriminant function analysis, the underlying seasonal (sinusoidal) trends in the data remained. 79 Our repeated observations of different temporal patterns in relative contributions over three years indicate persistent relationships within the microbial community and sensitivity to environmental changes. Some OTUs exhibited a summer high as seen by changes in OTU 353.3 or 288.5, (Figure 3-4), while other myoviral OTUs like OTU 222.7 spiked in abundance from <0.01%. Other OTUs were moderately abundant and showed little variation; these OTUs may represent viral types that infect bacterial hosts that are common year-round. Several plausible arguments could be made which support these observations. Although we posit that seasonal viral OTUs reflect the changing abundance of bacterial hosts, these contrasting temporal patterns suggest that even within the T4-like viruses, there is a range of strategies beyond host density dependence, probably including those analogous to r- and k- strategies (Suttle 2007). Additional interpretations include: (1) the hosts of the moderately abundant OTUs are similarly stable year- round (interestingly implying near-stable virus-host coexistence), (2) the persistent OTUs could encompass multiple myovirus types with opposite strategies (one peaks in summer and the other in fall, so that the overall distribution is even), or (3) the OTUs represented may be capable of infecting several hosts and preferentially infect one or the other based on host abundance or other parameters. Further investigation of these potential relationships will allow for a better understanding of virus-host interactions over time within the marine ecosystem. Conclusions Uncultivated viruses are amenable to molecular studies by marker genes for individual viral families from field samples. Through development of g23-TRFLP, we investigated monthly variability of T4-like myoviral OTUs over time. Our results indicate predictable seasonality in marine T4-like virus communities at our oligotrophic to mesotrophic sampling site. Both persistent and seasonal OTUs were observed throughout the three-year time-series in contrast to 80 expectations of both the Bank and “Kill the Winner” models. Individual T4-like OTUs peaked in opposite seasons (e.g. fall/winter or spring/summer). We showed that the contributions of individual OTUs to the T4-like myoviral community structure are consistent with a temporal adaptation of the occupancy-abundance relationship from classical ecology here replacing spatial with temporal range, where OTUS detected more frequently tend to be more abundant. This method is amenable to future studies that may aim to identify individual roles of T4-like myoviral OTUs in shaping co-occurring host communities. Our ability to identify repeatable patterns suggests that this approach can be applied to further discern viral response to the environment or variability in host abundance, host range and turnover rates. An improved understanding of both community-wide and individual variability of viruses and their related host organisms over time can provide insight into the resilience and stability of marine microbial ecosystems. 81 Chapter Three References Angly, F.E., Felts, B., Breitbart, M., Salamon, P., Edwards, R.A., Carlson, C.A., et al. (2006). The marine viromes of four oceanic regions. Plos Biol, 4, e368. Bellec, L., Grimsley, N., Derelle, E., Moreau, H. & Desdevises, Y. (2010). Abundance, spatial distribution and genetic diversity of Ostreococcus tauriviruses in two different environments. Environ Microbiol Rep, 2, 313–321. Bench, S.R., Hanson, T.E., Williamson, K.E., Ghosh, D., Radosovich, M., Wang, K., et al. (2007). Metagenomic Characterization of Chesapeake Bay Virioplankton. Appl. Environ. Microbiol., 73, 7629–7641. Breitbart, M. (2002). Genomic analysis of uncultured marine viral communities. Proc. Natl. Acad. Sci. U.S.A., 99, 14250–14255. Breitbart, M. (2012). Marine Viruses: Truth or Dare. Annu. Rev. Marine. Sci., 4, 425–448. Breitbart, M. & Rohwer, F. (2005). Here a virus, there a virus, everywhere the same virus? Trends Microbiol., 13, 278–284. Breitbart, M., Felts, B., Kelley, S., Mahaffy, J.M., Nulton, J., Salamon, P., et al. (2004). Diversity and population structure of a near-shore marine-sediment viral community. Proc R Soc Lond B, 271, 565–574. Brown, J.H. (1984). On the relationship between abundance and distribution of species. Am. Nat., 255–279. Butina, T.V., Belykh, O.I., Maksimenko, S.Y. & Belikov, S.I. (2010). Phylogenetic diversity of T4-like bacteriophages in Lake Baikal, East Siberia. FEMS Microbiol Lett, 309, 122-129. Børsheim, K.Y. (1993). Native marine bacteriophages. FEMS Microbiol Ecol, 102, 141–159. Clarke, K.R. (1993). Non‐parametric multivariate analyses of changes in community structure. Aust J Ecol, 18, 117–143. Clarke, K.R. & Gorley, R. (2006). PRIMER v6: User Manual/Tutorial. PRIMER-E, Plymouth, UK. Clokie, M.R.J., Millard, A.D., Mehta, J.Y. & Mann, N.H. (2006). Virus isolation studies suggest short-term variations in abundance in natural cyanophage populations of the Indian Ocean. J. Mar. Biol. Ass., 86, 499–505. Collins, R.E. & Rocap, G. (2007). REPK: an analytical web server to select restriction endonucleases for terminal restriction fragment length polymorphism analysis. Nucleic Acids Res, 35, W58–W62. Colombet, J., Robin, A., Lavie, L., Bettarel, Y., Cauchie, H.M. & Sime-Ngando, T. (2007). Virioplankton “pegylation”: Use of PEG (polyethylene glycol) to concentrate and purify viruses in pelagic ecosystems. J. Microbiol. Methods, 71, 212–219. 82 Comeau, A.M. & Krisch, H.M. (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, 1321–1332. Filee, J. (2005). Marine T4-type bacteriophages, a ubiquitous component of the dark matter of the biosphere. Proc. Natl. Acad. Sci. U.S.A., 102, 12471–12476. Frederickson, C.M., Short, S.M. & Suttle, C.A. (2003). The Physical Environment Affects Cyanophage Communities in British Columbia Inlets. Microb Ecol, 46, 348–357. Fuhrman, J.A. (1999). Marine viruses and their biogeochemical and ecological effects. Nature, 399, 541–548. Fuhrman, J.A. & Schwalbach, M. (2003). Viral influence on aquatic bacterial communities. The Biological Bulletin, 204, 192–195. Fuhrman, J.A. & Steele, J.A. (2008). Community structure of marine bacterioplankton: patterns, networks, and relationships to function. Aquat. Microb. Ecol., 53, 69–81. Fuhrman, J.A. & Suttle, C.A. (1993). Viruses in marine planktonic systems. Oceanography, 6, 51–63. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Fujii, T., Nakayama, N., Nishida, M., Sekiya, H., Kato, N., Asakawa, S., et al. (2008). Novel capsid genes (g23) of T4-type bacteriophages in a Japanese paddy field. Soil Biol. Biochem., 40, 1049–1058. Hewson, I. & Fuhrman, J.A. (2006). Viral impacts upon marine bacterioplankton assemblage structure. J. Mar. Biol. Ass., 86, 577–589. Hewson, I. & Fuhrman, J.A. (2007). Covariation of viral parameters with bacterial assemblage richness and diversity in the water column and sediments. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 54, 811–830. Holmfeldt, K., Titelman, J. & Riemann, L. (2010). Virus Production and Lysate Recycling in Different Sub-basins of the Northern Baltic Sea. Microb Ecol, 60, 572–580. Jameson, E., Mann, N.H., Joint, I., Sambles, C. & hling, M.M.U. (2011). The diversity of cyanomyovirus populations along a North-South Atlantic Ocean transect. ISME J, 5, 1713– 1721. Jiang, S., Fu, W., Chu, W. & Fuhrman, J.A. (2003). The Vertical Distribution and Diversity of Marine Bacteriophage at a Station off Southern California. Microb Ecol, 45, 399–410. Labonte, J.M., Reid, K.E. & Suttle, C.A. (2009). Phylogenetic Analysis Indicates Evolutionary Diversity and Environmental Segregation of Marine Podovirus DNA Polymerase Gene Sequences. Appl. Environ. Microbiol., 75, 3634–3640. 83 Larsen, A., Flaten, G.A.F., Sandaa, R.-A., Castberg, T., Thyrhaug, R., Erga, S.R., et al. (2004). Spring phytoplankton bloom dynamics in Norwegian coastal waters: Microbial community succession and diversity. Limnol Oceanogr., 49, 180–190. Lavigne, R., Darius, P., Summer, E.J., Seto, D., Mahadevan, P., Nilsson, A.S., et al. (2009). Classification of Myoviridae bacteriophages using protein sequence similarity. BMC Microbiol, 9, 224. Lawton, J.H. (1993). Range, population abundance and conservation. Trends Ecol Evol, 8, 409– 413. Lopez-Bueno, A., Tamames, J., Velazquez, D., Moya, A., Quesada, A. & Alcami, A. (2009). High Diversity of the Viral Community from an Antarctic Lake. Science, 326, 858–861. Millard, A.D., Zwirglmaier, K., Downey, M.J., Mann, N.H. & Scanlan, D.J. (2009). Comparative genomics of marine cyanomyoviruses reveals the widespread occurrence of Synechococcushost genes localized to a hyperplastic region: implications for mechanisms of cyanophage evolution. Environ Microbiol, 11, 2370–2387. Mühling, M., Fuller, N.J., Millard, A., Somerfield, P.J., Marie, D., Wilson, W.H., et al. (2005). Genetic diversity of marine Synechococcus and co-occurring cyanophage communities: evidence for viral control of phytoplankton. Environ Microbiol, 7, 499–508. Noble, R.T. & Fuhrman, J.A. (1998). Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Aquat. Microb. Ecol., 14, 113–118. Parsons, R.J., Breitbart, M., Lomas, M.W. & Carlson, C.A. (2012). Ocean time-series reveals recurring seasonal patterns of virioplankton dynamics in the northwestern Sargasso Sea. ISME J, 6, 273–284. Patel, A., Noble, R.T., Steele, J.A., Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2007). Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR Green I. Nat Protoc, 2, 269–276. Rodriguez-Brito, B., Li, L., Wegley, L., Furlan, M., Angly, F.E., Breitbart, M., et al. (2010). Viral and microbial community dynamics in four aquatic environments. ISME J, 4, 739–751. Rohwer, F. & Thurber, R.V. (2009). Viruses manipulate the marine environment. Nature, 459, 207–212. Ruan, Q., Steele, J.A., Schwalbach, M.S., Fuhrman, J.A. & Sun, F. (2006). A dynamic programming algorithm for binning microbial community profiles. Bioinformatics, 22, 1508– 1514. Sambrook, M., Fritsch, E.F. & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press, New York, USA. Sandaa, R.-A. (2008). Burden or benefit? Virus–host interactions in the marine environment. Res. Microbiol., 159, 374–381. 84 Sandaa, R.-A. & Larsen, A. (2006). Seasonal Variations in Virus-Host Populations in Norwegian Coastal Waters: Focusing on the Cyanophage Community Infecting Marine Synechococcus spp. Appl. Environ. Microbiol., 72, 4610–4618. Sandaa, R.-A., Clokie, M.R.J. & Mann, N.H. (2008). Photosynthetic genes in viral populations with a large genomic size range from Norwegian coastal waters. FEMS Microbiol Ecol, 63, 2–11. Sandaa, R.-A., Gómez-Consarnau, L., Pinhassi, J., Riemann, L., Malits, A., Weinbauer, M.G., et al. (2009). Viral control of bacterial biodiversity - evidence from a nutrient-enriched marine mesocosm experiment. Environ Microbiol, 11, 2585–2597. Sandaa, R.-A., Short, S.M. & Schroeder, D.C. (2010). Fingerprinting aquatic virus communities. MAVE, 9–18. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., et al. (2009). Introducing mothur: Open-Source, Platform-Independent, Community-Supported Software for Describing and Comparing Microbial Communities. Appl. Environ. Microbiol., 75, 7537– 7541. Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2004). Viral effects on bacterial community composition in marine plankton microcosms. Aquat. Microb. Ecol., 34, 117–127. Short, C.M. & Suttle, C.A. (2005). Nearly identical bacteriophage structural gene sequences are widely distributed in both marine and freshwater environments. Appl. Environ. Microbiol., 71, 480–486. Short, C.M., Rusanova, O. & Short, S.M. (2011). Quantification of virus genes provides evidence for seed-bank populations of phycodnaviruses in Lake Ontario, Canada. ISME J, 5, 810–821. Short, S.M. & Short, C.M. (2009). Quantitative PCR reveals transient and persistent algal viruses in Lake Ontario, Canada. Environ Microbiol, 11, 2639–2648. Short, S.M. & Suttle, C.A. (2002). Sequence analysis of marine virus communities reveals that groups of related algal viruses are widely distributed in nature. Appl. Environ. Microbiol., 68, 1290–1296. Short, S.M. & Suttle, C.A. (2003). Temporal dynamics of natural communities of marine algal viruses and eukaryotes. Aquat. Microb. Ecol., 32, 107–119. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. Steward, G.F. & Culley, A.I. (2010). Extraction and purification of nucleic acids from viruses. MAVE, 154–165. Steward, G.F., Montiel, J.L. & Azam, F. (2000). Genome size distributions indicate variability and similarities among marine viral assemblages from diverse environments. Limnol Oceanogr., 45, 1697–1706. 85 Sullivan, M.B., Huang, K.H., Ignacio-Espinoza, J.C., Berlin, A.M., Kelly, L., Weigele, P.R., et al. (2010). Genomic analysis of oceanic cyanobacterial myoviruses compared with T4-like myoviruses from diverse hosts and environments. Environ Microbiol, 12, 3035–3056. Suttle, C.A. (2007). Marine viruses — major players in the global ecosystem. Nat Rev Micro, 5, 801–812. Thingstad, T.F. (2000). Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol Oceanogr., 1320–1328. Thingstad, T.F. & Lignell, R. (1997). Theoretical models for the control of bacterial growth rate, abundance, diversity and carbon demand. Aquat. Microb. Ecol., 13, 19–27. Wang, K. & Chen, F. (2004). Genetic diversity and population dynamics of cyanophage communities in the Chesapeake Bay. Aquat. Microb. Ecol., 34, 105–116. Wang, K., Wommack, K.E. & Chen, F. (2011). Abundance and Distribution of Synechococcus spp. and Cyanophages in the Chesapeake Bay. Appl. Environ. Microbiol., 77, 7459–7468. Waterbury, J.B. & Valois, F.W. (1993). Resistance to co-occurring phages enables marine Synechococcus communities to coexist with cyanophages abundant in seawater. Appl. Environ. Microbiol., 59, 3393–3399. Weinbauer, M.G. & Rassoulzadegan, F. (2003). Are viruses driving microbial diversification and diversity? Environ Microbiol, 6, 1–11. Williamson, S.J., Cary, S.C., Williamson, K.E., Helton, R.R., Bench, S.R., Winget, D., et al. (2008). Lysogenic virus–host interactions predominate at deep-sea diffuse-flow hydrothermal vents. ISME J, 2, 1112–1121. Wommack, K.E., Hill, R., Kessel, M., Russek-Cohen, E. & Colwell, R.R. (1992). Distribution of viruses in the Chesapeake Bay. Appl. Environ. Microbiol., 58, 2965–2970. Wommack, K.E., Ravel, J., Hill, R., Chun, J. & Colwell, R.R. (1999). Population dynamics of chesapeake bay virioplankton: total-community analysis by pulsed-field gel electrophoresis. Appl. Environ. Microbiol., 65, 231–240. Yooseph, S., Sutton, G., Rusch, D.B., Halpern, A.L., Williamson, S.J., Remington, K., et al. (2007). The Sorcerer II Global Ocean Sampling Expedition: Expanding the Universe of Protein Families. Plos Biol, 5, e16. Zhong, Y., Chen, F., Wilhelm, S.W., Poorvin, L. & Hodson, R.E. (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, 1576–1584. 86 Chapter Four: Identifying controls on bacterial community structure: evaluating connections to viral and protistan communities Abstract Identifying and characterizing relationships - those between and among individuals, communities, and the environment - drives our understanding of the natural world. Current methods to understand the variability and relationships among viruses, bacteria, and protists in the ocean are limited and do not yet allow for systematic identification of multiple and specific relationships. Microbial communities were characterized monthly over three years in the surface ocean (0-5m) as part of the ongoing USC Microbial Observatory program at the San Pedro Ocean Time-series station. Concurrent shifts in community structure were observed by Bray-Curtis community similarity for bacteria, T4-like myoviruses, and protistan communities (as described by molecular fingerprinting assays: ARISA, g23-TRFLP, and 18s rRNA TRFLP, respectively), suggesting similar timing of responses to environmental or community composition shifts. Association networks based upon local similarity correlations (LS) identified individual links between T4-like myoviral, bacterial, and protistan OTUs over time, potentially indicative of synergistic and antagonistic relationships akin to viral lysis, grazing, competition, or other indirect interactions. Introduction Bacterial activity in the ocean is a key driver of biogeochemical cycles; this activity is mediated by bottom-up controls (e.g. resource availability and competition), top-down controls (e.g. predation and viral lysis), as well as possible bacteria-bacteria interactions (e.g. allelopathy). This semi-closed system where the bacteria consume biomass produced by other bacteria following natural death, grazing by protists or viral lysis links all of these organisms and creates a 87 complex microbial community (Azam et al. 1983; Sherr & Sherr 1988; Fuhrman & Suttle 1993; Bratbak et al. 1994; Fuhrman 1999). The dominant top-down forces, or sources of bacterial mortality, in the open ocean environment are viral lysis and protistan grazing. It is still not well- understood how these agents exert selective influences on bacterial community composition. Although widely accepted as inherently important, the proportional influence of each top- down control is still under debate – and may differ based on location and physiological status or identity of the bacteria under study. Many experiments have made attempts to quantify grazing and viral lysis in order to determine the relative impact of each process on structuring the co- occurring microbial communities (McManus & Fuhrman 1988; Fuhrman & Noble 1995; Simek et al. 2001; Sherr & Sherr 2002; Evans et al. 2003; Weinbauer et al. 2003; 2007; Zhang et al. 2007; Baudoux & Veldhuis 2008; Longnecker et al. 2010; Staniewski et al. 2012). Most of these studies enriched or removed the grazers and viruses to investigate short-term or episodic events, from which long-term influences were inferred. Off the Oregon coast, reduction in grazer activity affected bacterial diversity of active cells but removal of viruses only affected the activity rates and not the diversity of the active cells over a 4 day experiment (Longnecker et al. 2010). In an example from a freshwater ecosystem, high grazing rates were concurrent with high viral production rates, although nutrient enrichment had a more significant positive effect on viral production than grazing rates (Weinbauer et al. 2003). Infected cells were more abundant when grazers were also present in the community, than without. Bacterial communities in these experimental mesocosms, with or without viruses or flagellates, also suggested differential effects on the community structure – where the flagellates targeted specific sub-populations. Viruses may also play a role in keeping rare organisms ‘rare,’ by preventing growth through selective mortality – thus providing a foundation of organisms at low abundance within the community (Bouvier & del Giorgio 2007). Past studies of virus community dynamics have often considered 88 viruses as one population, despite its heterogeneity; as each virus-host system may interact differently within the environment, it is important to assess each one individually (Fuller et al., 1998). Indeed, the influence of viruses on bacterial communities has been studied with mixed results, when bulk seawater was enriched with a natural virus concentrate (Schwalbach et al., 2004; Hewson and Fuhrman, 2006). Bottom-up controls can also simultaneously affect bacterial diversity, however, in a manner that is still dependent upon viral or protistan activity (Moebus 1996; Middelboe 2000; Gasol et al. 2002; Corno & Jürgens 2008; Sandaa et al. 2009; Longnecker et al. 2010; Ory et al. 2010; Bouvy et al. 2011). These investigations collectively have revealed close couplings between viruses-bacteria-protists within the aquatic realm, in terms of abundances and activity, but many questions remain in establishing how shifts in top-down controls may affect microbial community structure at natural concentrations and in diverse communities. Ecological networks, based on trophic interactions, have historically been used to characterize complex food webs and both positive and negative interactions within (Sole & Montoya 2001; Dunne 2002; Montoya et al. 2006; Olesen et al. 2011). Stable patterns within the complex interconnected web can be discovered to describe ecologically meaningful interactions. The presence of stable predator-prey modules can give rise to equally stable ecological communities, especially when highly inter-connected (Allesina & Pascual 2007). Network analysis has only recently been applied to microbial communities, and our ability to interpret these networks is still under development (Fuhrman & Steele 2008; Chaffron et al. 2010; Steele et al. 2011; Eiler et al. 2012; Gilbert et al. 2012). Positive correlations may not only suggest co- occurrence but also a potential mutualistic relationship between organisms cooperating within the same niche. Indirect relationships, where a third party benefits from an interaction between two others, may also be detected within these microbial association networks. Negative correlations 89 between individual OTUs may suggest the presence of predation (protist-protist, protist-bacteria), viral lysis (virus-bacteria) and competition (any two nodes); all of these interactions are to be expected in nature. A succession of bacterial nodes may follow viral lysis, represented as time- shifted correlations between virus and bacteria. A series of time-delayed dominant bacterial OTUs linked to environmental, viral, or protistan nodes could indicate whether the environment, viruses, or grazing controls bacterial abundance and community structure. Here, we query a seasonally variable semi-oligotrophic bacterial community in the surface ocean to determine links between protistan, viral, and environmental factors with shifts in bacterial community composition though culture-independent approaches. Past research at the USC Microbial Observatory at the San Pedro Ocean Time-series (SPOT) has interrogated correlations among the smallest plankton, specifically bacterial, protistan, and archaeal operational taxonomic units (OTUs) over time (Fuhrman et al. 2006; Fuhrman & Steele 2008; Steele et al. 2011). Predictable relationships and direct positive and negative correlations were observed between all microbial OTUs. All microbial communities were assessed by culture- independent fingerprinting assays. We focused specifically on the T4-like myovirus family in lieu of the entire viral community, because it is diverse, abundant, detectable through cultivation- independent methods and includes known bacteriophages (Filee 2005; Comeau & Krisch 2008; Clokie et al. 2010; Chow & Fuhrman 2012). Correlations between individual bacterial, protistan, T4-like myoviral, and environmental parameters through local similarity analysis and construction of association networks revealed clusters of OTUs that likely reflect ecologically relevant interactions. We hypothesized that: (1) multiple T4-like virus OTUs will be significantly correlated to bacterial OTUs (as hosts may be susceptible to multiple viruses), (2) each viral OTU may be correlated with multiple bacterial OTUs (suggestive of a virus’ capability to infect multiple hosts), (3) most protistan OTUs may be correlated with multiple bacterial OTUs (as 90 evidence of non-selective grazing) while others may have only a single associated bacterial OTU or taxonomic group (selective grazing or only one prey source available), and (4) strongly inter- correlated clusters of parameters may identify potential ecological niches. Materials and Methods Sample Collection Seawater was collected monthly at 0m or 5m from March 2008 to January 2011 at the University of Southern California Microbial Observatory at the San Pedro Ocean Time Series site at 33’ 33º N, 118’ 24º W offshore from Los Angeles, CA, USA. Approximately, ten liters each were filtered and processed for protistan, bacterial and viral DNA as previously described (Countway et al. 2005; Fuhrman et al. 2006; Vigil et al. 2009; Countway et al. 2010; Steele et al. 2011; Chow & Fuhrman 2012). Microbial community data is unavailable for: 10/2008 (virus), 1/2009 (all), 3/2009 (bacteria), 10-11/2009 (bacteria), 1/2011 (protist). Additional samples were collected for bacterial and viral abundance by SYBR green epifluorescence microscopy, bacterial production by thymidine and leucine incorporation, and nutrient concentrations (Steele et al. 2011). Fingerprinting Microbial Communities Bacteria and Viruses. Bacterial community composition was determined by Automated Ribosomal Intergenic Spacer Analysis (Fisher & Triplett 1999; Brown et al. 2005). T4-like myovirus community structure was analyzed by terminal restriction fragment length polymorphism (TRFLP) of the major capsid gene, gp23 (Chow & Fuhrman 2012). Two viral community fingerprints were obtained each month by analyzing both terminal fragments (5' or 3') after HincII digestion. In all cases, each fragment length represented a single OTU. ARISA and 91 g23-TRFLP fragments were run in duplicate in non-adjacent lanes on an ABI 377 by slab gel electrophoresis, with internal size standards (Bioventures; Murfreesboro, TN) every 25 bp (50- 900bp) or every 50 bp (900-1400bp). Peaks were identified in DAx, and fragment sizes of 400- 1210 bp (ARISA) and 50-500bp (g23-TRFLP), rounded to the nearest 0.1bp, were dynamically binned (Ruan et al. 2006b; Steele et al. 2011; Chow and Fuhrman 2012). The resulting bins were manually curated to merge bins <0.1bp wide with the nearest neighbor; each assay was binned independently. ARISA OTUs were assigned a taxonomic identity based on matching ARISA lengths with known sequences (See Materials and Methods, Chapter 2); in silico analysis of g23 amplicons from whole genome sequences were linked to environmental g23-TRFLP OTUs. Protists. Protistan communities were surveyed by 18S rDNA-based TRFLP using the HaeIII restriction enzyme only (Countway et al. 2005; Vigil et al. 2009); TRFLP fragments were analyzed on a Beckman CEQ 8000. Protistan OTUs were identified based on in silico digestion of 1,341 full-length 18S rRNA gene sequences from October 2001 at SPOT (Diane Kim, pers. comm.). Post-processing. Each fingerprint was standardized by total area. All peaks were considered unique OTUs. Relative abundance of each OTU (each peak) was calculated by dividing peak area by total area within the fingerprint; peaks less than 0.1% of the community were removed from further analysis. Data Analysis Community Analysis. Bray-Curtis Similarity was determined individually for each community dataset, where monthly community composition was standardized so that the total relative abundance of each microbial community in each month equals 100 (PRIMER-E). RELATE function was used to compare resemblance matrices from similarities calculated between samples 92 (time points, in this case) by Spearman correlation coefficient (PRIMER-E). Correlations between Bray-Curtis similarities between adjacent months (i.e., only one month apart) were also calculated by Pearson-product-moment, to test for similarity between microbial communities (Sigmaplot 11). Co-correspondence analysis to determine covariance of microbial community data was completed using the coccorresp package in R (Wommack & Colwell 2000; Braak & Schaffers 2004; Simpson 2009; Thurber 2009) on log-transformed relative abundance data for all months where all microbial community datasets were collected (n=28 months). We excluded any microbial OTUs that occurred in less than 5 months and performed significance testing by cross- validation with the “leave-one-out” method followed by permutation tests (n=99). Covariance of microbial communities and environmental parameters were determined by canonical correspondence analysis using the cca function from the vegan package v2.0.2 (Oksanen et al. 2011). Models were calculated by a step-wise approach and were validated by ANOVA. Estimates for chlorophyll a concentrations and primary production were downloaded from NOAA Coastwatch for the grid area surrounding SPOT from (a) Chlorophyll-a, SeaWiFS, 0.04167 degrees, West US Science Quality + Chlorophyll -a and (b) Primary Productivity, SeaWiFS and Pathfinder, 0.1 degrees, Global, Experimental datasets. Environmental data were transformed as follows: log (value+0.01) – NO 2 , NO 3 , PO 4 , P*, bacterial production by thymidine and leucine incorporation, calculated turnover time, and satellite-based chlorophyll a; square-root – bacterial and viral abundance and the virus:bacteria ratio; none – salinity, temperature, seas surface height differential, primary production (satellite), day length, monthly change in day length. Missing environmental data were filled with the overall mean of the transformed data for this analysis only. Local Similarity and Network Analysis. We determined local similarity values (LS) by eLSA using 2000 permutations and linear interpolation of missing values (Ruan et al. 2006a; Steele et 93 al. 2011; Xia et al. 2011). We excluded any OTU or parameters that occurred in less than 5 months, resulting in 227 bacteria, 376 T4-like viruses (3’-HincII -171, 5’-HincII - 205), 70 protists, and 30 environmental parameters in the final network. Only LS values with a p-value <0.025 and q-value <0.10 were considered significant; these LS correlations were visualized in Cytoscape v2.8.2 (Shannon 2003; Cline et al. 2007; Smoot et al. 2011), as previously described (Fuhrman & Steele 2008; Steele et al. 2011). Example networks were filtered from the overall network by selecting nodes and/or edges based on taxonomic identification (e.g. cyanobacteria) or edge type (e.g. correlations between bacterial and protistan OTU). Network statistics were calculated with the Network Analyzer plugin (Azam et al. 1983; Sherr & Sherr 1988; Fuhrman & Suttle 1993; Bratbak et al. 1994; Fuhrman 1999; Assenov et al. 2008). Results Monthly Covariance in Microbial Communities and the Environment Monthly shifts in community composition were determined by Bray-Curtis community similarity over the three-year period for viral, bacterial, and protistan communities independently (Figure 4-1, Table 4-1). Simultaneous changes in the magnitude or direction of Bray-Curtis similarities were observed, as indicated by positive correlations, for one-month shifts in community composition. Protistan and bacterial communities exhibited greater changes in community structure from one month to the next than the T4-like viral communities (Figure 4- 1A). Changes in protistan community structure were significantly correlated to changes in bacterial community composition, although weakly, only when comparing across all months but not for one-month time lags (Table 4-1). The forward and reverse terminal fragments of the T4- like viral communities are two measures of the same taxa and were significantly correlated as expected for comparisons of community similarity by one-month time lags and all pairwise 94 comparisons (Table 4-1). Changes in bacterial and T4-like viral community structure (by 3’- HincII g23-TRFLP) were also significantly correlated for both 1-month shifts and all months (Figure 4-1D). Month to month shifts in bacterial community structure, specifically, was negatively correlated to sea surface temperature (r=-0.486, p=0.025) and positively correlated with bacterial abundance (r=0.543, p=0.0134) and bacterial production (by thymidine incorporation: r=0.458, p=0.0367). Protistan monthly shifts in Bray-Curtis similarity were positively correlated with temperature differences between months compared (r=0.473, p=0.0353), but not to absolute sea surface temperature. Another metric to determine covariance of communities is co-correspondence analysis (Weinbauer et al. 2003; Braak & Schaffers 2004). This method relies on the individual OTU counts, or relative abundances in this case, to determine covariance or impact of variability in one community to another and has been applied to soil microbial communities, vegetation, and insect studies (Moebus 1996; Middelboe 2000; Gasol et al. 2002; Corno & Jürgens 2008; Sandaa et al. 2009; Longnecker et al. 2010; Ory et al. 2010; Bouvy et al. 2011; Mitchell et al. 2011; Müller et al. 2011). Co-correspondence analyses were unable to predict overall bacterial community variance from T4-like virus or protist community data, despite overall correlation in Bray-Curtis community similarity. Although 47% of the variance of one virus TRFLP assay (3’-HincII) was predictable by the other (5’-HincII), only 20% of the T4-like myoviral community variance (for 3’-HincII) was significantly predictable by bacterial community composition and none was significantly predictable by protistan community composition. Protistan community variance was surprisingly significantly predictable by both T4-like viral communities (5’H, 9.5% and 3’H, 8.2%) but not by overall bacterial community composition. 95 Figure 4-1. Month – to – month shifts in Bray-Curtis Similarity within microbial communities. (A) Average similarity within microbial communities. Line, average similarity; box, 25 th and 75 th percentiles; error bars, 10 th and 90th percentiles. (B) Bray-Curtis similarity between adjacent months, plotted over time for each microbial group. (C) Correlation of bacterial Bray-Curtis similarity to sea surface temperature and (D) viral (3’H) Bray-Curtis similarity. 96 Table 4-1. Correlation of Bray-Curtis similarities between microbial communities. Correlation between Bray-Curtis resemblance matrices comparing all months to each other (white area, lower triangle) and just communities in adjacent months (i.e., only 1 month apart, gray shaded area, upper triangle). Bold text indicates statistically significant correlations . Bacteria Virus (3’-H) Virus (5’-H) Protist Bacteria r = 0.540 p = 0.01 r = 0.525 p = 0.01 r = 0.117 p = 0.60 T4-like Virus (3’-HincII) r = 0.21 p = 0.033 r = 0.883 p <0.01 r = 0.154 p = 0.493 T4-like Virus (5’-HincII) r = 0.174 p = 0.054 r = 0.771 p = 0.001 r = 0.0883 p = 0.696 Protist r = 0.246 p = 0.015 r = 0.046 p = 0.346 r =0.118 p =0.157 Microbial communities were predictable by environmental descriptors, primarily day- length, change in day-length, sea surface temperature, salinity and/or satellite-based estimates of primary productivity or chlorophyll a concentrations by Canonical Correspondence Analysis (CCA). T4-like viruses were significantly predictable (22.1% of variation explained) at p≤0.01 by day length, change in day length, and estimated primary production as compared to the null model. Neither bacteria nor protists were predictable by any of the measured environmental parameters at this significance level. All microbial communities were significantly predictable by one or up to four of the parameters at p≤0.05, when compared to the null model. Bacteria were predictable by chlorophyll a and salinity, which explained 12.3% of the variability; protists by day length and bacterial abundance for 11.6%, T4-like viruses by day length, change in day length, salinity, and temperature for 28.1%. Connections between individual microbial taxa in association networks Many significant LS correlations were observed between viral, bacterial, and protistan OTUs to each other and to environmental parameters (Table 4-2, Figure 4-S1). After significance testing by permutation tests and screening by both p-values and false-discovery estimates (q- 97 values), only 2.5% of all possible pairwise LS values were retained for subsequent analysis (Figure 4-S1). Most non-significant LS values ranged from -0.2 to 0.2; positive LS values were more frequent than negative ones. The number of significant correlations between bacterial OTUs and either group of T4-like virus OTUs outnumbered the protistan connections, although this may be due to the reduced number of protistan nodes retained in the final network (Table 4- 2); each OTU or environmental parameter is shown as a distinct shape, a node, within the networks. For simplicity, LS correlations for only one viral dataset, 3’-HincII g23-TRFLP was shown, as the distribution of connections for both viral assays were largely similar (Table 4-2). We focused on the 3’-HincII OTUs since its overall Bray-Curtis community similarity was significantly correlated to the bacterial community (Table 4-1). Figure 4-S1. Distribution of LS values, rounded to the nearest tenth. LS values are shown along the x- axis, and their frequency as log (total counts) for all LS values (grey) and for LS values with a q-value ≤ 0.1 (black). -0.5 0 0.5 1 10 1 10 2 10 3 10 4 log (Counts) 98 Table 4-2. Distribution of significant LSA correlations (edges) between all microbial OTUs and environmental parameters. Columns ‘Nodes’ indicate number of nodes each included in the global networks and the remaining columns and rows indicate node type. ‘Biological and Chemical’ includes: bacterial and viral abundances, nutrient concentrations, chlorophyll a, etc.; ‘Physical’ includes: salinity, temperature, day length, change in day length, etc. Nodes Bacteria Virus (3’-H) Virus (5’-H) Protist Bio+Chem Phys Bacteria 227 777 T4-like Virus (3’-HincII) 168 936 576 T4-like Virus (5’-HincII) 199 899 1133 630 Protist 60 96 92 118 31 Biological or Chemical 21 220 78 65 11 22 Physical Oceanography 6 36 69 74 7 13 9 Table 4-3. Description of LS correlations for exemplar networks in Figure 2. Interactions (Int.) are abbreviated up to three letters, based on whether correlations are: 1) positive (p) or negative (n); no delay (u), or delayed (dr, where Y is the lagging OTU). ^ indicates the OTU whose abundances were ‘fixed’ in time; * denotes which OTU’s relative abundances were shifted by one month in Figure 2. X and Y note the month where the LS correlation begins while ‘L,’ abbreviation for Length, indicates the length of the LS correlation in months. PCC and P PCC are the Pearson’s Correlation Coefficient with no time lag and associated p-value, respectively, for the OTUs listed. OTU (x) OTU (y) Int. LS x y L PCC P PCC Network 1 Dinoflagellate_180 OTU_607.4 pu 0.556 1 1 34 0.698 0 Network 2 SAR11_Aegean-169_653.1 Formos/SAR92_762.8 pu 0.544 1 1 34 0.544 9.00E-04 SAR11_Aegean-169_653.1 SAR11_S4_703.7 pu 0.521 2 2 33 0.363 0.0349 Formos/SAR92_762.8 3H_296.9* pdr 0.508 4 3 31 0.581 3.00E-04 Network 3 3H_372.6 SAR11_S1_686.9 nu -0.540 3 3 32 -0.641 0 3H_372.6 3H_430.2 pu 0.512 1 1 31 0.540 0.001 3H_207.5* 3H_372.6^ ndr -0.523 3 2 30 -0.404 0.0179 3H_207.5* 3H_430.2^ ndr -0.548 3 2 31 -0.449 0.0078 99 Figure 4-2. Three mini networks and the relative abundance of each OTU over time for LS correlations observed in the overall T4-like virus, protist, and bacteria microbial association network. Each mini network shows microbial OTUS as nodes (bacteria, circles; protists, diamonds; viruses, V-shapes). Lines represent LS correlations with LS values as edge labels: solid lines = positive correlation, dashed lines = negative correlations, arrows = delayed correlations, pointing towards lagging OTU. * denotes time-shifted OTU, as described in Table 3. Relative abundance of each node is shown as a percent of bacterial, protistan, or T4-like myoviral community from Mar 2008 – Jan 2011 (B-E). 0.512 −0.540 −0.523 −0.548 3H_372.6 SAR11_S1_686.9 3H_430.2 3H_207.5 Network 3 0.508 0.521 0.544 3H_296.9 SAR11_Aegean-169_653.1 Formos/SAR92_762.8 SAR11_S4_703.7 Network 2 0.556 OTU_607.4 Dinoflagellate_180 Network 1 A 0 1 2 3 4 5 6 7 8 Dinoflagellate_180 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 OTU_607.4 D S J M D S J M D S J M Rel. Abundance of Protist Rel. Abundance of Bacterium B E 0.0 0.1 0.2 0.3 0.4 0.5 3H_296.9* 3H_296.9 D S J M D S J M D S J M 0 1 2 3 4 5 6 7 8 Formos/SAR92_762.8 SAR11_S4_703.7 SAR11_Aegean-169_653.1 Rel. Abundance of Bacteria Rel. Abundance of 3H_296.9 Rel. Abundance of OTU 0 3 6 9 12 15 3H_372.6 3H_430.2 SAR11_S1_686.9 D S J M D S J M D S J M D Rel. Abundance of OTU 0 1 2 3 4 5 6 7 8 3H_207.5* 3H_207.5 3H_372.6 3H_430.2 D S J M D S J M D S J M C 100 To demonstrate how networks relate to original relative abundance data, Figure 4-2 depicts some simple networks and the data they derive from. These example networks illustrate the some of the types of correlations that occurred within the complex “whole community” network. All potential types of LS correlations were observed - same and different node types were correlated (i.e. bacteria-bacteria, protist-bacteria, virus-bacteria) by both positive and negative LS values. Co-occurring (not delayed) and time-shifted (delayed by one month) LS correlations were also observed. The Pearson’s correlation coefficient for all pairwise interactions, with no time delay, was calculated and in most cases is close to the LS value (Table 4-3). In Network 1, a bacterial and a protistan node were positively correlated – potentially indicative of co-occurrence, mutualistic or predator-prey interactions (Figures 4-2A, 4-2B). In Network 2, positive correlations were observed between bacterial nodes as well as a positive time-shifted LS correlation from a bacterial to a viral node – potentially reflective of a lytic relationship given the monthly sampling frequency (Figures 4-2A, 4-2C). In Network 3, negative and undelayed LS values are shown between one virus, 3H_372.6, and one bacterium, SAR11_S1_686.9 (Figure 4-2A, 4-2D). Negative delayed correlations are shown between three virus nodes (Figure 4-2A, 4-2D, 4-2E) – all of which may indicate a competitive relationship for host bacteria. The structure of bacteria-virus interactions and bacteria-protist interactions differ remarkably in their interconnectedness (Figure 4-3). When looking at just the LS correlations between bacterial and protistan OTUs, each bacterial node is typically only correlated to a single protist although a protist is often correlated with relatively few bacteria (115 nodes, 96 correlations, 23 sub clusters). In contrast, a typical viral node is connected to several bacterial nodes and vice versa, resulting in one large interconnected cluster and a few ancillary pairs when considering just virus-bacteria interactions (308 nodes, 936 correlations, 8 clusters). Only 2% 101 (n=2) of protist-bacteria correlations are negative, in comparison to 16.6% (n=156) of virus- bacteria correlations. On average, a node in the protist-bacteria networks has only 1.7 connections. That number increased to 6.1 connections in the virus-bacteria network. When including secondary connections between nodes of the same type (i.e., bacteria-bacteria and protist-protist), many of the smaller clusters become loosely connected; however, subgroups of interconnected nodes remain clear. The T4-like myovirus-bacteria network only becomes more interwoven through the addition of virus-virus and bacteria-bacteria interactions. Figure 4-3. Broad overview of interactions between (A) protists and bacteria and (B) T4-like viruses and bacteria. Microbial nodes are: bacteria, circles; protists, diamonds; viruses, V-shapes. Solid lines = positive LS correlations, dashed lines = negative LS correlations, arrows = delayed LS correlations, pointing towards lagging OTU. A B 102 Figure 4-S2. Bacteria-Protist correlations. Microbial nodes are: bacteria, circles; protists, diamonds. Solid lines = positive LS correlations, dashed lines = negative LS correlations, arrows = delayed LS correlations, pointing towards lagging OTU.. Pro_HL(I)_828.8 Dinoflagellate/Lingulo culture_336 NS9_639.8 Unk_77 SAR11_S2_718.4 Unk_124 Dinoflagellate/Prorocentrum culture/Gonyaulax culture_338 Unk_177 Syn_A.V_1053.2 Ciliate_590 OTU_406.7 OCS155_419.5 Ciliate_273 OCS155_423.3 OTU_416.4 SAR86_538.9 Ichtyosporea_593 Ciliate_272 OTU_431.4 SAR11_662 OTU_542.8 Telonema_331 Plastid_573.6 SAR324_519 Shewan_588.6 Unk_127 Fluvii/SAR406_622.5 SAR11_S1_692.2 OTU_519.6 OTU_481.8 Telonema/Dinoflagellate/Stramenopile_328 Unk_70 NS5_749.6 Pro_HL(I)_831.8 Unk_330 SAR11_682.4 Unk_448 OTU_742.6 SAR86_402.4 Microb_637.2 Plastid_541.5 SAR11_S1_670.5 Ostreococcus_259 Pro_HL(II)_820.5 SAR86_557.8 SAR406_624.5 Unk_222 SAR11_S2_721.2 OTU_404.8 Ciliate_592 SAR86_405.6 OCS116_852.1 OTU_1191 Unk_58 Plastid_589.7 Unk_99 OTU_436.3 OTU_591.2 Rhodob_897.8 ZD0417_759.8 OTU_1192.4 Unk_104 OTU_609.4 OM75_889.2 Unk_64 SAR11_694.1 OTU_607.4 Dinoflagellate_180 Unk_67 Shewan_644.9 SAR86_399.1 Shewan_868.8 OCS155_421.5 SAR11_S1/3_674.2 Unk_193 MAST/Stramenopile_228 Haptophyte/Acantharea/Fungi_274 OCS116_859.8 Telonema/Stramenopile/Dinoflagellate/Viridiplantae_332 Unk_268 SAR11_S2_716.8 OTU_662.8 OTU_884.3 SAR11_S2_713.9 Ciliate/Stramenopile/Metazoa_276 OM182_767.5 SAR11_S1_663.4 Ciliate/unk. Alveolate_179 Roseob_1185.8 Rhodob_951.6 NS7_613.3 Roseob_987.8 OTU_486.6 NS5_750.4 Metazoa_498 Mariba/Alcani_802.9 Dinoflagellate_337 OTU_614.9 Rhodob/Chromat_946.3 Cercozoa_596 Pseudo_1037.7 SAR86_531.5 NS2b_726.4 Dino Group I_236 Dinoflagellate/Micromonas_326 OTU_533.4 Alveolate_333 OTU_618.6 OTU_611 OTU_764.5 Haptophyte/Dino Group I_329 OCS155_417 NS9_529.5 NS2b_741.8 Dino/Alex sp._600 103 Figure 4-S3. Bacteria and 3H-Virus interactions. Microbial nodes are: bacteria, circles; viruses, V- shapes. Solid lines = positive LS correlations, dashed lines = negative LS correlations, arrows = delayed LS correlations, pointing towards lagging OTU. NS9_683.9 3H_373.5 OTU_630.6 OTU_854.8 Formos/SAR92_762.8 3H_443.6 Plastid_566.5 SAR406_627.8 3H_419.3 3H_376.3 OCS155_419.5 3H_287.4 SAR86_484.1 SAR11_S2_718.4 Owenwe_563.6 3H_418.3 SAR11_S2_721.2 3H_392.8 3H_402.4 OTU_436.3 Mariba/Alcani_802.9 SAR86_399.1 NS9_639.8 3H_370.7 OTU_407.7 3H_378.6 SAR11_S1_666.4 OTU_432 3H_137.3 3H_286.4 3H_428 NS4_729.4 3H_442.1 3H_427.5 3H_430.2 3H_341.1 SAR11_694.1 3H_370 OTU_671.2 SAR86_473 Plastid_570.1 OTU_798.7 SAR11_S2_716.8 3H_435.1 3H_412.7 Tricho_752.3 Flavo/Pro_874 OTU_804.9 3H_325.7 Shewan_588.6 3H_438.7 3H_279.2 3H_421.3 NS4/SAR86_579.3 3H_192.8 OM75_887.8 3H_193.8 NS2b_726.4 OTU_404.8 3H_471.4 3H_454.6 3H_433.6 OTU_535 3H_296.9 NS2b_738.8 3H_313.4 NS2b_741.8 OTU_632.6 3H_406.4 3H_395.4 3H_197.8 Fluvii/SAR406_622.5 SAR11_584 OCS155_420.5 3H_265.3 Sedimi/Punice_791.4 Plastid_589.7 3H_194.4 3H_362.1 3H_316.4 3H_261.4 3H_349 3H_382.2 Plastid_541.5 SAR406_624.5 OTU_454 Shewan_602.5 OTU_591.6 Roseob_1185.8 OTU_559.7 OTU_534.1 Pro_HL(II)_823.1 OTU_742.6 OTU_408.8 Pro_HL(II)_820.5 OTU_764.5 3H_404.6 3H_343.2 OCS155_414.2 SAR116_400.2 3H_128.3 3H_150.1 3H_400.4 3H_385.3 3H_369 3H_186.3 Owenwe/SAR116_654.9 3H_301.5 3H_148.3 OTU_607.4 SAR11_S1_692.2 Shewan_698.2 Piscir_833.8 SAR406_600.6 3H_207.5 Bacter_505.7 Shewan_868.8 OTU_515.9 3H_426.4 OCS155_891.9 3H_359.2 OTU_486.6 OTU_614.9 3H_416.2 OTU_487.4 OTU_611 3H_414.5 OCS116_859.8 3H_330.6 OM75_889.2 OTU_481.8 OTU_609.4 OTU_1192.4 Pseudo_1037.7 Pseudo/OM60_937.8 3H_153.6 OTU_542.8 OTU_403.8 SAR324_519 3H_366.9 3H_448.9 SAR11_667.6 OTU_564.5 OTU_491.7 OTU_392.9 3H_424.4 Acidi_575.6 OTU_884.3 SAR86_557.8 OTU_431.4 OTU_1191 OTU_438.2 Syn_1020.8 OM182_767.5 3H_340.3 SAR11_S1_686.9 NS7_613.3 Hellea_993.6 SAR116_765.7 3H_334.8 OTU_511 ZD0417_759.8 3H_360.4 SAR324_1023.2 3H_432.7 SAR86_469.4 Photob_857.1 3H_314.9 3H_345.1 OTU_963 3H_408.9 Owenwe_616.2 3H_389 3H_423.9 3H_283.7 SAR11_S1_663.4 3H_266.6 OTU_416.4 OCS116_795.9 OTU_788.5 OTU_756.5 OTU_782 OCS155_871.5 3H_393.8 3H_158.6 SAR86_531.5 NS5_749.6 SAR116_744.7 Bacter_478.8 OTU_846 3H_411.9 Microb_706.4 NS9_540.1 Syn_A.I_1056.1 OTU_618.6 NS5_969.6 Microb_637.2 OTU_522.8 OCS155_417 Plastid_596.8 3H_397.1 OTU_543.8 SAR11_S2_712.4 3H_387.1 3H_364 3H_388.1 3H_416.7 NS9_732.2 OTU_582.1 SAR11_AEGEAN169_676.9 Pro_HL(II)_795 3H_170.8 3H_130.3 3H_95.4 Formos_773.1 SAR86_618.3 3H_328.1 3H_446.5 3H_431.8 SAR11_S4_703.7 3H_431 OTU_763.9 OCS155_418.5 Comamo_964.8 SAR116_657.6 3H_444.9 OM43_836.8 3H_282.6 Plastid_567.5 OCS155_423.3 OTU_792.6 3H_422.9 OTU_439.2 3H_363.1 SAR11_662 NS9_529.5 OTU_545.8 Rhodob_897.8 Pro_HL(II)_933.7 Shewan_644.9 3H_380 3H_101.7 OCS155_421.5 3H_364.5 SAR406_521.6 SAR11_AEGEAN ï 169_653.1 OCS155_424.4 SAR86_538.9 3H_356.4 3H_222.7 OTU_490 OTU_480 Plastid_561.8 OTU_533.4 Plastid_573.6 3H_146.6 SAR86_532.4 3H_132.4 3H_336.4 3H_333.7 3H_372.6 3H_269.8 3H_391.9 Pro_LL(I)_912.5 SAR11_682.4 Roseob_987.8 3H_332.6 SAR11_S1_670.5 3H_172.5 3H_288.5 3H_336 3H_415.5 Pro_HL(I)_828.8 3H_353.3 NS5_750.4 SAR86_536.9 3H_403.2 3H_276.5 Rhodob_951.6 3H_365.7 SAR86_700.7 OTU_519.6 SAR86_405.6 OTU_591.2 SAR86_402.4 3H_347.1 SAR86_471.9 3H_371.6 3H_405.7 OTU_526.8 OTU_425.6 Pro_LL(IV)/OM60_907.8 Cronob_620.5 Pro_HL(I)_825.8 Sphing_1011.5 3H_401.4 OTU_808 3H_375.5 3H_407.3 3H_417.6 Rhodob/Chromat_946.3 3H_342.4 3H_284.9 3H_331.7 3H_399.2 OTU_704.7 3H_384.3 3H_323.1 104 Figure 4-4. Network of Top 5 bacterial OTUs and primary LS correlations to their first neighbors. Microbial nodes are: bacteria, circles; protists, diamonds; viruses, V-shapes. Solid lines = positive LS correlations, dashed lines = negative LS correlations, arrows = delayed LS correlations, pointing towards lagging OTU. Five ARISA OTUs dominate the surface ocean bacterial communities, based on average relative abundance and overall contribution to Bray-Curtis similarity between months. These five OTUs - 4 putative SAR11 and 1 Actinobacterium (435.5) - are significantly correlated to a number of other bacterial and viral OTUs, two protistan OTUs, to build a network of 2 connected components, 76 nodes, and 80 edges (Figure 4-4; 363 edges with secondary connections, Figure 4-S4). Note that these top 5 OTUs are only indirectly connected, although they are directly correlated in a parallel ten-year dataset for the surface ocean (not shown). The two protistan OTUs, identified as Ostreococcus_259 and Ichtyosporea_593, were both correlated with delays to two of the dominant SAR11 OTUs. Addition of the secondary connections between the 0.3024 0.3133 0.3954 0.3289 0.3318 0.3004 0.4096 0.2759 ï 0.2693 ï 0.273 ï 0.4041 ï 0.2601 ï 0.2733 ï 0.2564 ï 0.4028 ï 0.3191 ï 0.3517 ï 0.267 ï 0.5051 ï 0.3577 ï 0.2742 ï 0.28 ï 0.2641 ï 0.4169 ï 0.3705 ï 0.3739 ï 0.2958 ï 0.2875 ï 0.3097 ï 0.2623 0.4418 ï 0.4767 0.372 0.4386 0.2947 0.3545 0.3229 0.2613 0.3076 0.3558 0.3592 0.3633 0.2866 0.3165 0.3485 0.3171 0.3255 0.3179 ï 0.4412 0.3908 0.3061 0.3848 0.3384 0.2998 0.2823 0.3907 ï 0.4832 ï 0.33 0.4167 ï 0.2711 ï 0.3029 ï 0.3283 ï 0.4087 ï 0.3318 0.4129 ï 0.507 ï 0.4534 ï 0.4379 ï 0.3806 ï 0.4642 ï 0.266 ï 0.3303 0.3138 0.3109 0.2925 ï 0.5401 0.2758 ï 0.481 ï 0.2703 0.2738 Hellea_993.6 Ichtyosporea_593 3H_404.6 3H_148.3 Pseudo_1037.7 SAR11_662 OCS155_435.5 OTU_545.8 SAR116_744.7 Formos_770.5 NS2b_741.8 NS4_729.4 Shewan_868.8 OTU_625.9 Formos_785 OCS155_871.5 OTU_522.8 SAR11_S2_712.4 NS2b_724.3 SAR406_600.6 SAR11_AEGEAN ï 169_650.2 NS2b_738.8 SAR11_679.4 OCS155_424.4 SAR11_735.5 Plastid_567.5 SAR11_S1_666.4 OTU_591.6 Shewan_644.9 OTU_511 OTU_491.7 OTU_439.2 OTU_487.4 SAR86_399.1 SAR116_400.2 OCS155_418.5 3H_360.4 3H_342.4 Ostreococcus_259 OTU_614.9 NS9_639.8 3H_343.2 Plastid_566.5 OTU_481.8 OTU_630.6 3H_296.9 3H_428 SAR86_469.4 OCS155_423.3 Owenwe/SAR116_654.9 3H_384.3 SAR11_S1_670.5 3H_375.5 3H_186.3 3H_345.1 3H_314.9 OTU_392.9 OTU_564.5 3H_323.1 3H_359.2 3H_385.3 3H_334.8 OTU_416.4 OTU_438.2 Plastid_596.8 3H_372.6 OTU_432 SAR11_S1_686.9 OTU_431.4 Bacter_505.7 3H_366.9 NS9_732.2 SAR86_557.8 OCS155_417 OTU_486.6 OTU_533.4 105 ‘neighbors’ revealed two interconnected clusters with limited correlations between them (Figure 4-S4). In one cluster, Actino_435.5 is a negatively-correlated hub to a suite of other bacterial OTUs that are positively correlated to one another. The other top four SAR11 OTUs are correlated to both bacterial and viral OTUs, with a number of the connections showing a time delay of one month. Figure 4-S4. Top 5 OTUs and their first neighbors, including all edges. Microbial nodes are: bacteria, circles; protists, diamonds; viruses, V-shapes. Solid lines = positive LS correlations, dashed lines = negative LS correlations, arrows = delayed LS correlations, pointing towards lagging OTU. ï 0.3462 0.3584 ï 0.3151 0.3431 0.32 0.3848 0.3871 0.3211 0.294 0.3477 0.2947 0.6289 0.3898 0.5107 ï 0.4832 ï 0.2941 0.4611 0.3322 ï 0.2941 0.5513 0.2919 ï 0.2668 ï 0.3354 ï 0.274 0.2919 0.4096 0.3026 0.3135 0.2812 0.3049 ï 0.3318 ï 0.3134 ï 0.2713 0.5177 0.2987 0.3894 ï 0.2861 0.4608 0.3615 0.2815 ï 0.3564 ï 0.3501 ï 0.348 0.4992 0.3291 ï 0.3602 ï 0.2634 ï 0.2588 0.331 0.3954 0.5137 ï 0.3666 ï 0.3197 ï 0.3393 0.3712 0.4129 0.3079 0.3521 0.3217 0.3469 0.2866 0.3877 ï 0.4194 0.3684 0.2843 0.3811 0.3275 0.3185 0.3385 ï 0.393 0.2889 0.2994 0.3097 0.3276 0.3421 0.4386 0.3577 0.3302 ï 0.4087 0.2683 0.4096 0.3171 0.3681 0.3337 0.3789 0.5414 ï 0.2996 0.3109 0.3384 ï 0.2696 0.4922 0.4837 0.4541 0.5859 0.2738 0.3179 0.5528 0.5225 ï 0.3021 0.3545 0.2925 0.3633 ï 0.2933 ï 0.2784 ï 0.5401 0.3558 0.3292 1.0709 0.3908 0.9848 0.5845 0.6786 0.7846 0.372 ï 0.4767 0.3076 0.3165 ï 0.4412 0.3061 0.3907 0.2823 0.3255 0.3229 0.2866 0.2613 0.2998 0.3485 0.3592 0.2758 0.4448 0.5414 0.2792 0.2806 0.452 0.2883 0.5432 0.601 0.4685 0.3892 0.5002 0.3552 0.3105 0.2816 0.5882 0.3328 0.5105 ï 0.2641 0.5429 0.3049 0.3422 0.3808 0.299 ï 0.2958 0.3113 0.3531 ï 0.317 0.3382 0.2771 0.3276 0.3052 0.405 0.303 0.3319 ï 0.4028 0.3619 0.3389 0.3019 0.3467 ï 0.3747 0.2969 0.3862 0.3235 0.3106 0.2855 0.308 ï 0.3577 0.3149 0.2736 0.3113 ï 0.4025 0.3026 0.3073 0.3111 0.3778 0.3028 0.299 0.2952 0.2707 0.3699 0.3205 0.4241 ï 0.4041 0.3019 0.2953 0.3497 ï 0.3567 0.3419 0.2731 0.3077 0.3888 0.3852 0.3861 0.3472 ï 0.2875 0.3633 0.3323 0.3118 0.372 0.2898 0.3754 0.3859 0.2821 0.302 ï 0.2742 ï 0.273 ï 0.2601 ï 0.2564 ï 0.4169 ï 0.3191 ï 0.5051 ï 0.3517 ï 0.3705 ï 0.3097 ï 0.2733 ï 0.2623 ï 0.28 ï 0.2693 ï 0.267 ï 0.3739 0.4418 0.9848 ï 0.3083 0.3265 0.3229 0.3036 0.3247 0.3154 0.3385 0.3138 ï 0.481 ï 0.2703 0.2988 ï 0.2665 0.3018 0.3344 0.3035 0.3057 0.3193 0.5138 0.3024 0.4805 0.3946 ï 0.4534 0.3411 0.3101 0.2638 0.4328 0.2655 ï 0.3136 ï 0.316 0.3769 0.3599 0.3485 ï 0.3303 0.2643 0.2808 0.3601 ï 0.4642 0.4519 0.3034 0.3864 0.3015 0.3367 0.317 0.3995 0.2667 ï 0.4379 0.2606 0.2608 0.2975 0.2739 0.262 0.3437 0.3225 0.3894 0.3897 0.3338 0.3585 0.3492 0.3101 0.3002 0.2955 0.3499 0.3799 0.29 0.3718 0.3776 0.3659 0.3404 0.3456 0.3617 0.3574 0.3659 0.2993 0.2843 0.5603 0.6167 0.4657 0.4947 0.3148 0.3645 0.526 0.4562 0.5341 0.2941 0.3039 0.4982 0.3244 0.6162 0.5252 0.4621 0.3568 0.5349 0.4696 0.3253 0.3205 0.3194 0.3061 0.2954 0.3387 0.4652 ï 0.507 0.2804 0.3263 0.325 0.2973 0.2713 0.3306 0.2631 0.2962 0.3255 0.2691 ï 0.3806 0.315 0.316 0.2862 0.2873 0.3361 0.3625 0.3481 0.3128 0.343 0.5751 0.4729 0.4245 0.5021 0.7275 ï 0.2711 0.2759 0.5872 0.4953 0.4555 0.4353 ï 0.3029 0.4908 0.3975 0.2963 ï 0.33 0.3 0.3181 0.3289 ï 0.266 0.3512 0.3133 0.3523 0.3318 0.4167 ï 0.3283 0.3004 0.3101 0.2531 ï 0.47 0.3451 OCS155_418.5 OTU_439.2 3H_360.4 3H_323.1 3H_404.6 3H_385.3 3H_384.3 OCS155_423.3 OTU_416.4 OTU_431.4 Shewan_644.9 OTU_438.2 3H_372.6 Plastid_596.8 OTU_487.4 SAR11_S1_666.4 OTU_432 OTU_511 OCS155_417 OTU_486.6 OTU_533.4 SAR86_469.4 OTU_545.8 NS2b_738.8 NS9_732.2 SAR11_679.4 OTU_630.6 Formos_770.5 OCS155_435.5 OTU_591.6 OTU_491.7 3H_334.8 SAR11_S1_686.9 3H_366.9 3H_428 SAR86_399.1 Hellea_993.6 3H_345.1 SAR116_400.2 OTU_392.9 Bacter_505.7 3H_359.2 3H_375.5 Owenwe/SAR116_654.9 SAR11_735.5 Plastid_566.5 OTU_522.8 OCS155_871.5 NS2b_724.3 3H_186.3 3H_314.9 OTU_625.9 SAR406_600.6 OTU_614.9 OTU_481.8 SAR116_744.7 NS9_639.8 Plastid_567.5 Shewan_868.8 3H_296.9 3H_342.4 3H_148.3 OTU_564.5 Ichtyosporea_593 Pseudo_1037.7 Ostreococcus_259 SAR86_557.8 3H_343.2 SAR11_662 SAR11_S1_670.5 NS4_729.4 NS2b_741.8 SAR11_S2_712.4 Formos_785 OCS155_424.4 SAR11_AEGEAN ï 169_650.2 106 Figure 4-5. Cyanobacteria OTUs and their first neighbors. Prochlorococcus OTUs are noted in light green and Synechococcus OTUs in pink. Microbial nodes are: bacteria, circles; protists, diamonds; viruses, V-shapes; biological or chemical oceanographic parameters, squares; physical oceanographic parameters, hexagons. Solid lines = positive LS correlations, dashed lines = negative LS correlations, arrows = delayed LS correlations, pointing towards lagging OTU. 0.3264 0.3115 0.3441 0.278 0.3086 0.358 0.258 0.2758 0.2639 0.2541 0.4921 0.3357 0.3573 0.3131 0.2867 ï 0.2744 ï 0.2714 ï 0.3296 0.2979 0.2705 0.3094 0.2868 0.3255 0.2779 0.325 0.3031 0.2771 0.3298 0.268 0.2729 0.2968 0.3208 0.3042 0.2432 0.2912 0.2666 0.3202 0.2522 0.2676 0.3591 0.2513 0.3303 0.2868 0.3124 0.2915 0.2648 0.2536 0.2544 0.3336 0.6716 ï 0.3048 0.3662 0.4214 0.2645 0.3159 0.3494 0.319 0.4349 0.6741 0.245 0.3435 0.273 0.3405 0.4032 0.3115 0.3025 0.37 0.2835 0.3249 0.3321 0.3678 0.3143 0.3004 0.3069 0.2857 0.3271 0.563 0.321 0.3662 0.7369 0.3044 0.458 0.4822 0.2606 Formos_785 Formos_770.5 3H_130.3 3H_170.8 Pro_HL(II)_795 OM43_836.8 OTU_662.8 Pro_HL(I)_831.8 Unk_70 Unk_330 3H_330.6 Dinoflagellate/Lingulo culture_336 3H_332.6 3H_128.3 Prim_Prod 3H_282.6 3H_148.3 NS5_969.6 3H_360.4 NS5_747.3 3H_369 3H_150.1 Pro_HL(II)_823.1 Comamo_964.8 3H_345.1 3H_359.2 MAST/Stramenopile_228 Pro_LL(I)_912.5 OTU_571.4 3H_416.2 3H_342.4 OCS155_424.4 Pro_HL(II)_820.5 3H_404.6 NS9_639.8 3H_153.6 3H_265.3 Thioth_787.7 3H_427.5 SSHD Syn_1020.8 OTU_792.6 3H_430.2 3H_422.9 Sedimi/Punice_791.4 OTU_963 OTU_671.2 SAR86_700.7 3H_400.4 Syn_A.V_1053.2 Syn_A.I_1056.1 3H_408.9 3H_407.3 OM75_887.8 3H_365.7 3H_276.5 OM75_889.2 OTU_474.9 Pro_HL(I)_828.8 SAR86_471.9 Ciliate_590 Pro_HL(II)_933.7 Pseudo/OM60_937.8 Salinity Pro_HL(I)_825.8 Hellea_993.6 Roseob_987.8 3H_186.3 107 Case Study: Cyanobacteria, their grazers, and their viruses In looking at cyanobacterial OTUs and their correlations, we observed a number of potential virus-host relationships (Figure 4-5). Some cyanobacterial nodes were connected to a number of viral OTUs and others were also connected to protistan or other bacterial nodes – 67 nodes with 80 correlations. Many of the virus-bacteria correlations shown here were delayed with both cyanobacterial and T4-like virus nodes as the lagging partners. In silico analysis of g23 sequences from known cyanophage isolates matched one of associated T4-like viral TRFLP OTUs (3H_416.7) in this network to a Synechococcus WH8109 phage isolate (S-SSM7), which has an estimated fragment length of 416bp. This node, 3H_416.7 was positively correlated with the high-light Prochlorococcus OTU 820.5. Protistan nodes included three potential cyanobacterial grazers: 1) dinoflagellate or Lingulodinium-relative, 2) a ciliate, and 3) MAST/Stramenopile; all correlations to protists were delayed except to the MAST/Stramenopile. Two additional protistan nodes were unidentified but all five were positively correlated. Environmental nodes were also correlated to the cyanobacteria and represented: salinity, nitrite, and satellite-based estimates of chlorophyll a and primary production. Discussion By integrating our assessment of bacterial, protistan, and T4-like viral communities, we can complete a more holistic analysis of the complex microbe-microbe associations occurring in the surface ocean at SPOT, and the relationship to the changing environmental conditions as it relates to seasonal stratification. All three communities were significantly correlated with one another across all months based on overall Bray-Curtis community similarity, suggesting similar timing of community responses to either one another or a shift in the environment (Table 4-1, Figure 4-1). A positive correlation in similarity between bacterial and viral Bray-Curtis 108 similarities from month to month also suggests that if the bacterial community composition changes, then the viral community does as well (Table 4-1, Figure 4-1). T4-like myoviral and bacterial communities were more intricately linked on the OTU level (Table 4-2) than protistan and bacterial communities – by the number of significant LS correlations and concurrent shifts in monthly community similarity (Table 4-2, Figure 4-1). Variability in community structure from month to month was higher in the protists and bacteria, although this may reflect the resolution of our community fingerprinting assays. ARISA and 18S TRFLP both target the entire domain, while the g23-TRFLP focuses solely on a specific viral family. Bacterial community similarity was also negatively correlated with sea surface temperature, such that during periods of winter mixing (i.e. colder months), the communities were more similar from month to month. During the warmer, stratified months, bacterial communities were less similar from one month to the next. Thus, at the community level, variation of community structure was apparent but a more detailed inquiry would reveal the underlying individual relationships that comprise these net patterns and correlations between overall community structure of T4-like viruses, bacteria, and protists. Few environmental parameters were required to predict variability (albeit mostly a small percent) in the bacterial, T4-like myoviral, and protistan communities as observed by community fingerprinting methods. Variability within the T4-like family (28.1%) was more predictable than both protists (11.6%) and bacteria (6.5%), yet communities relied on different combinations of environmental variables. Bacterial variance was predicted by satellite-based estimates of chlorophyll a where as protists relied on both day length and bacterial abundance. Day length may indicate the type of autotrophs as opposed to prey availability for the heterotrophs or mixotrophs. T4-like virus community shifts were predicted by a more complex combination of day length, change in day length, salinity, and temperature. Cyanophages have been linked to a diel cycle and dependence upon light (Clokie & Mann 2006; Clokie et al. 2006; Bouvier & del 109 Giorgio 2007), and salinity and temperature may also reflect the need for a physiologically active host for a virus to replicate. It is not possible, to our knowledge, to determine whether the same variance (or OTUs) was predictable by both environmental parameters and the co-occurring microbial communities by these metrics. Inference of ecological interactions from LS correlation-based networks Although the networks shown here did not represent ‘true’ ecological networks, in that we presented correlations as opposed to direct observations, they identify common correlations from which the ecological implications of these potential relationships can be discussed. LSA increases the resolution with which to define potential relationships over time by permitting detection of time-lagged correlations (Ruan et al. 2006a). Although correlations are not equivalent to causal relationships, numerous statistically significant microbe-microbe interactions that potentially occur in nature were observed. Our inability to significantly predict bacterial community structure from T4-like myoviral or protistan community structure reflects the complex nature of the ecological relationships surrounding virus-bacteria-protist interactions. For example, subpopulations exhibiting opposite behaviors may be masked by general patterns. Example networks, taken from this three year-dataset, illustrate the breadth of interactions detected (Figure 4-2, Table 4-3). Both direct and indirect connections between individual OTUs or environmental parameters can be indicative (Miki & Jacquet 2008). Thus, assessment at the level of individual relationships between OTUs (i.e. by local similarity analysis) can clarify these underlying patterns and inform our interpretation of community-level patterns. Due to the response and turnover times of microbes, our interpretation of these long-term correlations as potential grazing or viral lysis activity would differ based on sampling frequency of the underlying data as the effects of both top-down controls could occur within days or weeks. 110 Given our monthly sampling frequency, the top-down relationships may appear as both positive or negative correlations, with or without time-delays (Figure 4-2) and the interpretation is reliant upon our identity of the OTUs and associated data from the field and laboratory experiments on the activities of these specific organisms. Grazing and viral lysis can be selective processes, but in different ways (Miki & Jacquet 2008). For example, grazing can be more influenced by the size rather than the taxonomy of the targeted prey – which is seen as a protistan OTU linked to a few bacterial OTUs (Figures 4-3A, 4-S2) or none at all. In fact, a number of small clusters of one protist and several bacteria were observed from SPOT (Figures 4-3A, 4-S2). Similarly, a virus’ host range may be indicated by the number of bacteria “host” OTUs correlated to a virus OTU such that more connections would suggest a broader host range. The increased number of correlations between virus-bacteria supports our observation on the correlation of community-level similarity for these two groups. Connectivity was much lower between protist-bacteria than in T4-like virus-bacteria interactions (Figures 4-3, 4-S2, 4-S3), lending further support to the idea that these top-down controls do not elicit a similar response by the bacterial community at large. Virus-bacterial interactions may define some niches – yet others may be bound by protistan or bacteria-bacteria interactions. Virus-bacteria interactions were not, however, evenly dispersed throughout the community, but rather specific bacterial OTUs or clusters of them were more highly connected than others. Investigation into the dominant bacteria that define the surface ocean bacterial community at SPOT over a ten-year period suggests that three SAR11 OTUs may be equally influenced by interactions with viruses and other bacteria, while other bacteria (like the Actinobacteria 435 and SAR11_S1_666.4) are more closely coupled with bacteria-bacteria interactions (Figure 4). Gross changes in viral abundance and community 111 structure may not have a universal impact on the bacterial community itself – but rather the effect may be partitioned. Case study: Connections between Cyanobacteria, co-occurring microbes, and the environment Prochlorococcus and Synechococcus, two genera of marine cyanobacteria, are dominant members of the ocean and integral to the marine ecosystems as key autotrophs in the microbial loop(Chisholm et al. 1988; 1992; Li 1994; Campbell et al. 1997; Liu et al. 1997; Partensky et al. 1999; DuRand et al. 2001; Giovannoni & Vergin 2012). It has been suggested that spatial, temporal, and vertical differences in distribution of Prochlorococcus and Synechococcus ecotypes are reflective of their capabilities or adaptation to nutrient utilization and differential mortality (Moore et al. 1998; Martiny et al. 2009; Malmstrom et al. 2010; Partensky & Garczarek 2010). Our association network revealed specific virus-cyanobacteria-protist interactions although not all cyanobacterial OTUs were equally linked to viruses or protistan grazers (Figure 4-5). Ciliates and nanoflagellates are thought to be the predominant grazers of cyanobacteria, yet new groups like new lineages of marine stramenopiles (MAST) continue to be discovered (Christaki et al. 1999; Worden & Binder 2003; Christaki et al. 2005; Frias-Lopez et al. 2009; Gong et al. 2012). Of note, although a lineage of marine stramenopiles was recently observed to prey on Synechococcus (Gong et al. 2012), the correlation we observed was from a Prochlorococcus high-light (II) OTU_820.5 to a MAST node. It is unknown, to our knowledge, whether MAST lineages also graze upon Prochlorococcus sp. Four other protistan OTUs were also correlated to bacterial OTUs, but their identities were not well-resolved. Additional sequencing efforts or analysis with high-throughput sequencing data may aid definition of these relationships. 112 Cyanophage-host systems are some of the best-characterized models from the ocean. Stable co-existing populations of viruses and their hosts have been documented, as well as seasonal variation from fieldwork (Waterbury & Valois 1993; Suttle 1994; Marston & Sallee 2003; Mühling et al. 2005; Sandaa & Larsen 2006; Wang et al. 2011) in addition the culture and isolate-based laboratory experiments (Sullivan et al. 2003; Lindell et al. 2005; Zinser et al. 2009). Growth rates of Synechococcus were positively affected by simultaneous lysis of co-occurring heterotrophic bacteria (Weinbauer et al. 2011) and the presence of heterotrophic bacteria in culture with Prochlorococcus either significantly improved or inhibited growth depending on the identity of the co-cultured taxa (Sher et al. 2011). Each cyanobacterial OTU was correlated to at least one other bacterial (non-cyanobacterial) OTU (Figure 4-5), many of which were known heterotrophs (i.e. Roseobacter (Moran & Miller 2007)). Only one viral OTU was identified based on the available sequence data from cultured isolates – which was OTU 3H_416.7 as cyanophage S-SSM7, isolated from Synechococcus WH8109. Although data on S-SSM7’s host range, to our knowledge, is unavailable (Sullivan et al. 2008), other cyanomyophages have broad host ranges and are capable of infecting both Prochlorococcus and Synechococcus isolates (Sullivan et al. 2003). These results, taken together, might suggest that cyanobacterial ecotypes are differentially responsive to these top-down controls, similar to their response to environmental forcings. Conclusions Ecological relationships in the ocean are dynamic, and the association networks presented here, likely represent stable relationships between microbes observed at in situ concentrations. The seasonal dynamics of the bacterial community provides a foundation for understanding the intricate relationships between bacteria and their co-occurring microbes (McManus & Fuhrman 1988; Fuhrman & Noble 1995; Simek et al. 2001; Sherr & Sherr 2002; Evans et al. 2003; 113 Weinbauer et al. 2003; Fuhrman et al. 2006; Weinbauer et al. 2007; Zhang et al. 2007; Baudoux & Veldhuis 2008; Longnecker et al. 2010; Steele et al. 2011; Staniewski et al. 2012). Seasonal variability in both T4-like viruses and protist communities had also been observed previously at SPOT (Countway & Caron 2006; Countway et al. 2010; Chow & Fuhrman 2012). Microbial relative abundances varied over time, as clusters of microbes rose and fell together or with one- month time delays. Our findings on connectivity of OTUs continue to suggest that microbial communities are complex, highly interconnected, and full of potential niches that warrant further investigation. Protistan-bacterial associations were far fewer than virus-bacteria associations, and illustrated unique patterns in connectivity that may reflect relative non-selective or size-selective grazing habits and complex but more narrow host ranges for viruses. Phytoplankton, specifically the cyanobacteria, are known to be regulated by both bottom-up and top-down controls – our association network continues to support those findings yet adds another layer of complexity to the bacterial response to changing microbial counterparts and environmental conditions. These networks can help to sort through which factors are of increased importance to different OTUs, both cyanobacteria are others, and the dependency of those relationships over time. Although shorter time-scale resolution is still needed to understand the intricacies of these relationships, LSA can identify which pairings of OTUs may be of increased ecological significance over long- term temporal or spatial scales and merit future targeted research. 114 Chapter Four References Allesina, S. & Pascual, M. (2007). Network structure, predator–prey modules, and stability in large food webs. Theor Ecol, 1, 55–64. Assenov, Y., Ramírez, F., Schelhorn, S., Lengauer, T. & Albrecht, M. (2008). Computing topological parameters of biological networks. Bioinformatics, 24, 282–284. Azam, F., Fenchel, T., Field, J., Gray, J., Meyer-Reil, L. & Thingstad, F. (1983). The ecological role of water-column microbes in the sea. Mar. Ecol. Prog. Ser., 10, 257–263. Baudoux, A. & Veldhuis, M. (2008). Estimates of virus-vs. grazing induced mortality of picophytoplankton in the North Sea during summer. Aquat. Microb. Ecol. Bouvier, T. & del Giorgio, P.A. (2007). Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol, 9, 287–297. Bouvy, M., Bettarel, Y., Bouvier, C., Domaizon, I., Jacquet, S., Le Floc'h, E., et al. (2011). Trophic interactions between viruses, bacteria and nanoflagellates under various nutrient conditions and simulated climate change. Environ Microbiol, 13, 1842–1857. Braak, Ter, C.J.F. & Schaffers, A.P. (2004). Co-correspondence analysis: a new ordination method to relate two community compositions. Ecology, 85, 834–846. Bratbak, G., Thingstad, F. & Heldal, M. (1994). Viruses and the microbial loop. Microb Ecol, 28, 209–221. Brown, M.V., Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2005). Coupling 16S-ITS rDNA clone libraries and automated ribosomal intergenic spacer analysis to show marine microbial diversity: development and application to a time series. Environ Microbiol, 7, 1466–1479. Campbell, L., Liu, H., Nolla, H.A. & Vaulot, D. (1997). Annual variability of phytoplankton and bacteria in the subtropical North Pacific Ocean at Station ALOHA during the 1991-1994 ENSO event. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 44, 167–192. Chaffron, S., Rehrauer, H., Pernthaler, J. & Mering, von, C. (2010). A global network of coexisting microbes from environmental and whole-genome sequence data. Genome Research, 20, 947–959. Chisholm, S.W., Frankel, S.L., Goericke, R., Olson, R.J., Palenik, B., Waterbury, J.B., et al. (1992). Prochlorococcus marinus nov. gen. nov. sp.: an oxyphototrophic marine prokaryote containing divinyl chlorophyll a and b. Arch. Microbiol., 157, 297–300. Chisholm, S.W., Olson, R.J., Zettler, E.R., Goericke, R., Waterbury, J.B. & Welschmeyer, N.A. (1988). A novel free-living prochlorophyte abundant in the oceanic euphotic zone. Nature, 334, 340–343. Chow, C.-E.T. & Fuhrman, J.A. (2012). Seasonality and monthly dynamics of marine myovirus communities. Environ Microbiol. In press. 115 Christaki, U., Jacquet, S., Dolan, J., Vaulot, D. & Rassoulzadegan, F. (1999). Growth and Grazing on Prochlorococcus and Synechococcus by Two Marine Ciliates. Limnol Oceanogr., 44, 52–61. Christaki, U., Vazquez-Dominguez, E., Courties, C. & Lebaron, P. (2005). Grazing impact of different heterotrophic nanoflagellates on eukaryotic (Ostreococcus tauri ) and prokaryotic picoautotrophs (Prochlorococcus and Synechococcus). Environ Microbiol, 7, 1200–1210. Cline, M.S., Smoot, M., Cerami, E., Kuchinsky, A., Landys, N., Workman, C., et al. (2007). Integration of biological networks and gene expression data using Cytoscape. Nat Protoc, 2, 2366–2382. Clokie, M.R.J. & Mann, N.H. (2006). Marine cyanophages and light. Environ Microbiol, 8, 2074–2082. Clokie, M.R.J., Millard, A.D. & Mann, N.H. (2010). T4 genes in the marine ecosystem: studies of the T4-like cyanophages and their role in marine ecology. Virol J, 7, 291. Clokie, M.R.J., Millard, A.D., Mehta, J.Y. & Mann, N.H. (2006). Virus isolation studies suggest short-term variations in abundance in natural cyanophage populations of the Indian Ocean. J. Mar. Biol. Ass., 86, 499–505. Comeau, A.M. & Krisch, H.M. (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, 1321–1332. Corno, G. & Jürgens, K. (2008). Structural and functional patterns of bacterial communities in response to protist predation along an experimental productivity gradient. Environ Microbiol, 10, 2857–2871. Countway, P.D. & Caron, D.A. (2006). Abundance and distribution of Ostreococcus sp. in the San Pedro Channel, California, as revealed by quantitative PCR. Appl. Environ. Microbiol., 72, 2496–2506. Countway, P.D., Gast, R.J., Savai, P. & Caron, D.A. (2005). Protistan diversity estimates based on 18S rDNA from seawater incubations in the western north Atlantic1. J. Eukaryot. Microbiol., 52, 95–106. Countway, P.D., Vigil, P.D., Schnetzer, A., Moorthi, S.D. & Caron, D.A. (2010). Seasonal analysis of protistan community structure and diversity at the USC Microbial Observatory (San Pedro Channel, North Pacific Ocean). Limnol Oceanogr., 55, 2381–2396. Dunne, J.A. (2002). Food-web structure and network theory: The role of connectance and size. Proc. Natl. Acad. Sci. U.S.A., 99, 12917–12922. DuRand, M.D., Olson, R.J. & Chisholm, S.W. (2001). Phytoplankton population dynamics at the Bermuda Atlantic Time-series station in the Sargasso Sea. Deep-Sea Res. (2 Top. Stud. Oceanogr.), 48, 1983–2003. 116 Eiler, A., Heinrich, F. & Bertilsson, S. (2012). Coherent dynamics and association networks among lake bacterioplankton taxa. ISME J, 6, 330–342. Evans, C., Archer, S.D., Jacquet, S. & Wilson, W.H. (2003). Direct estimates of the contribution of viral lysis and microzooplankton grazing to the decline of a Micromonas spp. population. Aquat. Microb. Ecol., 30, 207–219. Filee, J. (2005). Marine T4-type bacteriophages, a ubiquitous component of the dark matter of the biosphere. Proc. Natl. Acad. Sci. U.S.A., 102, 12471–12476. Fisher, M.M. & Triplett, E.W. (1999). Automated approach for ribosomal intergenic spacer analysis of microbial diversity and its application to freshwater bacterial communities. Appl. Environ. Microbiol., 65, 4630–4636. Frias-Lopez, J., Thompson, A., Waldbauer, J. & Chisholm, S.W. (2009). Use of stable isotope- labelled cells to identify active grazers of picocyanobacteria in ocean surface waters. Environ Microbiol, 11, 512–525. Fuhrman, J.A. (1999). Marine viruses and their biogeochemical and ecological effects. Nature, 399, 541–548. Fuhrman, J.A. & Noble, R.T. (1995). Viruses and protists cause similar bacterial mortality in coastal seawater. Limnol Oceanogr., 1236–1242. Fuhrman, J.A. & Steele, J.A. (2008). Community structure of marine bacterioplankton: patterns, networks, and relationships to function. Aquat. Microb. Ecol., 53, 69–81. Fuhrman, J.A. & Suttle, C.A. (1993). Viruses in marine planktonic systems. Oceanography, 6, 51–63. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Gasol, J.M., Pedrós-Alió, C. & Vaqué, D. (2002). Regulation of bacterial assemblages in oligotrophic plankton systems: results from experimental and empirical approaches. Antonie Leeuwenhoek, 81, 435–452. Gilbert, J.A., Steele, J.A., Caporaso, J.G., Steinbrück, L., Reeder, J., Temperton, B., et al. (2012). Defining seasonal marine microbial community dynamics. ISME J, 6, 298–308. Giovannoni, S.J. & Vergin, K.L. (2012). Seasonality in Ocean Microbial Communities. Science, 335, 671–676. Gong, G., Chiang, K. & Worden, A. (2012). Distribution patterns and phylogeny of marine stramenopiles (MAST) in the North Pacific Ocean. Appl. Environ. Microbiol. Li, W.K.W. (1994). Primary production of prochlorophytes, cyanobacteria, and eucaryotic ultraphytoplankton: measurements from flow cytometric sorting. Limnol Oceanogr., 39, 169–175. 117 Lindell, D., Jaffe, J.D., Johnson, Z.I., Church, G.M. & Chisholm, S.W. (2005). Photosynthesis genes in marine viruses yield proteins during host infection. Nature, 438, 86–89. Liu, H., Nolla, H.A. & Campbell, L. (1997). Prochlorococcus growth rate and contribution to primary production in the equatorial and subtropical North Pacific Ocean. Aquat. Microb. Ecol., 12, 39–47. Longnecker, K., Wilson, M., Sherr, E. & Sherr, B.F. (2010). Effect of top-down control on cell- specific activity and diversity of active marine bacterioplankton. Aquat. Microb. Ecol., 58, 153–165. Malmstrom, R.R., Coe, A., Kettler, G.C., Martiny, A.C., Frias-Lopez, J., Zinser, E.R., et al. (2010). Temporal dynamics of Prochlorococcus ecotypes in the Atlantic and Pacific oceans. ISME J, 4, 1252–1264. Marston, M.F. & Sallee, J.L. (2003). Genetic diversity and temporal variation in the cyanophage community infecting marine Synechococcus species in Rhode Island's coastal waters. Appl. Environ. Microbiol., 69, 4639–4647. Martiny, A.C., Tai, A.P.K., Veneziano, D., Primeau, F. & Chisholm, S.W. (2009). Taxonomic resolution, ecotypes and the biogeography of Prochlorococcus. Environ Microbiol, 11, 823– 832. McManus, G.B. & Fuhrman, J.A. (1988). Control of marine bacterioplankton populations: Measurement and significance of grazing. Hydrobiologia, 159, 51–62. Middelboe, M. (2000). Bacterial Growth Rate and Marine Virus-Host Dynamics. Microb Ecol, 40, 114–124. Miki, T. & Jacquet, S. (2008). Complex interactions in the microbial world: underexplored key links between viruses, bacteria and protozoan grazers in aquatic environments. Aquat. Microb. Ecol., 51, 195–208. Mitchell, R.J., Hester, A.J., Campbell, C.D., Chapman, S.J., Cameron, C.M., Hewison, R.L., et al. (2011). Explaining the variation in the soil microbial community: do vegetation composition and soil chemistry explain the same or different parts of the microbial variation? Plant Soil, 351, 355–362. Moebus, K. (1996). Marine bacteriophage reproduction under nutrient-limited growth of host bacteria. I. Investigations with six phage-host systems. Mar. Ecol. Prog. Ser., 14, 1–12. Montoya, J.M., Pimm, S.L. & Solé, R.V. (2006). Ecological networks and their fragility. Nature, 442, 259–264. Moore, L.R., Rocap, G. & Chisholm, S.W. (1998). Physiology and molecular phylogeny of coexisting Prochlorococcus ecotypes. Nature, 393, 464–467. Mühling, M., Fuller, N.J., Millard, A., Somerfield, P.J., Marie, D., Wilson, W.H., et al. (2005). Genetic diversity of marine Synechococcus and co-occurring cyanophage communities: evidence for viral control of phytoplankton. Environ Microbiol, 7, 499–508. 118 Müller, J., Stadler, J., Jarzabek-Müller, A., Hacker, H., Braak, ter, C. & Brandl, R. (2011). The Predictability of Phytophagous Insect Communities: Host Specialists as Habitat Specialists. PLoS ONE, 6, e25986. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O'Hara, R.B., et al. (2011). vegan: Community Ecology Package (R package version 2.0.2). R package version 2.0.2. Olesen, J.M., Stefanescu, C. & Traveset, A. (2011). Strong, Long-Term Temporal Dynamics of an Ecological Network. PLoS ONE, 6, e26455. Ory, P., Hartmann, H.J., Jude, F., Dupuy, C., Del Amo, Y., Catala, P., et al. (2010). Pelagic food web patterns: do they modulate virus and nanoflagellate effects on picoplankton during the phytoplankton spring bloom? Environ Microbiol, 12, 2755–2772. Partensky, F. & Garczarek, L. (2010). Prochlorococcus: Advantages and Limits of Minimalism. Annu. Rev. Marine. Sci., 2, 305–331. Partensky, F., Hess, W.R. & Vaulot, D. (1999). Prochlorococcus, a marine photosynthetic prokaryote of global significance. Microbiol. Mol. Biol. Rev., 63, 106–127. Ruan, Q., Dutta, D., Schwalbach, M.S., Steele, J.A., Fuhrman, J.A. & Sun, F. (2006a). Local similarity analysis reveals unique associations among marine bacterioplankton species and environmental factors. Bioinformatics, 22, 2532–2538. Ruan, Q., Steele, J.A., Schwalbach, M.S., Fuhrman, J.A. & Sun, F. (2006b). A dynamic programming algorithm for binning microbial community profiles. Bioinformatics, 22, 1508– 1514. Sandaa, R.-A. & Larsen, A. (2006). Seasonal Variations in Virus-Host Populations in Norwegian Coastal Waters: Focusing on the Cyanophage Community Infecting Marine Synechococcus spp. Appl. Environ. Microbiol., 72, 4610–4618. Sandaa, R.-A., Gómez-Consarnau, L., Pinhassi, J., Riemann, L., Malits, A., Weinbauer, M.G., et al. (2009). Viral control of bacterial biodiversity - evidence from a nutrient-enriched marine mesocosm experiment. Environ Microbiol, 11, 2585–2597. Shannon, P. (2003). Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Research, 13, 2498–2504. Sher, D., Thompson, J.W., Kashtan, N., Croal, L. & Chisholm, S.W. (2011). Response of Prochlorococcus ecotypes to co-culture with diverse marine bacteria. ISME J, 5, 1125–1132. Sherr, E. & Sherr, B.F. (1988). Role of microbes in pelagic food webs: a revised concept. Limnol Oceanogr., 33, 1225–1227. Sherr, E. & Sherr, B.F. (2002). Significance of predation by protists in aquatic microbial food webs. Antonie Leeuwenhoek, 81, 293–308. 119 Simek, K., Pernthaler, J., Weinbauer, M., Hornák, K., Dolan, J.R., Nedoma, J., et al. (2001). Changes in Bacterial Community Composition and Dynamics and Viral Mortality Rates Associated with Enhanced Flagellate Grazing in a Mesoeutrophic Reservoir. Appl. Environ. Microbiol., 67, 2723–2733. Simpson, G.L. (2009). cocorresp: Co-correspondence analysis ordination methods (R package version 0.1-9). R package version 0.1-9. Smoot, M.E., Ono, K., Ruscheinski, J., Wang, P.-L. & Ideker, T. (2011). Cytoscape 2.8: new features for data integration and network visualization. Bioinformatics, 27, 431–432. Sole, R.V. & Montoya, M. (2001). Complexity and fragility in ecological networks. Proc R Soc Lond B, 268, 2039–2045. Staniewski, M.A., Short, C.M. & Short, S.M. (2012). Contrasting Community versus Population- Based Estimates of Grazing and Virus-Induced Mortality of Phytoplankton. Microb Ecol. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. Sullivan, M.B., Coleman, M.L., Quinlivan, V., Rosenkrantz, J.E., DeFrancesco, A.S., Tan, G., et al. (2008). Portal protein diversity and phage ecology. Environ Microbiol, 10, 2810–2823. Sullivan, M.B., Waterbury, J.B. & Chisholm, S.W. (2003). Cyanophages infecting the oceanic cyanobacterium Prochlorococcus. Nature, 424, 1047–1051. Suttle, C.A. (1994). Dynamics and Distribution of Cyanophages and Their Effect on Marine Synechococcus spp. Appl. Environ. Microbiol. Thurber, R.V. (2009). Current insights into phage biodiversity and biogeography. Curr. Opin. Microbiol, 12, 582–587. Vigil, P., Countway, P., Rose, J., Lonsdale, D., Gobler, C. & Caron, D. (2009). Rapid shifts in dominant taxa among microbial eukaryotes in estuarine ecosystems. Aquat. Microb. Ecol., 54, 83–100. Wang, K., Wommack, K.E. & Chen, F. (2011). Abundance and Distribution of Synechococcus spp. and Cyanophages in the Chesapeake Bay. Appl. Environ. Microbiol., 77, 7459–7468. Waterbury, J.B. & Valois, F.W. (1993). Resistance to co-occurring phages enables marine Synechococcus communities to coexist with cyanophages abundant in seawater. Appl. Environ. Microbiol., 59, 3393–3399. Weinbauer, M.G., Bonilla-Findji, O., Chan, A.M., Dolan, J.R., Short, S.M., Simek, K., et al. (2011). Synechococcus growth in the ocean may depend on the lysis of heterotrophic bacteria. J. Plankton Res., 33, 1465–1476. Weinbauer, M.G., Christaki, U., Nedoma, J. & Simek, K. (2003). Comparing the effects of resource enrichment and grazing on viral production in a meso-eutrophic reservoir. Aquat. Microb. Ecol., 31, 137–144. 120 Weinbauer, M.G., Hornák, K., Jezbera, J., Nedoma, J., Dolan, J.R. & Šimek, K. (2007). Synergistic and antagonistic effects of viral lysis and protistan grazing on bacterial biomass, production and diversity. Environ Microbiol, 9, 777–788. Wommack, K.E. & Colwell, R.R. (2000). Virioplankton: viruses in aquatic ecosystems. Microbiol. Mol. Biol. Rev., 64, 69–114. Worden, A.Z. & Binder, B.J. (2003). Application of dilution experiments for measuring growth and mortality rates among Prochlorococcus and Synechococcus populations in oligotrophic environments. Aquat. Microb. Ecol., 30, 159–174. Xia, L.C., Steele, J.A., Cram, J.A., Cardon, Z.G., Simmons, S.L., Vallino, J.J., et al. (2011). Extended local similarity analysis (eLSA) of microbial community and other time series data with replicates. BMC Syst. Biol., 5, S15. Zhang, R., Weinbauer, M.G. & Qian, P.-Y. (2007). Viruses and flagellates sustain apparent richness and reduce biomass accumulation of bacterioplankton in coastal marine waters. Environ Microbiol, 9, 3008–3018. Zinser, E.R., Lindell, D., Johnson, Z.I., Futschik, M.E., Steglich, C., Coleman, M.L., et al. (2009). Choreography of the Transcriptome, Photophysiology, and Cell Cycle of a Minimal Photoautotroph, Prochlorococcus. PLoS ONE, 4, e5135. 121 Chapter Five: Metagenomic depth profile of viral communities highlights depth-dependent variation Abstract The importance of microbes in the ocean is undeniable; viruses are not only tremendously abundant (~1 million in a drop of seawater) but also key players in global marine food webs. Changes in viral diversity over space and time are thought to have significant influence on microbial processes, yet we still have limited information on how virus communities vary - especially over the water column. Metagenomics (high-throughput sequencing of DNA from a community) does not rely on cultivation and results have shown that a wide diversity of marine viruses encode many unexpected and interesting potential functions. We described viral metagenomes collected at five depths (5m, deep chlorophyll maximum (20-30m), 150m, 500m, and 890m) in March, May and August 2009 at the San Pedro Ocean Time Series station (SPOT), located off Southern California. Our results suggest that significant differences in the taxonomic distribution of viruses occur between the mesopelagic and the euphotic zone. Myoviridae and Siphoviridae traded dominance, such that the Myoviridae were a larger fraction in the euphotic zone while Siphoviridae were more dominant below the euphotic zone – an observation consistent with previous views of the role of lysogeny in the ocean. Functionally, a key difference was the presence of photosynthesis genes in the photic zone library, presumably from cyanophages. These four metagenomic libraries represent a first look at the variability in viral community composition and function through the water column – from which we identified fundamental distinctions in taxonomic composition and functional annotation that separate the euphotic zone from the mesopelagic. 122 Introduction In the past 30 years, the fields of microbial ecology and microbial oceanography have been transformed with the development of molecular biology tools. Marine virology, in particular, has directly benefited from significant sequencing efforts – from single genes, to genomics and metagenomics (e.g. Rohwer & Thurber 2009; Jacquet et al. 2010; Clokie et al. 2011; Breitbart 2012). A survey of viromes throughout the ocean basins and of viral sequences from the Sorcerer II Global Ocean Survey (GOS) confirmed the prevalence of viruses as reservoirs of genetic diversity and varying global distributions by family (Angly et al. 2006; Williamson et al. 2008b). Additional insight into the role and function of viruses in the marine environment stemmed from the discovery of auxiliary metabolic genes, viral-encoded functional genes with host origins, and has shed light on the processes that occur when viruses infect marine bacteria (as reviewed in Sharon et al. 2011; Breitbart 2012). Single viral gene surveys combined with other molecular methods had previously increased the detected diversity of virus groups present in the environment, which was later confirmed during the explosion of viral metagenomes in the last ten years. To date, viral metagenomes have been sampled from the ocean, sediment, soil, deep seafloor, freshwater lakes, hot springs and hypersaline environments (Breitbart 2002; Breitbart et al. 2003; 2004; Angly et al. 2006; Bench et al. 2007; Fierer et al. 2007; Schoenfeld et al. 2008; Dinsdale et al. 2008a; Williamson et al. 2008a; Dinsdale et al. 2008b; 2008b; Helton & Wommack 2009; Rodriguez-Brito et al. 2010; Steward & Preston 2011; Roux et al. 2012). Viral metagenomics is still quite exploratory, and to date have suggested that viruses have a surprisingly high potential to not only influence community composition but also a variety of metabolic processes in the sea. Viral metagenomics, as a field, has challenged our view of the extent of viral influence - leaving much to be discovered, especially below the photic zone. 123 In one of the few microbial metagenomic examples of a depth profile, an abundance of viral sequences were uncovered in the cellular fraction (DeLong et al. 2006). It was suggested that the viral sequences were likely from microbes under active infection and they consisted of cyanophage-like sequences in the photic zone. These cyanophage-like sequences included viral- encoded genes for photosynthesis, phosphate scavenging, and other processes that are consistent with the oligotrophic environment at station ALOHA and their widespread distribution (Sharon et al. 2007). More recently, a single unamplified metagenomic sample from the mesopelagic in Monterey Bay, representing a single depth and point in time, suggested similarity in overall genetic composition between this sample and prior viral metagenomes largely from coastal or surface waters (Steward & Preston 2011). Prior research has suggested a stratification of viral community composition and activity below the photic zone that mirrors changes within the host bacterial communities (Hara et al. 1996). Viruses in the dark ocean can be significant sources of bacterial mortality (Danovaro et al. 2008). At SPOT, in particular, viral communities at depth were dominated by different viral genomes than in the surface water, based on pulsed field gel electrophoresis although genetically- related phages were isolated from throughout the water column (Jiang et al. 2003). The relative abundance and significance of lytic versus lysogenic phages in the dark ocean is still under debate. Lysogeny has been associated with waters of low host abundance and less productivity while lytic infection dominated waters with high bacterial productivity and abundance (Weinbauer et al. 2003; Long et al. 2007) and so their distribution may depend most on the productivity and abundance of susceptible hosts in the environment. However, the dominant viral activity in the deep-sea benthos was reported to be lytic (Danovaro et al. 2008), although lysogeny reportedly dominated in waters surrounding hydrothermal vents (Williamson et al. 124 2008a). Overall, the virus-host relationship and thus the viral communities themselves may differ between the euphotic zone and the dark ocean (De Corte et al. 2010). We present here what is, to our knowledge, the first metagenomic depth profile of viral communities with discrete sampling depths using virus-centric sampling methods and bioinformatic analyses. Specifically, we analyzed four metagenomic libraries covering the entire water column at the San Pedro Ocean Time-series station (SPOT) and the resulting viromes were annotated with VIROME: Viral Informatics Resource for Metagenome Exploration (Bhavsar et al. 2009). We discuss the relative distributions of viral families and functional genes using SEED subsystems, a curated collection of sequence annotation across genomes based on related functional roles. Lastly, we compared the diversity of our metagenomic libraries using Shotgun UniFrac, a new method to assess and compare taxonomic composition of metagenomes based on sequence similarity to known viral genomes (Caporaso et al. 2011); however, database limitations and biases limited the applicability of the Shotgun UniFrac results. Materials and Methods Sample Collection and Virus Concentration Seawater was collected from the following regions of the water column: surface mixed layer (5m and deep chlorophyll maximum at 20, 23, and 30m), mid-water column (150 and 500m), and ~ 10m above the seafloor (890m) in March, May and August 2009 at the San Pedro Ocean Time-series (SPOT). SPOT is located midway between Los Angeles and Santa Catalina Island (33˚33’N, 118˚24’W). Samples were collected by Niskin bottles each month as follows: ~10L for 5m, ~20L at the deep chlorophyll maximum depth and ~30L each for the 150m, 500m, and 890m depths. Thus, approximately 90L seawater was collected for each metagenome (see Table 5-1). Each depth was processed independently each month. Environmental data was 125 collected as part of the USC Microbial Observatory and Wrigley Time-series Station programs (Fuhrman et al. 2006; Countway et al. 2010; Steele et al. 2011). Samples were serially filtered by positive pressure at 10-15psi through a 142mm Type A/E glass fiber filter (1.0µm nominal pore size, Pall) and 0.22µm Durapore (GVWP, Millipore) to remove cellular material. Filtrate was collected and concentrated for viral particles: (1) viral particles from March 2009 were further concentrated by 30kDa tangential flow filtration (Millipore) followed by precipitation with polyethylene glycol (Thurber et al. 2009); (2) samples collected in May and August 2009 were concentrated by iron chloride flocculation (John et al. 2010). By TFF+PEG concentration, only ~10 8 (150, 500, 890m) to 10 9 (5m, DCM) viral particles were recovered per sample of 30 (150, 500, or 890m), 20 (DCM), or 10 (5m) liters of seawater; in contrast, ~10 10 (150, 500, 890m) to ~10 11 (5m, DCM) viral particles were recovered by iron chloride flocculation per sample of 30 (150, 500, or 890m), 20 (DCM), or 10 (5m) liters of seawater. Molecular Methods DNA extraction. Viral particles were further purified from cells and free DNA by cesium chloride purification. DNA was subsequently extracted using a CTAB, formamide and phenol- chloroform purification, according to the protocol for DNA viruses (Thurber et al. 2009). All samples were processed individually for each depth and month up to this point. DNA was pooled to increase starting material for library preparation and sequence, to try to minimize potential biases introduced during the subsequent amplification steps, and also reduce any temporal biases. DNA extracts were pooled for the 5m and chlorophyll maximum depths across all three months to create a single “euphotic zone” sample. DNA for the 150m, 500m, and 890m were pooled across sampling months to build a library for each depth. 126 Library Preparation and Sequencing. Pooled DNA (for each metagenomic library) was ultrasonically sheared to 300-800bp by Covaris, concentrated, end-repaired, cleaned, ligated to adaptor (Linker A primer), cleaned, size-fractionated by agarose gel electrophoresis to select for 300-800bp, cleaned, amplified by PCR, reconditioned by PCR, cleaned and quantified (Henn et al. 2010). Modifications were: (1) PCR amplification with LS Takara HS instead of PFU Turbo and (2) addition of a reconditioning PCR following initial amplification (Duhaime et al. in press). Samples G4009 (5DCM), G4010 (890m), G4011 (500m) and G4012 (150m) were sequenced at the Broad Institute using 454 Titanium sequencing chemistry. Sequences are available in CAMERA as: 5DCM, CAM_SMPL_000990; 150m, CAM_SMPL_000961; 500m, CAM_SMPL_000962; 890m, CAM_SMPL_001014; sequences are available in VIROME as SPOT Virome 2009 5mDCM (SPA), SPOT Virome 2009 150m (SPB), SPOT Virome 2009 500m (SPC), SPOT Virome 2009 890m (SPD). Metagenomic Sequence Analysis VIROME. Viral Informatics Resource for Metagenome Exploration (VIROME) is a new web- based bioinformatics platform optimized for viral metagenomic sequence data. Briefly, sequences were trimmed of linker primers and searched for any ribosomal or transfer RNA gene sequences; reads with significant similarity (e-value <0.001) were removed from further analysis. For each remaining read, open-reading frames (ORFs) were predicted by MetaGene and each ORF was assigned to one of the following categories, listed in order of priority assignment, based on hits with e-value <0.001: 1) possible functional proteins in UniRef 100 Plus (UniRef100P), 2) unassigned proteins – no function listed in UniRef100P, 3) top-hit viral protein – top hit to an environmental sequence of viral origin from MetaGenomes On-line database (MGOL), 4) viral only environmental protein – all hits to MGOL were of viral origin, 5) top-hit microbial (non- viral) environmental protein – top hit was of microbial (non-viral) origin from MGOL, 6) 127 microbial only environmental protein – all hits were of microbial (non-viral) origin from MGOL, and lastly 7) true ORFan – BLAST analysis against over 50 million proteins resulted in no significant matches to either UniRef100P or MGOL. MGOL was constructed from predicted environmental proteins from publicly available metagenomic libraries. ORFs that were homologous to proteins in UniRef100 were also assigned a potential taxonomic origin, based on the taxonomy of the best BLASTP hit; consensus taxonomy from the common terms was assembled for multiple hits. VIROME is available at: http://virome.dbi.udel.edu. Shotgun UniFrac. Sequence reads were trimmed for quality (VIROME) and linker primer sequences. For analysis by sequence reads, reads were excluded that did not meet specific requirements: minimum length of 200 bp, maximum of 1 ambiguous base, maximum of 6 homopolymers. For ORF-based analysis, ORFs were downloaded from VIROME. Reads or ORFs were clustered using Shotgun UniFrac, as described, with the script: pick_reference_otus_by_otu_table.py (Caporaso et al. 2011). The Phage Proteomic Tree with 651 phages represented was used as the reference database (Rohwer & Edwards 2002; Edwards & Rohwer 2005; Caporaso et al. 2011). Adjustments to default QIIME parameters included: 1) blast as the OTU picking algorithm, 2) minimum 50% alignment identity and 3) minimum 90% sequence similarity. Results Seawater samples were collected over a six-month period to reduce any potential temporal biases and increase total DNA yield. Four discrete metagenomic libraries were constructed representing: 1) “euphotic zone” – 0-5m and chlorophyll maximum depths, 2) 150m – upper mesopelagic, 3) 500 - mid-water column, and 4) 890m -just above sediment (Table 5-1). 128 Table 5-1. Overview of Metagenome Sample Libraries. Sample ID Depth (m) Date Volume filtered (L) Virus Conc. Method Number of reads Avg read length (bp) Called ORFs (partial) Upper mixed layer 5 3/11/09 10 TFF, PEG 194,220 389 251,379 13 3/11/09 20 5 5/13/09 10 FeCl 30 5/13/09 20 5 8/19/09 10 22 8/19/09 20 150m 150 3/11/09 30 TFF, PEG 201,002 331 232,447 150 5/13/09 30 FeCl 150 8/19/09 30 500m 500 3/11/09 30 TFF, PEG 128,849 390 159,678 500 5/13/09 30 FeCl 500 8/11/09 30 890m 890 3/11/09 30 TFF, PEG 168,217 398 210,191 890 5/13/09 30 FeCl 890 8/11/09 30 Hydrographic features of the water column at SPOT during the months sampled were consistent with ten-year averages observed from August 2000 – January 2011 (Figure 5-1). In general, salinity increased with depth while temperature decreased (Figure 5-1, F and G). Average seawater temperature was 17ºC and 13.7ºC in the surface waters (0-5m) and DCM (20- 30m), respectively; temperature ranged from 12-20ºC for the photic zone when sampled. The 150m, 500m, and 890m depths were sampled at 9.5, 6.6, and 5.2ºC, respectively, which equaled the ten-year average temperature. The 150m samples were collected considerably below the euphotic zone, as seen with chlorophyll a distribution (Figure 5-1, G). Bacterial production and abundance of both cells and viruses were much higher in the euphotic zone than the deeper waters (Figure 5-1, B-E). Bacterial abundance in the euphotic zone was on average almost 6-fold higher than the deeper samples; bacterial production was 1-2 orders of magnitude higher in the euphotic zone. Viral abundances were at least one order of magnitude higher in the euphotic zone as well (Figure 5-1, E). Over the ten-year period, a nitrite maximum was typically observed near the chlorophyll maximum and both nitrate and phosphate concentrations increased with depth (Figure 5-1. H-J). 129 Figure 5-1. Environmental characteristics of the water column at SPOT. Panels indicate variation in depth for: salinity (a), bacterial production by tritiated thymidine (b) and leucine (c) incorporation, bacterial abundance (d), viral abundance (e), temperature (f), chlorophyll a (g), nitrite (h), nitrate (i), and phosphate (j). Values for the DCM are plotted at the average depth of 29m. Black X’s and solid lines indicate the ten-year average values. Error bars are standard error of the mean for the ten-year averages only. Open shapes with dotted lines indicate the sampled months: circles, March; triangles, May; squares, August. C AB E D H FG J I 130 Through the VIROME pipeline, potential ORFs were determined for each sequence read based on the presence of either a start or stop codon; these sequences were translated and then searched against publicly available metagenomes. As these sequences reads are too short to represent complete genes, coding sequence that lacked a start or stop codon within the sequenced region may be overlooked and ORFs may have been overestimated or truncated too early. Sequencing results suggested that our viromes were almost exclusively viral in origin based on limited quantification of 16S rDNA by quantitative PCR from DNA extracts and lack of reads detected with homology to rRNA (only 4 of 692,288 sequence reads) or tRNA (n = 1126). For all libraries, half of all sequences did not match to either the UniRef 100 Plus or other publicly available metagenomic samples (422,595/853,695 predicted ORFs = 49.5%) and 37% were observed in other metagenomes only. Only 13.5% of predicted ORFs had a significant hit with an e-value <0.001 to any genome in UniRef 100 Plus (n=115,412); of these, 41,463 ORFs had no assigned function. The functional annotations and taxonomic origins for each ORF were summarized for each library where an annotation was present, excluding sequences that had no match in any database (Figure 5-2 and 5-3). The taxonomic distribution of ORF’s covered all domains of life but the majority of ORFs were homologous to bacterial sequences (Figure 5-2). Virus ORFS composed ~20% below the euphotic zone and 36.5% in the pooled 5m and DCM library, an increase of around 15% in the euphotic zone as compared to the other depths. Eukarya, Archaea, and other unclassified sequences comprised the remaining portion, <8%, of each library. When considering just these viral ORFS, the taxonomic composition at the family level differed between the euphotic zone and the three deep depths (Figure 5-3). Specifically, the photic zone virome was composed of 56.2% Myoviridae in contrast to the 34.8%-37.2% in 150, 500, 890m. Podoviridae were a consistent proportion across all depths – from 19.9-24.2%. Thus, the difference in Myoviridae was balanced by an increase in Siphoviridae at depth, which 131 increased from 12% in the photic zone library to 28.9%, 30.3%, and 33.3% in 150m, 500m, and 890m, respectively. Other viral families that comprised the remaining 7.2-11.7% of the annotated viral community included (in descending order of representation): Phycodnaviridae, Mimiviridae, Iridoviridae, Poxviridae, Reoviridae, Tectiviridae, Retroviridae, Adenoviridae, Ampullaviridae, Asfarviridae, Baculoviridae, Bunyaviridae, and Parvoviridae. Figure 5-2. Taxonomic identification of ORFs by domain per library. Each ring represents a metagenomic library: the pooled 5m and DCM library is shown in the center, and is surrounded by 150m, 500m, and 890m (moving outwards). Each domain is represented by a different color as labeled. Unclassified Eukaryota Archaea Bacteria Viruses 890m 500m 150m 5DCM 132 Figure 5-3. Taxonomic distribution of phage ORFs differs with depth. Each pie charts reflect an individual metagenome: A) 5DCM, euphotic zone, B) 150m, C) 500m, and D) 890m. The wedge labels indicated the number and proportion of ORFs assigned to each viral family from those which had a database hit. The functional diversity of these viral metagenomic libraries covered many of the functional SEED subsystems (Figure 5-4). The most abundant subsystems across all libraries included: DNA metabolism, Nucleosides and Nucleotides, Cell Wall and Capsule, Clustering- based Subsystems and Protein Metabolism. Few instances of the subsystems for Dormancy and Sporulation; Fatty Acids, Lipids, and Isoprenoids; Membrane Transport; Metabolism of Aromatics, and other miscellaneous genes were observed. The main clear distinction between the 133 libraries by presence/absence was the observation of photosynthesis-related genes, primarily psbA and psbD, in the euphotic zone library and their absence in the remaining three metagenomes. Additionally, the relative proportion of genes related to “Cofactors, Vitamins, Prosthetic Groups, and Pigments” was approximately double in the photic zone as compared to the other metagenomes. In all metagenomes, this group was represented by only two genes, 1) thymidylate synthase and 2) GTP cyclohydrolase I; the former converts uracil to thymidine while the latter hydrolyzes GTP. Otherwise, the distribution of SEED subsystems across depths and their relative proportions are remarkably similar. We utilized Shotgun UniFrac, a new method for comparing metagenomic datasets developed from comparative tools for single-gene sequence analysis, as an alternative approach to assess differences between the viral metagenomes (Caporaso et al. 2011). The relative differences in representation of tree branches in the Phage Proteomic Tree was used to differentiate the metagenomic libraries as in the original UniFrac or weighted UniFrac metrics for ribosomal rRNA gene diversity (Lozupone & Knight 2005; Lozupone et al. 2006; Hamady et al. 2009). Each sequence read was compared to the genomic content of 651 individual virus genomes included in the Phage Proteomic Tree, built from sequence similarity of overall genomic content where each branch tip represents a single viral genome (Rohwer & Edwards 2002; Edwards & Rohwer 2005). Our application of this new UniFrac-based method resulted in few significant sequence hits to any known viruses (Table 2), far fewer than the minimum of 200 matches required in the original comparison of host-associated and other environmental metagenomes. Most of our hits were to known cyanophages (Pro_xx or Syn_xx) and from the 5DCM metagenome. Utilization of ORFs instead of sequence reads resulted in an increase in absolute number of hits, but the relative proportion of sequences with hits remained low at 0.06%. 134 Figure 5-4. SEED Subsystem Functional Gene Distribution by metagenomic library. Each colored bar represents a metagenomic library, and the x-axis indicates the percentage of ORFs assigned to each SEED subsystem of all assigned ORFs. 0.00 0.05 0.10 0.15 0.20 0.25 0.30 890m 500m 150m 5DCM Virulence Sulfur Metabolism Stress Response Secondary Metabolism RNA Metabolism Respiration Regulation and Cell Signaling Protein Metabolism Potassium Metabolism Plant-Prokaryote Doe Project Photosynthesis Phosphorus Metabolism Phages, Prophages, Transposable Elements Nucleosides and Nucleotides Nitrogen Metabolism Motility and Chemotaxis Miscellaneous Metabolism of Aromatic Compounds Membrane Transport Fatty Acids, Lipids, and Isoprenoids Dormancy and Sporulation DNA Metabolism Cofactors, Vitamins, Prosthetic Groups, Pigments Clustering-Based Subsystems Cell Wall and Capsule Cell Division and Cell Cycle Carbohydrates Amino Acids and Derivatives 135 Table 5-2. Phylogenetic Assignment by Shotgun UniFrac. The first column denotes the viral genome that each read or ORF was assigned to. Columns 2-5 list sequence hits to viral genomes based on the sequence reads, while columns 6-9 indicate sequence hits based on the called ORFs. All non-zero numbers were highlighted in bold. The last four rows summarize: the number of sequence matches, the number of unassigned sequences, total sequences or ORFs in each library, and the percent of reads or ORFs assigned by shotgun UniFrac. A) Sequence Reads B) Predicted ORFs Viral Genome 5DCM 150m 500m 890m 5DCM 150m 500m 890m Bac_SPM2 0 0 0 0 3 0 0 0 Bac_phBC6A52 0 0 0 0 0 0 0 1 Bac_phiYeO312 0 0 0 0 2 0 0 0 Clo_phiSM101 0 0 0 0 0 0 1 0 Cya_P60 0 0 0 0 2 0 0 0 Cya_PSSM2 12 0 0 0 19 0 0 0 Cya_PSSP7 2 0 0 0 4 0 0 0 Ent_EcoDS1 0 0 0 0 1 0 0 0 Ent_RB51 0 0 0 0 0 0 0 1 Lac_SL4 0 0 0 0 0 1 0 0 Pro_PHM1 2 0 0 0 2 1 0 0 Pro_PHM2 2 0 0 0 6 0 0 0 Pro_PRSM4 1 0 0 0 3 0 0 0 Pro_PSSM4 13 0 0 0 5 0 0 0 Pro_PSSM7 0 0 0 0 8 1 0 1 Pro_Syn33 24 0 0 0 22 0 0 0 Syn_SRSM4 11 0 0 0 15 0 0 0 Syn_SSM1 0 0 0 0 5 0 0 0 Syn_SSM2 12 0 1 0 34 1 0 0 Syn_SSSM5 0 0 0 0 5 0 0 0 Syn_SSSM7 0 0 0 0 3 0 0 0 Syn_Syn5 0 0 0 0 1 0 0 0 Syn_Syn9 0 0 0 0 6 0 0 0 Vib_VP93 0 0 0 0 0 1 0 0 Hits 79 0 1 0 146 5 1 3 No Match 122721 118162 80892 108732 251230 232441 159676 210188 Total 122800 118162 80893 108732 251379 232446 159677 210191 % of library 0.064% Na 0.001% na 0.058% 0.002% 0.001% 0.001% 136 Discussion The data presented here represent the first metagenomic look at a depth profile of free virus particles in the open ocean (Table 5-1). The majority of viral metagenomic analysis to date has focused on the surface ocean, extreme environments or from analyses of viral genes found in the metagenomes of the microbial fraction. Our sampling site is in some way typical of an open- ocean site and also has features characteristic of an oxygen minimum zone at 500 and 890m (Figure 5-1). The San Pedro Basin has a sill, leading to limited circulation in the deep waters and persistent low oxygen concentrations in the basin except for occasional flushing events (Berelson 1991; Countway et al. 2010; Collins et al. 2011). Water in the San Pedro basin has a residence time of 12+/-6 months between flushing events that brings in cooler, oxygen- and nitrate-rich waters in place of warmer, depleted bottom water (Berelson 1991). These libraries can provide a broad insight into viral community composition and function within disparate regions of the water column. Although at first glance most ORFs were identified as bacterial sequences, these bacterial hits do include matches to completed bacterial genomes that could be prophage or other viral-like genes encoded in the genome (Figure 5-2). Less than 10% of the ORFS matched Eukarya or Archaea, potentially reflecting a bias in the available databases to bacteria and bacteriophage for marine environments. Fewer viral hits in the deeper depths may be related to the limited information from isolates, genomes or metagenomes that originated from below the photic zone. By taxonomic annotation, we observed an increased proportion of siphoviruses below the photic zone, coupled with a reduction of the proportion of myoviruses (Figure 5-3). While not all siphoviruses are lysogenic, these observations were consistent with prior evidence suggesting that lytic phages are less abundant below the euphotic zone (e.g. De Corte et al. 2010). Our results do not suggest that each viral life strategy exists at the exclusion of the other; it does however reflect 137 that the abundance, and potential activity of lytic or temperate phages may relate to the productivity of the environment or potential host availability (presence of a susceptible host at high enough abundance to sustain consistent infection). In the less productive, deeper depths, temperate phages may take advantage of this alternative survival strategy if it is less likely to encounter a new host. Functionally, the distribution of SEED subsystems varied little between the four viral metagenomes. One notable exception was the lack of photosynthesis-related genes in the 150m, 500m, and 890m libraries. It is presumed that an abundance of cyanophage in the photic zone library would account for these sequences as core photosystem genes have been found to be common in cyanophages. Both psbA and psbD were discovered through whole genome sequencing of Synechococcus and Prochlorococcus phage isolates (Chen and Lu, 2002; Paul and Sullivan, 2005; Sullivan et al., 2005; Sullivan et al., 2006). It has been shown that these genes are transcribed during the infection process to aid viral replication and proteins are produced after the virus terminates host regulation of host-encoded genes (Lindell et al. 2005; Clokie et al. 2006; Lindell et al. 2007). Sequencing-based studies has also shown that significant diversity exists between cyanophages and that this diversity is reflected in the ecological adaptations and species of the susceptible host cyanobacteria (Sullivan et al. 2006; Sharon et al. 2007; 2011). Secondly, thymidylate synthase has been observed in a number of the cyanophage genomes sequences to date including cyanosiphovirues and cyanomyoviruses (Chen & Lu 2002; Weigele et al. 2007) as well as in other viral metagenomes (Schoenfeld et al. 2008) while GTP cyclohydrolase was also previously observed in the Sargasso Sea virome (Angly et al. 2006) suggesting that both genes are quite common in marine viruses. As these two genes were observed in all metagenomes, it remains to be determined if there is a functional or ecological impact of this differential representation or merely a byproduct of a database bias towards cyanophage-like sequences. 138 Comparative metagenomics, as a field is still in development especially for viral samples where there are particularly significant database biases and limitations. Bacterial genomes outnumber viral genomes; host- or human-associated virus genomes also outnumber marine phages, although the situation is improving with new sequencing efforts, increased research interest, and development of new comparative metrics. It was therefore not surprising that our tests revealed few hits overall by Shotgun UniFrac with the Phage Proteomic Tree that included known human pathogenic viruses and other non-marine viruses in addition to the environmental ones. Prior successful comparison with Shotgun UniFrac for marine metagenomes was based on libraries with increased sequencing effort or longer sequence read length, and required a minimum number of 200 sequence hits for successful comparisons (Caporaso et al. 2011). Most of our sequence hits were to cyanophages and confirmed the presence of cyanophage-like sequences in the photic zone; most phage sequences from the “dark” ocean were uncharacterized (Table 5-2). The remaining sequences hits included a vibriophage-like sequence from VP93; although likely not identical, several vibriophages have previously been isolated from SPOT below the euphotic zone (Jiang et al. 2003). It is probable that many of the phages within and below the photic zone infect heterotrophic bacteria, and would include vibriophage. Vibrio sp. are common and found at low abundances throughout the ocean so it is reasonable to conclude that their viruses would be similarly well-distributed. Other non-cyanophage open ocean and mesopelagic viruses are rather poorly represented in the currently available database for comparison. Conclusions Through continued investigation into the unexpected diversity and functions of marine viral communities by metagenomics and other methods, we can continue to untangle the role, 139 activity and importance of viruses in the sea. In this metagenomic depth profile, we highlighted differences in functional and taxonomic distribution of viruses between the euphotic zone and the mesopelagic. The euphotic zone metagenome likely include cyanophages and other marine viruses, as was observed by the presence of photosynthesis genes. The ‘dark’ depths included sequences that were less likely to match sequences previously detected in viruses, marine or otherwise. Our results are consistent with increased abundance of temperate phage and lysogeny in the less productive, deeper zones of the water column. This first look at depth-dependent variation will provide a comparative sample for viral communities elsewhere and to better understand viral function and activity in the light, productive euphotic zone versus the less productive dark ocean. 140 Chapter Five References Angly, F.E., Felts, B., Breitbart, M., Salamon, P., Edwards, R.A., Carlson, C.A., et al. (2006). The marine viromes of four oceanic regions. Plos Biol, 4, e368. Bench, S.R., Hanson, T.E., Williamson, K.E., Ghosh, D., Radosovich, M., Wang, K., et al. (2007). Metagenomic Characterization of Chesapeake Bay Virioplankton. Appl. Environ. Microbiol., 73, 7629–7641. Berelson, W. (1991). The flushing of two deep-sea basins, southern California borderland. Limnol Oceanogr. Bhavsar, J., Polson, S.W. & Wommack, K.E. (2009). VIROME: An informatics resource for viral metagenome exploration. Breitbart, M. (2002). Genomic analysis of uncultured marine viral communities. Proc. Natl. Acad. Sci. U.S.A., 99, 14250–14255. Breitbart, M. (2012). Marine Viruses: Truth or Dare. Annu. Rev. Marine. Sci., 4, 425–448. Breitbart, M., Felts, B., Kelley, S., Mahaffy, J.M., Nulton, J., Salamon, P., et al. (2004). Diversity and population structure of a near-shore marine-sediment viral community. Proc R Soc Lond B, 271, 565–574. Breitbart, M., Hewson, I., Felts, B., Mahaffy, J.M., Nulton, J., Salamon, P., et al. (2003). Metagenomic analyses of an uncultured viral community from human feces. J. Bacteriol., 185, 6220–6223. Caporaso, J., Knight, R. & al, E. (2011). Host-associated and free-living phage communities differ profoundly in phylogenetic composition. PLoS ONE. Clokie, M.R.J., Millard, A.D., Letarov, A.V. & Heaphy, S. (2011). Phages in nature. Bacteriophage, 1, 31–45. Clokie, M.R.J., Shan, J., Bailey, S., Jia, Y., Krisch, H.M., West, S., et al. (2006). Transcription of a “photosynthetic” T4-type phage during infection of a marine cyanobacterium. Environ Microbiol, 8, 827–835. Collins, L.E., Berelson, W., Hammond, D.E., Knapp, A., Schwartz, R. & Capone, D. (2011). Particle fluxes in San Pedro Basin, California A four-year record of sedimentation and physical forcing. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 58, 898–914. Countway, P.D., Vigil, P.D., Schnetzer, A., Moorthi, S.D. & Caron, D.A. (2010). Seasonal analysis of protistan community structure and diversity at the USC Microbial Observatory (San Pedro Channel, North Pacific Ocean). Limnol Oceanogr., 55, 2381–2396. Danovaro, R., Dell'Anno, A., Corinaldesi, C., Magagnini, M., Noble, R., Tamburini, C., et al. (2008). Major viral impact on the functioning of benthic deep-sea ecosystems. Nature, 454, 1084–1087. 141 De Corte, D., Sintes, E., Winter, C., Yokokawa, T., Reinthaler, T. & Herndl, G.J. (2010). Links between viral and prokaryotic communities throughout the water column in the (sub)tropical Atlantic Ocean. ISME J, 4, 1431–1442. DeLong, E.F., Preston, C.M., Mincer, T., Rich, V., Hallam, S.J., Frigaard, N.-U., et al. (2006). Community genomics among stratified microbial assemblages in the ocean's interior. Science, 311, 496–503. Dinsdale, E.A., Edwards, R.A., Hall, D., Angly, F.E., Breitbart, M., Brulc, J.M., et al. (2008a). Functional metagenomic profiling of nine biomes. Nature, 452, 629–632. Dinsdale, E.A., Pantos, O., Smriga, S., Edwards, R.A., Angly, F.E., Wegley, L., et al. (2008b). Microbial Ecology of Four Coral Atolls in the Northern Line Islands. PLoS ONE, 3, e1584. Duhaime, M.B., Deng, L., Poulos, B. & Sullivan, M.B. (n.d.). Linker amplification for quantitative metagenomics of wild viruses and other ultra-low concentration DNA samples, in prep, 1–46. Edwards, R.A. & Rohwer, F. (2005). Viral metagenomics. Nat Rev Micro, 3, 504–510. Fierer, N., Breitbart, M., Nulton, J., Salamon, P., Lozupone, C., Jones, R., et al. (2007). Metagenomic and Small-Subunit rRNA Analyses Reveal the Genetic Diversity of Bacteria, Archaea, Fungi, and Viruses in Soil. Appl. Environ. Microbiol., 73, 7059–7066. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Hamady, M., Lozupone, C. & Knight, R. (2009). Fast UniFrac: facilitating high-throughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J, 4, 17–27. Hara, S., Koike, I., Terauchi, K. & Kamiya, H. (1996). Abundance of viruses in deep oceanic waters. Mar. Ecol. Prog. Ser., 145, 269–277. Helton, R.R. & Wommack, K.E. (2009). Seasonal Dynamics and Metagenomic Characterization of Estuarine Viriobenthos Assemblages by Randomly Amplified Polymorphic DNA PCR. Appl. Environ. Microbiol., 75, 2259–2265. Henn, M.R., Sullivan, M.B., Stange-Thomann, N., Osburne, M.S., Berlin, A.M., Kelly, L., et al. (2010). Analysis of High-Throughput Sequencing and Annotation Strategies for Phage Genomes. PLoS ONE, 5, e9083. Jacquet, S., Miki, T., Noble, R., Peduzzi, P. & Wilhelm, S. (2010). Viruses in aquatic ecosystems: important advancements of the last 20 years and prospects for the future in the field of microbial oceanography and limnology. Advances in Oceanography and Limnology, 1, 97– 141. Jiang, S., Fu, W., Chu, W. & Fuhrman, J.A. (2003). The Vertical Distribution and Diversity of Marine Bacteriophage at a Station off Southern California. Microb Ecol, 45, 399–410. 142 John, S.G., Mendez, C.B., Deng, L., Poulos, B., Kauffman, A.K.M., Kern, S., et al. (2010). A simple and efficient method for concentration of ocean viruses by chemical flocculation. Environ Microbiol Rep, Virus concentration by flocculation with iron, 3, 195–202. Lindell, D., Jaffe, J.D., Coleman, M.L., Futschik, M.E., Axmann, I.M., Rector, T., et al. (2007). Genome-wide expression dynamics of a marine virus and host reveal features of co- evolution. Nature, 449, 83–86. Lindell, D., Jaffe, J.D., Johnson, Z.I., Church, G.M. & Chisholm, S.W. (2005). Photosynthesis genes in marine viruses yield proteins during host infection. Nature, 438, 86–89. Long, A., McDaniel, L.D., Mobberley, J. & Paul, J.H. (2007). Comparison of lysogeny (prophage induction) in heterotrophic bacterial and Synechococcus populations in the Gulf of Mexico and Mississippi river plume. ISME J, 2, 132–144. Lozupone, C. & Knight, R. (2005). UniFrac: a new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol., 71, 8228–8235. Lozupone, C., Hamady, M. & Knight, R. (2006). UniFrac–an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinformatics, 7, 371. Rodriguez-Brito, B., Li, L., Wegley, L., Furlan, M., Angly, F.E., Breitbart, M., et al. (2010). Viral and microbial community dynamics in four aquatic environments. ISME J, 4, 739–751. Rohwer, F. & Edwards, R.A. (2002). The Phage Proteomic Tree: a genome-based taxonomy for phage. J. Bacteriol., 184, 4529–4535. Rohwer, F. & Thurber, R.V. (2009). Viruses manipulate the marine environment. Nature, 459, 207–212. Roux, S., Enault, F., Robin, A., Ravet, V., Personnic, S., Theil, S., et al. (2012). Assessing the Diversity and Specificity of Two Freshwater Viral Communities through Metagenomics. PLoS ONE, 7, e33641. Schoenfeld, T., Patterson, M., Richardson, P.M., Wommack, K.E., Young, M. & Mead, D. (2008). Assembly of Viral Metagenomes from Yellowstone Hot Springs. Appl. Environ. Microbiol., 74, 4164–4174. Sharon, I., Battchikova, N., Aro, E.-M., Giglione, C., Meinnel, T., Glaser, F., et al. (2011). Comparative metagenomics of microbial traits within oceanic viral communities. ISME J, 5, 1178–1190. Sharon, I., Tzahor, S., Williamson, S.J., Shmoish, M., Man-Aharonovich, D., Rusch, D.B., et al. (2007). Viral photosynthetic reaction center genes and transcripts in the marine environment. ISME J, 1, 492–501. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. 143 Steward, G.F. & Preston, C.M. (2011). Analysis of a viral metagenomic library from 200 m depth in Monterey Bay, California constructed by direct shotgun cloning. Virol J, 8, 287. Sullivan, M.B., Lindell, D., Lee, J.A., Thompson, L.R., Bielawski, J.P. & Chisholm, S.W. (2006). Prevalence and Evolution of Core Photosystem II Genes in Marine Cyanobacterial Viruses and Their Hosts. Plos Biol, 4, e234. Thurber, R., Haynes, M., Breitbart, M. & Wegley, L. (2009). Laboratory procedures to generate viral metagenomes : Abstract : Nature Protocols. Nat Protoc. Weigele, P.R., Pope, W.H., Pedulla, M.L., Houtz, J.M., Smith, A.L., Conway, J.F., et al. (2007). Genomic and structural analysis of Syn9, a cyanophage infecting marine Prochlorococcus and Synechococcus. Environ Microbiol, 9, 1675–1695. Weinbauer, M.G., Brettar, I. & Höfle, M.G. (2003). Lysogeny and virus-induced mortality of bacterioplankton in surface, deep, and anoxic marine waters. Limnol Oceanogr., 48, 1457– 1465. Williamson, S.J., Cary, S.C., Williamson, K.E., Helton, R.R., Bench, S.R., Winget, D., et al. (2008a). Lysogenic virus–host interactions predominate at deep-sea diffuse-flow hydrothermal vents. ISME J, 2, 1112–1121. Williamson, S.J., Rusch, D.B., Yooseph, S., Halpern, A.L., Heidelberg, K.B., Glass, J.I., et al. (2008b). The Sorcerer II Global Ocean Sampling Expedition: Metagenomic Characterization of Viruses within Aquatic Microbial Samples. PLoS ONE, 3, e1456. 144 Chapter Six: Synthesis In the last thirty years, scientific research on the ecology and role of viruses in the ocean has developed alongside the methodology to assess these previously unseen microbes (Azam et al. 1983; Fuhrman 1999; Azam & Worden 2004; Edwards & Rohwer 2005; Fenchel 2008; Rohwer & Thurber 2009; Breitbart 2012). After overwhelming evidence that viruses were both ubiquitous and highly abundant in the ocean (Bergh et al. 1989; Proctor & Fuhrman 1990; Børsheim 1993), initial surveys of morphology and abundance were soon followed by the development of ecological theories based on results from traditional field experiments and sequencing efforts to explain (a) the dynamics of individual viruses - kill the winner (Fuhrman & Suttle 1993; Thingstad & Lignell 1997; Thingstad 2000; Wommack & Colwell 2000) and the Cheshire Cat or Red Queen (Frada et al. 2008; Rohwer & Thurber 2009; Bidle & Vardi 2011), (b) the distribution of viral types by abundance as in the Bank model (Breitbart & Rohwer 2005; Bellec et al. 2010; Campbell et al. 2011; Chow & Fuhrman 2012), and (c) how virus-host interactions could influence biogeochemical cycles (Bratbak et al. 1994; Fuhrman 1999; Wilhelm & Suttle 1999; Suttle 2005; Sullivan et al. 2010; Breitbart 2012). However, an understanding of how changes in one microbial community may affect another on the scale of individual(s) or populations is still an unknown piece that is essential to our understanding of microbes in the sea. The presented work addressed this knowledge gap by describing shifts in bacterial and viral community structure at a long-term time-series, concurrently with an analysis of individual correlations between viruses, bacteria, and protists. Microbial community dynamics were investigated through molecular fingerprinting methods (Chapters 2-4), metagenomics (Chapters 5), and ecological networks (Chapters 2 and 4) at the USC Microbial Observatory at the San Pedro Ocean Time-series station over a ten-year period. T4-like virus communities were investigated by a newly developed terminal restriction fragment length polymorphism (TRFLP) 145 method for g23, a gene that encodes the virus’ major capsid protein (Chapters 3), and those results were coupled to community fingerprints of bacterial and protistan communities to describe potential microbe-microbe interactions (Chapter 4). This work, and related Microbial Observatory research, has built a microbial dataset that can be compared with other long-term time-series sites – such as the Western English Channel, Station ALOHA at the Hawaii Ocean Time-series (HOT), the Bermuda Atlantic Time Series (BATS). By examining microbes as individual OTUs and as a community, new insights into the ecology of viruses and bacteria in the ocean were developed in terms of community membership, temporal variation, persistence and connectivity. Community Membership, Temporal Variability, and Persistence Knowledge of the constituents, how many there are and how often they appear are critical first steps to understand how a community works (Fuhrman 2009; Shade & Handelsman 2012). In this case, organisms were defined at the level of operational taxonomic units and observed over ten years at SPOT for the bacterial communities in the surface ocean and deep chlorophyll maximum depths (Chapters 2 and 4) and over three years for the T4-like viral communities (Chapters 3 and 4). A few hundred unique OTUs were observed in total by g23-TRFLP, with around 100 OTUs detected each month, Additional OTUs may be present that were not detected by our fingerprinting methods, yet the overall results presented here are consistent with prior observations of a high diversity of microbes in the ocean. Metagenomic characterization of viral communities over spatial scales identified key differences in taxonomic identities within the communities, despite pooling samples from a few different months (Chapter 5). Depth-dependent relationships were interrogated for bacteria by comparing surface (0-5m) and deep chlorophyll maximum (DCM) communities over ten years and for viruses over the entire water column (5- 890m) at a single time point (Chapter 2 and 5). Bacterial community membership was generally 146 consistent within the euphotic zone, but viral communities exhibited different taxonomic distributions at depth compared to the euphotic zone. Viral and bacterial communities were found to be seasonally variable, and significantly predictable from year to year in the surface ocean (Chapters 2 and 3). Shifts in bacterial community composition were correlated to protistan and viral communities, confirming prior assumptions that these communities are intricately linked via the microbial loop (Chapter 4). Each community contained OTUs that peaked repeatedly in spring-summer and others in fall- winter. Some of the most persistent T4-like viral OTUs were steady year-round while others exhibited seasonal patterns in variation. Thus, bacterial and viral communities were not entirely transient, but rather followed specified patterns at SPOT. SPOT and BATS have showed clear signs of recurrent and seasonal bacterial communities in the euphotic zone in contrast to less pronounced trends at HOT (Fuhrman et al. 2006; Carlson et al. 2009; Eiler et al. 2011; Giovannoni & Vergin 2012). The extent of mixing and temporal variability in bacterial structure would suggest that SPOT exhibits trends “in- between” HOT and BATS (Chapter 2). Annual recurrence of community structure was demonstrated by discriminant function analysis and trends in Bray-Curtis similarities - all of which suggested that communities one month or one year apart were more similar than communities of opposite seasons (i.e. summer versus winter) for the surface ocean. Community similarity between the surface and DCM bacterial communities at SPOT was lowest during periods of late summer stratification, as shown by a negative correlation to the temperature difference between the two depths (Chapter 2). Seasonality in viral community structure is not inherently a new observation, but few previous studies extended beyond a year or demonstrated consistent annual recurrence of viral community composition (Chow & Fuhrman 2012). Although viral abundances were not seasonally variable at SPOT as has been observed at BATS 147 (Parsons et al. 2012), the observed seasonality in viral community structure at SPOT would suggest that viral community composition at BATS might also have a seasonal pattern. Phylogeny The coherence of bacterial communities in the surface ocean and its underlying deep chlorophyll maximum layer revealed a core microbiome of common OTUs between both depths with similar taxonomic composition. Less than 10% of all OTUs were not observed at least five times in the other sampling depth. The distribution of taxonomic identities of these OTUs, at the class level, was also consistent between depths – even when separate by frequency of detection into persistent, intermittent, and ephemeral OTUs (Chapter 2). Future high-resolution sequencing will provide the necessary information to investigate these trends in more detail, as has been completed in the Western English Channel (Gilbert et al. 2009; 2012) and expected from the Earth Microbiome Project. Consistency in viral phylogeny over time was not specifically addressed as we primarily investigated one group, the T4-like viruses, for our monthly sampling (Chapter 3) however metagenomic characterization of viral communities at different depths revealed spatial differences in taxonomic distribution of the Order Caudovirales, which includes the T4-like viruses (Chapter 5). Myoviruses (often lytic) were more dominant in the surface ocean compared to siphoviruses (which includes the lysogenic viruses), which dominated below the photic zone. Previous sequencing efforts had suggested significant similarity in viral communities throughout the ocean, although these samples were limited in coverage of discrete depths or time points (Angly et al. 2006; Steward & Preston 2011). Frequent sampling and metagenomic characterization of a human-controlled aquatic environments, similarly suggested that higher- level taxonomic composition was quite stable and that individuals more likely varied at the strain 148 level over time (Rodriguez-Brito et al. 2010). Additional research is required to determine if these depth-based trends are consistent across ocean basins, specific to oxygen minimum zones, or unique to SPOT. Connectivity Just having similar community members does not necessarily indicate that the same processes or interactions always occur; rather the strength or type of interaction may vary depending on their context (environmental or activity of other microbial community members). From aquatic mesocosm experiments, microbe-microbe relationships were shown to be flexible in relation to an organism’s flexible nutrient stoichiometry, level of predatory controls, or some combination of the two for both grazing and viral effects (Weinbauer et al. 2003; Sandaa 2008; Thingstad & Cuevas 2010; Bouvy et al. 2011). Connections and co-occurrence patterns within the microbial community and their transitions on a monthly scale were investigated by Bray-Curtis similarity patterns over time, local similarity analysis and development of association networks (Chapters 2, 3 and 4), which confirmed that the microbial community at SPOT is quite complex and inter-related. In essence, this work begins to describe how all connections between all individuals (or OTUs) contribute to the broader community-level relationships and builds from prior research from four-year surveys of bacterial and archaeal communities at SPOT (Fuhrman & Steele 2008; Beman et al. 2011; Steele et al. 2011). Through these networks, it is possible to begin deconstructing the individual relationships that form the general link between microbial groups in the microbial loop. A network of virus- bacteria-protist interactions (for 0-5m only) and a pair of networks (one for each sampling depth – 0-5m and deep chlorophyll maximum) were used to interrogate these complex relationships over a three and ten-year period, respectively (Chapters 2 and 4). The ability to define specific 149 microbe-microbe interactions with traditional ecological principles and our understanding of how to interpret these networks is still developing. Mostly positive relationships, potentially indicative of mutualistic and predatory relationships were observed, although significant negative correlations, reflective of competition, were present as well. Freshwater ecosystem research identified similarly stable clusters of bacteria that co-occurred overtime (Eiler et al. 2012). Clusters of inter-connected bacteria from these networks may similarly represent “tribes” or modules that often co-occur in the ocean. The same degree of connectivity within the protists themselves was not seen as in an earlier study, these results may be a key difference between the DCM (Steele et al. 2011) and the surface ocean (Chapter 2). Overall, the overabundance of positive correlations and modules in both bacteria-only networks by depths and the virus- bacteria-protist network suggests overall stability in the microbial realm over time at this location. Many relationships were identified within the association networks between viral, protistan and bacterial OTUs. Some relationships such as those between a ciliate and a Prochlorococcus OTU (Chapter 4) might be expected from past culture-based studies and field observations (Christaki et al. 1999); however, other relationships may be new – such as a MAST- lineage OTU linked to a Prochlorococcus OTU (Chapter 4). Relatively few studies have been completed on understanding modularity or descriptions of host ranges for broad groups of viruses – leading to the question of whether genetically-related viruses primarily infect phylogenetically- related hosts (e.g. Sullivan et al. 2003; Flores et al. 2011). The correlations observed in these networks can be used to similarly identify and describe potential virus-host relationships – perhaps to more fully develop models of a virus’ host range, infectivity, or stable partners. These findings for both protists and viruses can lead to a better characterization of protist-bacterial relationships, following subsequent laboratory or field-based observations. 150 Concluding Remarks and Future Directions Following the advice of the virologist, Salvador Luria, this research sought to identify and describe consistent patterns observed in nature – specifically in relation to temporal variability of T4-like viral and bacterial communities. All evidence from community similarity estimates, predictability by individual OTU variation, and presence of ~100 OTU’s every month for years suggests that there is a remarkably stable and diverse microbial population in both the surface ocean and the deep chlorophyll maximum layers. For so many OTUs to co-exist repeatedly, it would seem as if there is an abundance of beneficial or cooperative direct or indirect interactions between microbes – as seen in the high number of inter-connected modules in the association networks. At the very least, common OTUs do not appear to be extremely antagonistic to one another over long time-scale, under the observed conditions. From this understanding of how individual members of the microbial community relate to one another, it would suggest that the communities overall would be remarkably resilient to future changes in the marine environment despite potentially more drastic shifts in relative abundances on seasonal or other shorter-term scales. However, continuous cultures of viruses and their hosts suggest that co- evolution of both may limit overall net effects on nutrient cycling (Lennon et al. 2007; Lennon & Martiny 2008; Marston et al. 2012). This assessment of individual connections between OTUs in the microbial realm are consistent with current knowledge of microbial interactions, yet provides additional detail into the complexity and strength of relationships at the individual level that together sum up to a community-wide consensus interaction. Our understanding of what organisms are present and their (relative) abundances in the ocean has drastically improved over the last several years, and with this knowledge it has become equally relevant to define not just which OTUs are cooperative or antagonistic but also the tenacity of those interactions. This research addressed the complex nature of interactions in the 151 environment – and the networks can be used to identify altogether unexpected or surprisingly strong correlations between microbes for further analysis. Defining the controlling factors of these microbe-microbe relationships will ultimately aid parameterization of global models of nutrient and energy flow in the ocean and continue to improve estimates on viral contributions to nutrient cycling. Future directions for this research include answering the following questions: • How does host physiological status or environment influence the efficacy of viral infection and the lytic cycle? o Are virus-host relationships and their limitations (or advantages) consistent across time or location? • Do siphoviruses consistently dominate in the mesopelagic or other oxygen minimum zones? o What genomic or other physiological adaptations contribute to their increased abundance below the photic zone? 152 Chapter Six References Angly, F.E., Felts, B., Breitbart, M., Salamon, P., Edwards, R.A., Carlson, C.A., et al. (2006). The marine viromes of four oceanic regions. Plos Biol, 4, e368. Azam, F. & Worden, A.Z. (2004). Microbes, molecules, and marine ecosystems. Science, 303, 1622–1624. Azam, F., Fenchel, T., Field, J., Gray, J., Meyer-Reil, L. & Thingstad, F. (1983). The ecological role of water-column microbes in the sea. Mar. Ecol. Prog. Ser., 10, 257–263. Bellec, L., Grimsley, N., Derelle, E., Moreau, H. & Desdevises, Y. (2010). Abundance, spatial distribution and genetic diversity of Ostreococcus tauriviruses in two different environments. Environ Microbiol Rep, 2, 313–321. Beman, J.M., Steele, J.A. & Fuhrman, J.A. (2011). Co-occurrence patterns for abundant marine archaeal and bacterial lineages in the deep chlorophyll maximum of coastal California. ISME J, 5, 1077–1085. Bergh, Ø., Børsheim, K.Y., Bratbak, G. & Heldal, M. (1989). High abundance of viruses found in aquatic environments. Nature, 340, 467–468. Bidle, K.D. & Vardi, A. (2011). A chemical arms race at sea mediates algal host-virus interactions. Curr. Opin. Microbiol, 14, 449–457. Bouvy, M., Bettarel, Y., Bouvier, C., Domaizon, I., Jacquet, S., Le Floc'h, E., et al. (2011). Trophic interactions between viruses, bacteria and nanoflagellates under various nutrient conditions and simulated climate change. Environ Microbiol, 13, 1842–1857. Bratbak, G., Thingstad, F. & Heldal, M. (1994). Viruses and the microbial loop. Microb Ecol, 28, 209–221. Breitbart, M. (2012). Marine Viruses: Truth or Dare. Annu. Rev. Marine. Sci., 4, 425–448. Breitbart, M. & Rohwer, F. (2005). Here a virus, there a virus, everywhere the same virus? Trends Microbiol., 13, 278–284. Børsheim, K.Y. (1993). Native marine bacteriophages. FEMS Microbiol Ecol, 102, 141–159. Campbell, B.J., Yu, L., Heidelberg, J.F. & Kirchman, D.L. (2011). Activity of abundant and rare bacteria in a coastal ocean. Proc. Natl. Acad. Sci. U.S.A., 108, 12776. Carlson, C.A., Morris, R., Parsons, R., Treusch, A.H., Giovannoni, S.J. & Vergin, K.L. (2009). Seasonal dynamics of SAR11 populations in the euphotic and mesopelagic zones of the northwestern Sargasso Sea. ISME J, 3, 283–295. Chow, C.-E.T. & Fuhrman, J.A. (2012). Seasonality and monthly dynamics of marine myovirus communities. Environ Microbiol. In press. 153 Christaki, U., Jacquet, S., Dolan, J., Vaulot, D. & Rassoulzadegan, F. (1999). Growth and Grazing on Prochlorococcus and Synechococcus by Two Marine Ciliates. Limnol Oceanogr., 44, 52–61. Edwards, R.A. & Rohwer, F. (2005). Viral metagenomics. Nat Rev Micro, 3, 504–510. Eiler, A., Hayakawa, D.H. & Rappé, M.S. (2011). Non-random assembly of bacterioplankton communities in the subtropical north pacific ocean. Front. Microbio., 2, 1–10. Eiler, A., Heinrich, F. & Bertilsson, S. (2012). Coherent dynamics and association networks among lake bacterioplankton taxa. ISME J, 6, 330–342. Fenchel, T. (2008). The microbial loop – 25 years later. J. Exp. Mar. Biol. Ecol., 366, 99–103. Flores, C.O., Meyer, J.R., Valverde, S., Farr, L. & Weitz, J.S. (2011). Statistical structure of host–phage interactions. Proc. Natl. Acad. Sci. U.S.A., 108, E288–297. Frada, M., Probert, I., Allen, M.J., Wilson, W.H. & De Vargas, C. (2008). The “Cheshire Cat” escape strategy of the coccolithophore Emiliania huxleyi in response to viral infection. Proc. Natl. Acad. Sci. U.S.A., 105, 15944–15949. Fuhrman, J.A. (1999). Marine viruses and their biogeochemical and ecological effects. Nature, 399, 541–548. Fuhrman, J.A. (2009). Microbial community structure and its functional implications. Nature, 459, 193–199. Fuhrman, J.A. & Steele, J.A. (2008). Community structure of marine bacterioplankton: patterns, networks, and relationships to function. Aquat. Microb. Ecol., 53, 69–81. Fuhrman, J.A. & Suttle, C.A. (1993). Viruses in marine planktonic systems. Oceanography, 6, 51–63. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Gilbert, J.A., Field, D., Swift, P., Newbold, L., Oliver, A., Smyth, T., et al. (2009). The seasonal structure of microbial communities in the Western English Channel. Environ Microbiol, 11, 3132–3139. Gilbert, J.A., Steele, J.A., Caporaso, J.G., Steinbrück, L., Reeder, J., Temperton, B., et al. (2012). Defining seasonal marine microbial community dynamics. ISME J, 6, 298–308. Giovannoni, S.J. & Vergin, K.L. (2012). Seasonality in Ocean Microbial Communities. Science, 335, 671–676. Lennon, J.T. & Martiny, J.B.H. (2008). Rapid evolution buffers ecosystem impacts of viruses in a microbial food web. Ecology Letters, 11, 1178–1188. 154 Lennon, J.T., Khatana, S.A.M., Marston, M.F. & Martiny, J.B.H. (2007). Is there a cost of virus resistance in marine cyanobacteria? ISME J, 1, 300–312. Marston, M.F., Pierciey, F.J., Shepard, A., Gearin, G., Qi, J., Yandava, C., et al. (2012). Rapid diversification of coevolving marine Synechococcus and a virus. Proc. Natl. Acad. Sci. U.S.A., 109, 4544–4549. Parsons, R.J., Breitbart, M., Lomas, M.W. & Carlson, C.A. (2012). Ocean time-series reveals recurring seasonal patterns of virioplankton dynamics in the northwestern Sargasso Sea. ISME J, 6, 273–284. Proctor, L.M. & Fuhrman, J.A. (1990). Viral mortality of marine bacteria and cyanobacteria. Nature, 343, 60–62. Rodriguez-Brito, B., Li, L., Wegley, L., Furlan, M., Angly, F.E., Breitbart, M., et al. (2010). Viral and microbial community dynamics in four aquatic environments. ISME J, 4, 739–751. Rohwer, F. & Thurber, R.V. (2009). Viruses manipulate the marine environment. Nature, 459, 207–212. Sandaa, R.-A. (2008). Burden or benefit? Virus–host interactions in the marine environment. Res. Microbiol., 159, 374–381. Shade, A. & Handelsman, J. (2012). Beyond the Venn diagram: the hunt for a core microbiome. Environ Microbiol, 14, 4–12. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. Steward, G.F. & Preston, C.M. (2011). Analysis of a viral metagenomic library from 200 m depth in Monterey Bay, California constructed by direct shotgun cloning. Virol J, 8, 287. Sullivan, M.B., Huang, K.H., Ignacio-Espinoza, J.C., Berlin, A.M., Kelly, L., Weigele, P.R., et al. (2010). Genomic analysis of oceanic cyanobacterial myoviruses compared with T4-like myoviruses from diverse hosts and environments. Environ Microbiol, 12, 3035–3056. Sullivan, M.B., Waterbury, J.B. & Chisholm, S.W. (2003). Cyanophages infecting the oceanic cyanobacterium Prochlorococcus. Nature, 424, 1047–1051. Suttle, C.A. (2005). Viruses in the sea. Nature, 437, 356–361. Thingstad, T.F. (2000). Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol Oceanogr., 1320–1328. Thingstad, T.F. & Cuevas, L. (2010). Nutrient pathways through the microbial food web: principles and predictability discussed, based on five different experiments. Aquat. Microb. Ecol., 61, 249251–249262. 155 Thingstad, T.F. & Lignell, R. (1997). Theoretical models for the control of bacterial growth rate, abundance, diversity and carbon demand. Aquat. Microb. Ecol., 13, 19–27. Weinbauer, M.G., Christaki, U., Nedoma, J. & Simek, K. (2003). Comparing the effects of resource enrichment and grazing on viral production in a meso-eutrophic reservoir. Aquat. Microb. Ecol., 31, 137–144. Wilhelm, S.W. & Suttle, C.A. (1999). Viruses and nutrient cycles in the sea. Bioscience, 49, 781– 788. Wommack, K.E. & Colwell, R.R. (2000). Virioplankton: viruses in aquatic ecosystems. Microbiol. Mol. Biol. Rev., 64, 69–114. 156 Bibliography (Alphabetized) Acinas, S.G., Rodrı ́ guez Valera, F. & Pedrós-Alió, C. (1997). Spatial and temporal variation in marine bacterioplankton diversity as shown by RFLP fingerprinting of PCR amplified 16S rDNA. FEMS Microbiol Ecol, 24, 27–40. Ahlgren, N.A. & Rocap, G. (2006). Culture Isolation and Culture-Independent Clone Libraries Reveal New Marine Synechococcus Ecotypes with Distinctive Light and N Physiologies. Appl. Environ. Microbiol., 72, 7193–7204. Allesina, S. & Pascual, M. (2008). Network structure, predator–prey modules, and stability in large food webs. Theor Ecol, 1, 55–64. Alonso Sáez, L., Balagué, V., Sà, E.L., Sánchez, O., González, J.M., Pinhassi, J., et al. (2007). Seasonality in bacterial diversity in north‐west Mediterranean coastal waters: assessment through clone libraries, fingerprinting and FISH. FEMS Microbiol Ecol, 60, 98–112. Angly, F.E., Felts, B., Breitbart, M., Salamon, P., Edwards, R.A., Carlson, C.A., et al. (2006). The marine viromes of four oceanic regions. Plos Biol, 4, e368. Angly, F.E., Rodriguez-Brito, B., Bangor, D., McNairnie, P., Breitbart, M., Salamon, P., et al. (2005). PHACCS, an online tool for estimating the structure and diversity of uncultured viral communities using metagenomic information. BMC Bioinformatics, 6, 41. Assenov, Y., Ramírez, F., Schelhorn, S., Lengauer, T. & Albrecht, M. (2008). Computing topological parameters of biological networks. Bioinformatics, 24, 282–284. Azam, F. & Worden, A.Z. (2004). Microbes, molecules, and marine ecosystems. Science, 303, 1622–1624. Azam, F., Fenchel, T., Field, J., Gray, J., Meyer-Reil, L. & Thingstad, F. (1983). The ecological role of water-column microbes in the sea. Mar. Ecol. Prog. Ser., 10, 257–263. Baudoux, A. & Veldhuis, M. (2008). Estimates of virus-vs. grazing induced mortality of picophytoplankton in the North Sea during summer. Aquat. Microb. Ecol. Bellec, L., Grimsley, N., Derelle, E., Moreau, H. & Desdevises, Y. (2010). Abundance, spatial distribution and genetic diversity of Ostreococcus tauriviruses in two different environments. Environ Microbiol Rep, 2, 313–321. Beman, J.M., Sachdeva, R. & Fuhrman, J.A. (2010). Population ecology of nitrifying Archaea and Bacteria in the Southern California Bight. Environ Microbiol, 12, 1282–1292. Beman, J.M., Steele, J.A. & Fuhrman, J.A. (2011). Co-occurrence patterns for abundant marine archaeal and bacterial lineages in the deep chlorophyll maximum of coastal California. ISME J, 5, 1077–1085. 157 Bench, S.R., Hanson, T.E., Williamson, K.E., Ghosh, D., Radosovich, M., Wang, K., et al. (2007). Metagenomic Characterization of Chesapeake Bay Virioplankton. Appl. Environ. Microbiol., 73, 7629–7641. Berelson, W. (1991). The flushing of two deep-sea basins, southern California borderland. Limnol Oceanogr. Bergh, Ø., Børsheim, K.Y., Bratbak, G. & Heldal, M. (1989). High abundance of viruses found in aquatic environments. Nature, 340, 467–468. Bhavsar, J., Polson, S.W. & Wommack, K.E. (2009). VIROME: An informatics resource for viral metagenome exploration. Bidle, K.D. & Vardi, A. (2011). A chemical arms race at sea mediates algal host-virus interactions. Curr. Opin. Microbiol, 14, 449–457. Bouskill, N.J., Eveillard, D., O'Mullan, G., Jackson, G.A. & Ward, B.B. (2010). Seasonal and annual reoccurrence in betaproteobacterial ammonia-oxidizing bacterial population structure. Environ Microbiol, 13, 872–886. Børsheim, K.Y. (1993). Native marine bacteriophages. FEMS Microbiol Ecol, 102, 141–159. Bouvier, T. & del Giorgio, P.A. (2007). Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol, 9, 287–297. Bouvy, M., Bettarel, Y., Bouvier, C., Domaizon, I., Jacquet, S., Le Floc'h, E., et al. (2011). Trophic interactions between viruses, bacteria and nanoflagellates under various nutrient conditions and simulated climate change. Environ Microbiol, 13, 1842–1857. Braak, Ter, C.J.F. & Schaffers, A.P. (2004). Co-correspondence analysis: a new ordination method to relate two community compositions. Ecology, 85, 834–846. Bratbak, G., Thingstad, F. & Heldal, M. (1994). Viruses and the microbial loop. Microb Ecol, 28, 209–221. Breitbart, M. (2002). Genomic analysis of uncultured marine viral communities. Proc. Natl. Acad. Sci. U.S.A., 99, 14250–14255. Breitbart, M. (2012). Marine Viruses: Truth or Dare. Annu. Rev. Marine. Sci., 4, 425–448. Breitbart, M. & Rohwer, F. (2005). Here a virus, there a virus, everywhere the same virus? Trends Microbiol., 13, 278–284. Breitbart, M., Felts, B., Kelley, S., Mahaffy, J.M., Nulton, J., Salamon, P., et al. (2004). Diversity and population structure of a near-shore marine-sediment viral community. Proc R Soc Lond B, 271, 565–574. Breitbart, M., Hewson, I., Felts, B., Mahaffy, J.M., Nulton, J., Salamon, P., et al. (2003). Metagenomic analyses of an uncultured viral community from human feces. J. Bacteriol., 185, 6220–6223. 158 Brown, J.H. (1984). On the relationship between abundance and distribution of species. Am. Nat., 255–279. Brown, M.V. & Fuhrman, J.A. (2005). Marine bacterial microdiversity as revealed by internal transcribed spacer analysis. Aquat. Microb. Ecol., 41, 15–23. Brown, M.V., Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2005). Coupling 16S-ITS rDNA clone libraries and automated ribosomal intergenic spacer analysis to show marine microbial diversity: development and application to a time series. Environ Microbiol, 7, 1466–1479. Brussaard, C.P.D., Wilhelm, S.W., Thingstad, F., Weinbauer, M.G., Bratbak, G., Heldal, M., et al. (2008). Global-scale processes with a nanoscale drive: the role of marine viruses. ISME J, 2, 575–578. Buchan, A. & Gonzalez, J. (2005). Overview of the Marine Roseobacter Lineage. Appl. Environ. Microbiol. Butina, T.V., Belykh, O.I., Maksimenko, S.Y. & Belikov, S.I. (2010). Phylogenetic diversity of T4-like bacteriophages in Lake Baikal, East Siberia. FEMS Microbiol Lett, 309, 122-129. Børsheim, K.Y. (1993). Native marine bacteriophages. FEMS Microbiol Ecol, 102, 141–159. Cai, H., Wang, K., Huang, S., Jiao, N. & Chen, F. (2010). Distinct Patterns of Picocyanobacterial Communities in Winter and Summer in the Chesapeake Bay. Appl. Environ. Microbiol., 76, 2955–2960. Campbell, B.J., Yu, L., Heidelberg, J.F. & Kirchman, D.L. (2011). Activity of abundant and rare bacteria in a coastal ocean. Proc. Natl. Acad. Sci. U.S.A., 108, 12776. Campbell, L., Liu, H., Nolla, H.A. & Vaulot, D. (1997). Annual variability of phytoplankton and bacteria in the subtropical North Pacific Ocean at Station ALOHA during the 1991-1994 ENSO event. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 44, 167–192. Caporaso, J., Knight, R. & al, E. (2011). Host-associated and free-living phage communities differ profoundly in phylogenetic composition. PLoS ONE. Caporaso, J.G., Paszkiewicz, K., Field, D., Knight, R. & Gilbert, J.A. (2011). The Western English Channel contains a persistent microbial seed bank. ISME J. Carlson, C.A., Morris, R., Parsons, R., Treusch, A.H., Giovannoni, S.J. & Vergin, K.L. (2009). Seasonal dynamics of SAR11 populations in the euphotic and mesopelagic zones of the northwestern Sargasso Sea. ISME J, 3, 283–295. Chaffron, S., Rehrauer, H., Pernthaler, J. & Mering, von, C. (2010). A global network of coexisting microbes from environmental and whole-genome sequence data. Genome Research, 20, 947–959. Chisholm, S.W., Frankel, S.L., Goericke, R., Olson, R.J., Palenik, B., Waterbury, J.B., et al. (1992). Prochlorococcus marinus nov. gen. nov. sp.: an oxyphototrophic marine prokaryote containing divinyl chlorophyll a and b. Arch. Microbiol., 157, 297–300. 159 Chisholm, S.W., Olson, R.J., Zettler, E.R., Goericke, R., Waterbury, J.B. & Welschmeyer, N.A. (1988). A novel free-living prochlorophyte abundant in the oceanic euphotic zone. Nature, 334, 340–343. Chow, C.-E.T. & Fuhrman, J.A. (2012). Seasonality and monthly dynamics of marine myovirus communities. Environ Microbiol. In press. Christaki, U., Jacquet, S., Dolan, J., Vaulot, D. & Rassoulzadegan, F. (1999). Growth and Grazing on Prochlorococcus and Synechococcus by Two Marine Ciliates. Limnol Oceanogr., 44, 52–61. Christaki, U., Vazquez-Dominguez, E., Courties, C. & Lebaron, P. (2005). Grazing impact of different heterotrophic nanoflagellates on eukaryotic (Ostreococcus tauri ) and prokaryotic picoautotrophs (Prochlorococcus and Synechococcus). Environ Microbiol, 7, 1200–1210. Clarke, K.R. (1993). Non‐parametric multivariate analyses of changes in community structure. Aust J Ecol, 18, 117–143. Clarke, K.R. & Gorley, R. (2006). PRIMER v6: User Manual/Tutorial. PRIMER-E, Plymouth, UK. Cline, M.S., Smoot, M., Cerami, E., Kuchinsky, A., Landys, N., Workman, C., et al. (2007). Integration of biological networks and gene expression data using Cytoscape. Nat Protoc, 2, 2366–2382. Clokie, M.R.J. & Mann, N.H. (2006). Marine cyanophages and light. Environ Microbiol, 8, 2074–2082. Clokie, M.R.J., Millard, A.D. & Mann, N.H. (2010). T4 genes in the marine ecosystem: studies of the T4-like cyanophages and their role in marine ecology. Virol J, 7, 291. Clokie, M.R.J., Millard, A.D., Letarov, A.V. & Heaphy, S. (2011). Phages in nature. Bacteriophage, 1, 31–45. Clokie, M.R.J., Millard, A.D., Mehta, J.Y. & Mann, N.H. (2006). Virus isolation studies suggest short-term variations in abundance in natural cyanophage populations of the Indian Ocean. J. Mar. Biol. Ass., 86, 499–505. Clokie, M.R.J., Shan, J., Bailey, S., Jia, Y., Krisch, H.M., West, S., et al. (2006). Transcription of a “photosynthetic” T4-type phage during infection of a marine cyanobacterium. Environ Microbiol, 8, 827–835. Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R.J., et al. (2009). The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res, 37, D141–D145. Collins, L.E., Berelson, W., Hammond, D.E., Knapp, A., Schwartz, R. & Capone, D. (2011). Particle fluxes in San Pedro Basin, California A four-year record of sedimentation and physical forcing. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 58, 898–914. 160 Collins, R.E. & Rocap, G. (2007). REPK: an analytical web server to select restriction endonucleases for terminal restriction fragment length polymorphism analysis. Nucleic Acids Res, 35, W58–W62. Comeau, A.M. & Krisch, H.M. (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, 1321–1332. Corno, G. & Jürgens, K. (2008). Structural and functional patterns of bacterial communities in response to protist predation along an experimental productivity gradient. Environ Microbiol, 10, 2857–2871. Countway, P.D. & Caron, D.A. (2006). Abundance and distribution of Ostreococcus sp. in the San Pedro Channel, California, as revealed by quantitative PCR. Appl. Environ. Microbiol., 72, 2496–2506. Countway, P.D., Gast, R.J., Savai, P. & Caron, D.A. (2005). Protistan diversity estimates based on 18S rDNA from seawater incubations in the western north Atlantic1. J. Eukaryot. Microbiol., 52, 95–106. Countway, P.D., Vigil, P.D., Schnetzer, A., Moorthi, S.D. & Caron, D.A. (2010). Seasonal analysis of protistan community structure and diversity at the USC Microbial Observatory (San Pedro Channel, North Pacific Ocean). Limnol Oceanogr., 55, 2381–2396. Culley, A. & Lang, A. (2006). Metagenomic Analysis of Coastal RNA Virus Communities. Science. Danovaro, R., Dell'Anno, A., Corinaldesi, C., Magagnini, M., Noble, R., Tamburini, C., et al. (2008). Major viral impact on the functioning of benthic deep-sea ecosystems. Nature, 454, 1084–1087. De Corte, D., Sintes, E., Winter, C., Yokokawa, T., Reinthaler, T. & Herndl, G.J. (2010). Links between viral and prokaryotic communities throughout the water column in the (sub)tropical Atlantic Ocean. ISME J, 4, 1431–1442. DeLong, E.F., Preston, C.M., Mincer, T., Rich, V., Hallam, S.J., Frigaard, N.-U., et al. (2006). Community genomics among stratified microbial assemblages in the ocean's interior. Science, 311, 496–503. Dinsdale, E.A., Edwards, R.A., Hall, D., Angly, F.E., Breitbart, M., Brulc, J.M., et al. (2008a). Functional metagenomic profiling of nine biomes. Nature, 452, 629–632. Dinsdale, E.A., Pantos, O., Smriga, S., Edwards, R.A., Angly, F.E., Wegley, L., et al. (2008b). Microbial Ecology of Four Coral Atolls in the Northern Line Islands. PLoS ONE, 3, e1584. Ducklow, H. (1983). Production and Fate of Bacteria in the Oceans. Bioscience. Ducklow, H.W., Doney, S.C. & Steinberg, D.K. (2009). Contributions of Long-Term Research and Time-Series Observations to Marine Ecology and Biogeochemistry. Annu. Rev. Marine. Sci., 1, 279–302. 161 Duhaime, M.B., Deng, L., Poulos, B. & Sullivan, M.B. (n.d.). Linker amplification for quantitative metagenomics of wild viruses and other ultra-low concentration DNA samples, in prep, 1–46. Dunne, J.A. (2002). Food-web structure and network theory: The role of connectance and size. Proc. Natl. Acad. Sci. U.S.A., 99, 12917–12922. DuRand, M.D., Olson, R.J. & Chisholm, S.W. (2001). Phytoplankton population dynamics at the Bermuda Atlantic Time-series station in the Sargasso Sea. Deep-Sea Res. (2 Top. Stud. Oceanogr.), 48, 1983–2003. Edwards, R.A. & Rohwer, F. (2005). Viral metagenomics. Nat Rev Micro, 3, 504–510. Eiler, A., Hayakawa, D.H. & Rappé, M.S. (2011). Non-random assembly of bacterioplankton communities in the subtropical north pacific ocean. Front. Microbio., 2, 1–10. Eiler, A., Hayakawa, D.H., Church, M.J., Karl, D.M. & Rappé, M.S. (2009). Dynamics of the SAR11 bacterioplankton lineage in relation to environmental conditions in the oligotrophic North Pacific subtropical gyre. Environ Microbiol, 11, 2291–2300. Eiler, A., Heinrich, F. & Bertilsson, S. (2012). Coherent dynamics and association networks among lake bacterioplankton taxa. ISME J, 6, 330–342. Evans, C., Archer, S.D., Jacquet, S. & Wilson, W.H. (2003). Direct estimates of the contribution of viral lysis and microzooplankton grazing to the decline of a Micromonas spp. population. Aquat. Microb. Ecol., 30, 207–219. Fenchel, T. (2008). The microbial loop – 25 years later. J. Exp. Mar. Biol. Ecol., 366, 99–103. Fierer, N., Breitbart, M., Nulton, J., Salamon, P., Lozupone, C., Jones, R., et al. (2007). Metagenomic and Small-Subunit rRNA Analyses Reveal the Genetic Diversity of Bacteria, Archaea, Fungi, and Viruses in Soil. Appl. Environ. Microbiol., 73, 7059–7066. Filee, J. (2005). Marine T4-type bacteriophages, a ubiquitous component of the dark matter of the biosphere. Proc. Natl. Acad. Sci. U.S.A., 102, 12471–12476. Fisher, M.M. & Triplett, E.W. (1999). Automated approach for ribosomal intergenic spacer analysis of microbial diversity and its application to freshwater bacterial communities. Appl. Environ. Microbiol., 65, 4630–4636. Flores, C.O., Meyer, J.R., Valverde, S., Farr, L. & Weitz, J.S. (2011). Statistical structure of host–phage interactions. Proc. Natl. Acad. Sci. U.S.A., 108, E288–297. Frada, M., Probert, I., Allen, M.J., Wilson, W.H. & De Vargas, C. (2008). The “Cheshire Cat” escape strategy of the coccolithophore Emiliania huxleyi in response to viral infection. Frederickson, C.M., Short, S.M. & Suttle, C.A. (2003). The Physical Environment Affects Cyanophage Communities in British Columbia Inlets. Microb Ecol, 46, 348–357. 162 Frias-Lopez, J., Thompson, A., Waldbauer, J. & Chisholm, S.W. (2009). Use of stable isotope- labelled cells to identify active grazers of picocyanobacteria in ocean surface waters. Environ Microbiol, 11, 512–525. Fuhrman, J.A. (1999). Marine viruses and their biogeochemical and ecological effects. Nature, 399, 541–548. Fuhrman, J.A. (2009). Microbial community structure and its functional implications. Nature, 459, 193–199. Fuhrman, J.A. & Noble, R.T. (1995). Viruses and protists cause similar bacterial mortality in coastal seawater. Limnol Oceanogr., 1236–1242. Fuhrman, J.A. & Schwalbach, M. (2003). Viral influence on aquatic bacterial communities. The Biological Bulletin, 204, 192–195. Fuhrman, J.A. & Steele, J.A. (2008). Community structure of marine bacterioplankton: patterns, networks, and relationships to function. Aquat. Microb. Ecol., 53, 69–81. Fuhrman, J.A. & Suttle, C.A. (1993). Viruses in marine planktonic systems. Oceanography, 6, 51–63. Fuhrman, J.A., Hewson, I., Schwalbach, M.S., Steele, J.A., Brown, M.V. & Naeem, S. (2006). Annually reoccurring bacterial communities are predictable from ocean conditions. Proc. Natl. Acad. Sci. U.S.A., 103, 13104–13109. Fuhrman, J.A., Steele, J.A., Hewson, I., Schwalbach, M.S., Brown, M.V., Green, J.L., et al. (2008). A latitudinal diversity gradient in planktonic marine bacteria. Proc. Natl. Acad. Sci. U.S.A., 105, 7774–7778. Fujii, T., Nakayama, N., Nishida, M., Sekiya, H., Kato, N., Asakawa, S., et al. (2008). Novel capsid genes (g23) of T4-type bacteriophages in a Japanese paddy field. Soil Biol. Biochem., 40, 1049–1058. Fuller, N.J., Wilson, W.H., Joint, I.R. & Mann, N.H. (1998). Occurrence of a sequence in marine cyanophages similar to that of T4 g20 and its application to PCR-based detection and quantification techniques. Appl. Environ. Microbiol., 64, 2051–2060. Gasol, J.M., Pedrós-Alió, C. & Vaqué, D. (2002). Regulation of bacterial assemblages in oligotrophic plankton systems: results from experimental and empirical approaches. Antonie Leeuwenhoek, 81, 435–452. Gilbert, J.A., Field, D., Swift, P., Newbold, L., Oliver, A., Smyth, T., et al. (2009). The seasonal structure of microbial communities in the Western English Channel. Environ Microbiol, 11, 3132–3139. Gilbert, J.A., Steele, J.A., Caporaso, J.G., Steinbrück, L., Reeder, J., Temperton, B., et al. (2012). Defining seasonal marine microbial community dynamics. ISME J, 6, 298–308. Giovannoni, S.J. & Vergin, K.L. (2012). Seasonality in Ocean Microbial Communities. Science, 335, 671–676. 163 Goldsmith, D.B., Crosti, G., Dwivedi, B., McDaniel, L.D., Varsani, A., Suttle, C.A., et al. (2011). Development of phoH as a Novel Signature Gene for Assessing Marine Phage Diversity. Appl. Environ. Microbiol., 77, 7730–7739. Gong, G., Chiang, K. & Worden, A. (2012). Distribution patterns and phylogeny of marine stramenopiles (MAST) in the North Pacific Ocean. Appl. Environ. Microbiol. Hamady, M., Lozupone, C. & Knight, R. (2009). Fast UniFrac: facilitating high-throughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J, 4, 17–27. Hara, S., Koike, I., Terauchi, K. & Kamiya, H. (1996). Abundance of viruses in deep oceanic waters. Mar. Ecol. Prog. Ser., 145, 269–277. Helton, R.R. & Wommack, K.E. (2009). Seasonal Dynamics and Metagenomic Characterization of Estuarine Viriobenthos Assemblages by Randomly Amplified Polymorphic DNA PCR. Appl. Environ. Microbiol., 75, 2259–2265. Henn, M.R., Sullivan, M.B., Stange-Thomann, N., Osburne, M.S., Berlin, A.M., Kelly, L., et al. (2010). Analysis of High-Throughput Sequencing and Annotation Strategies for Phage Genomes. PLoS ONE, 5, e9083. Hewson, I. & Fuhrman, J.A. (2006). Viral impacts upon marine bacterioplankton assemblage structure. J. Mar. Biol. Ass., 86, 577–589. Hewson, I. & Fuhrman, J.A. (2007). Covariation of viral parameters with bacterial assemblage richness and diversity in the water column and sediments. Deep-Sea Res. (1 Oceanogr. Res. Pap.), 54, 811–830. Hoffmann, K.H., Rodriguez-Brito, B., Breitbart, M., Bangor, D., Angly, F.E., Felts, B., et al. (2007). Power law rankâ“abundance models for marine phage communities. FEMS Microbiol Lett, 273, 224–228. Holmfeldt, K., Titelman, J. & Riemann, L. (2010). Virus Production and Lysate Recycling in Different Sub-basins of the Northern Baltic Sea. Microb Ecol, 60, 572–580. Ings, T.C., Montoya, J.M., Bascompte, J., Blüthgen, N., Brown, L., Dormann, C.F., et al. (2009). Review: Ecological networks - beyond food webs. J Anim Ecology, 78, 253–269. Jacquet, S., Miki, T., Noble, R., Peduzzi, P. & Wilhelm, S. (2010). Viruses in aquatic ecosystems: important advancements of the last 20 years and prospects for the future in the field of microbial oceanography and limnology. Advances in Oceanography and Limnology, 1, 97– 141. Jameson, E., Mann, N.H., Joint, I., Sambles, C. & hling, M.M.U. (2011). The diversity of cyanomyovirus populations along a North-South Atlantic Ocean transect. ISME J, 5, 1713– 1721. Jiang, S., Fu, W., Chu, W. & Fuhrman, J.A. (2003). The Vertical Distribution and Diversity of Marine Bacteriophage at a Station off Southern California. Microb Ecol, 45, 399–410. 164 John, S.G., Mendez, C.B., Deng, L., Poulos, B., Kauffman, A.K.M., Kern, S., et al. (2010). A simple and efficient method for concentration of ocean viruses by chemical flocculation. Environ Microbiol Rep, Virus concentration by flocculation with iron, 3, 195–202. Kan, J., Suzuki, M.T., Wang, K., Evans, S.E. & Chen, F. (2007). High Temporal but Low Spatial Heterogeneity of Bacterioplankton in the Chesapeake Bay. Appl. Environ. Microbiol., 73, 6776–6789. Kottmann, R., Kostadinov, I., Duhaime, M.B., Buttigieg, P.L., Yilmaz, P., Hankeln, W., et al. (2009). Megx.net: integrated database resource for marine ecological genomics. Nucleic Acids Res, 38, D391–D395. Kristensen, D.M., Mushegian, A.R., Dolja, V.V. & Koonin, E.V. (2010). New dimensions of the virus world discovered through metagenomics. Trends Microbiol., 18, 11–19. Labonte, J.M., Reid, K.E. & Suttle, C.A. (2009). Phylogenetic Analysis Indicates Evolutionary Diversity and Environmental Segregation of Marine Podovirus DNA Polymerase Gene Sequences. Appl. Environ. Microbiol., 75, 3634–3640. Larsen, A., Flaten, G.A.F., Sandaa, R.-A., Castberg, T., Thyrhaug, R., Erga, S.R., et al. (2004). Spring phytoplankton bloom dynamics in Norwegian coastal waters: Microbial community succession and diversity. Limnol Oceanogr., 49, 180–190. Lavigne, R., Darius, P., Summer, E.J., Seto, D., Mahadevan, P., Nilsson, A.S., et al. (2009). Classification of Myoviridae bacteriophages using protein sequence similarity. BMC Microbiol, 9, 224. Lawton, J.H. (1993). Range, population abundance and conservation. Trends Ecol Evol, 8, 409– 413. Lennon, J.T. & Martiny, J.B.H. (2008). Rapid evolution buffers ecosystem impacts of viruses in a microbial food web. Ecology Letters, 11, 1178–1188. Lennon, J.T., Khatana, S.A.M., Marston, M.F. & Martiny, J.B.H. (2007). Is there a cost of virus resistance in marine cyanobacteria? ISME J, 1, 300–312. Li, W.K.W. (1994). Primary production of prochlorophytes, cyanobacteria, and eucaryotic ultraphytoplankton: measurements from flow cytometric sorting. Limnol Oceanogr., 39, 169–175. Li, W. (1998). Annual average abundance of heterotrophic bacteria and Synechococcus in surface ocean waters. Limnol Oceanogr., 1746–1753. Lindell, D., Jaffe, J.D., Coleman, M.L., Futschik, M.E., Axmann, I.M., Rector, T., et al. (2007). Genome-wide expression dynamics of a marine virus and host reveal features of co- evolution. Nature, 449, 83–86. Lindell, D., Jaffe, J.D., Johnson, Z.I., Church, G.M. & Chisholm, S.W. (2005). Photosynthesis genes in marine viruses yield proteins during host infection. Nature, 438, 86–89. 165 Liu, H., Nolla, H.A. & Campbell, L. (1997). Prochlorococcus growth rate and contribution to primary production in the equatorial and subtropical North Pacific Ocean. Aquat. Microb. Ecol., 12, 39–47. Long, A., McDaniel, L.D., Mobberley, J. & Paul, J.H. (2007). Comparison of lysogeny (prophage induction) in heterotrophic bacterial and Synechococcus populations in the Gulf of Mexico and Mississippi river plume. ISME J, 2, 132–144. Longnecker, K., Wilson, M., Sherr, E. & Sherr, B.F. (2010). Effect of top-down control on cell- specific activity and diversity of active marine bacterioplankton. Aquat. Microb. Ecol., 58, 153–165. Lopez-Bueno, A., Tamames, J., Velazquez, D., Moya, A., Quesada, A. & Alcami, A. (2009). High Diversity of the Viral Community from an Antarctic Lake. Science, 326, 858–861. Lozupone, C. & Knight, R. (2005). UniFrac: a new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol., 71, 8228–8235. Lozupone, C., Hamady, M. & Knight, R. (2006). UniFrac–an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinformatics, 7, 371. Malmstrom, R.R., Coe, A., Kettler, G.C., Martiny, A.C., Frias-Lopez, J., Zinser, E.R., et al. (2010). Temporal dynamics of Prochlorococcus ecotypes in the Atlantic and Pacific oceans. ISME J, 4, 1252–1264. Marston, M.F. & Sallee, J.L. (2003). Genetic diversity and temporal variation in the cyanophage community infecting marine Synechococcus species in Rhode Island's coastal waters. Appl. Environ. Microbiol., 69, 4639–4647. Martiny, A.C., Tai, A.P.K., Veneziano, D., Primeau, F. & Chisholm, S.W. (2009). Taxonomic resolution, ecotypes and the biogeography of Prochlorococcus. Environ Microbiol, 11, 823– 832. Martiny, J.B.H., Bohannan, B.J.M., Brown, J.H., Colwell, R.K., Fuhrman, J.A., Green, J.L., et al. (2006). Microbial biogeography: putting microorganisms on the map. Nat Rev Micro, 4, 102–112. Marston, M.F., Pierciey, F.J., Shepard, A., Gearin, G., Qi, J., Yandava, C., et al. (2012). Rapid diversification of coevolving marine Synechococcus and a virus. Proc. Natl. Acad. Sci. U.S.A., 109, 4544–4549. McDaniel, L. & Delarosa, M. (2006). Cambridge Journals Online - Abstract - Temperate and lytic cyanophages from the Gulf of Mexico. J. Mar. Biol. Ass. McDonald, D., Price, M.N., Goodrich, J., Nawrocki, E.P., DeSantis, T.Z., Probst, A., et al. (2011). An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J, 6, 610–618. McManus, G.B. & Fuhrman, J.A. (1988). Control of marine bacterioplankton populations: Measurement and significance of grazing. Hydrobiologia, 159, 51–62. 166 Middelboe, M. (2000). Bacterial Growth Rate and Marine Virus-Host Dynamics. Microb Ecol, 40, 114–124. Miki, T. & Jacquet, S. (2008). Complex interactions in the microbial world: underexplored key links between viruses, bacteria and protozoan grazers in aquatic environments. Aquat. Microb. Ecol., 51, 195–208. Millard, A.D., Zwirglmaier, K., Downey, M.J., Mann, N.H. & Scanlan, D.J. (2009). Comparative genomics of marine cyanomyoviruses reveals the widespread occurrence of Synechococcushost genes localized to a hyperplastic region: implications for mechanisms of cyanophage evolution. Environ Microbiol, 11, 2370–2387. Mitchell, R.J., Hester, A.J., Campbell, C.D., Chapman, S.J., Cameron, C.M., Hewison, R.L., et al. (2011). Explaining the variation in the soil microbial community: do vegetation composition and soil chemistry explain the same or different parts of the microbial variation? Plant Soil, 351, 355–362. Moebus, K. (1996). Marine bacteriophage reproduction under nutrient-limited growth of host bacteria. I. Investigations with six phage-host systems. Mar. Ecol. Prog. Ser., 14, 1–12. Monier, A., Welsh, R.M., Gentemann, C., Weinstock, G., Sodergren, E., Armbrust, E.V., et al. (2011). Phosphate transporters in marine phytoplankton and their viruses: cross-domain commonalities in viral-host gene exchanges. Environ Microbiol, 14, 162–176. Montoya, J.M., Pimm, S.L. & Solé, R.V. (2006). Ecological networks and their fragility. Nature, 442, 259–264. Moore, L.R. & Chisholm, S.W. (1999). Photophysiology of the marine cyanobacterium Prochlorococcus: Ecotypic differences among cultured isolates. Limnol Oceanogr., 628–638. Moore, L.R., Rocap, G. & Chisholm, S.W. (1998). Physiology and molecular phylogeny of coexisting Prochlorococcus ecotypes. Nature, 393, 464–467. Morris, R.M., Frazar, C.D. & Carlson, C.A. (2012). Basin-scale patterns in the abundance of SAR11 subclades, marine Actinobacteria (OM1), members of the Roseobacter clade and OCS116 in the South Atlantic. Environ Microbiol, In press. Morris, R.M., Rappé, M.S., Connon, S.A., Vergin, K.L., Siebold, W.A., Carlson, C.A., et al. (2002). SAR 11 clade dominates ocean surface bacterioplankton communities. Nature, 420, 806–810. Morris, R.M., Vergin, K.L., Cho, J.C., Rappé, M.S., Carlson, C.A. & Giovannoni, S.J. (2005). Temporal and spatial response of bacterioplankton lineages to annual convective overturn at the Bermuda Atlantic Time-series Study site. Limnol Oceanogr., 1687–1696. Mühling, M., Fuller, N.J., Millard, A., Somerfield, P.J., Marie, D., Wilson, W.H., et al. (2005). Genetic diversity of marine Synechococcus and co-occurring cyanophage communities: evidence for viral control of phytoplankton. Environ Microbiol, 7, 499–508. 167 Müller, J., Stadler, J., Jarzabek-Müller, A., Hacker, H., Braak, ter, C. & Brandl, R. (2011). The Predictability of Phytophagous Insect Communities: Host Specialists as Habitat Specialists. PLoS ONE, 6, e25986. Noble, R.T. & Fuhrman, J.A. (1998). Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Aquat. Microb. Ecol., 14, 113–118. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O'Hara, R.B., et al. (2011). vegan: Community Ecology Package (R package version 2.0.2). R package version 2.0.2. Olesen, J.M., Stefanescu, C. & Traveset, A. (2011). Strong, Long-Term Temporal Dynamics of an Ecological Network. PLoS ONE, 6, e26455. Ory, P., Hartmann, H.J., Jude, F., Dupuy, C., Del Amo, Y., Catala, P., et al. (2010). Pelagic food web patterns: do they modulate virus and nanoflagellate effects on picoplankton during the phytoplankton spring bloom? Environ Microbiol, 12, 2755–2772. Paerl, R., Turk, K., Beinart, R., Chavez, F.P. & Zehr, J. (2012). Seasonal change in the abundance of Synechococcus and multiple distinct phylotypes in Monterey Bay determined by rbcL and narB quantitative PCR. Environ Microbiol, 14, 580–593. Parsons, R.J., Breitbart, M., Lomas, M.W. & Carlson, C.A. (2012). Ocean time-series reveals recurring seasonal patterns of virioplankton dynamics in the northwestern Sargasso Sea. ISME J, 6, 273–284. Partensky, F. & Garczarek, L. (2010). Prochlorococcus: Advantages and Limits of Minimalism. Annu. Rev. Marine. Sci., 2, 305–331. Partensky, F., Hess, W.R. & Vaulot, D. (1999). Prochlorococcus, a marine photosynthetic prokaryote of global significance. Microbiol. Mol. Biol. Rev., 63, 106–127. Patel, A., Noble, R.T., Steele, J.A., Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2007). Virus and prokaryote enumeration from planktonic aquatic environments by epifluorescence microscopy with SYBR Green I. Nat Protoc, 2, 269–276. Pedrós-Alió, C. (2012). The Rare Bacterial Biosphere. Annu. Rev. Marine. Sci., 4, 449–466. Pomeroy, L.R., Williams, P.I., Azam, F. & Hobbie, J.E. (2007). The Microbial Loop. Oceanography, 20, 28–33. Pommier, T., Canbäck, B., Riemann, L., Boström, K.H., Simu, K., Lundberg, P., et al. (2007). Global patterns of diversity and community structure in marine bacterioplankton. Mol. Ecol., 16, 867–880. Proctor, L.M. & Fuhrman, J.A. (1990). Viral mortality of marine bacteria and cyanobacteria. Nature, 343, 60–62. Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J., et al. (2007). SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res, 35, 7188–7196. 168 Robidart, J.C., Preston, C.M., Paerl, R.W., Turk, K.A., Mosier, A.C., Francis, C.A., et al. (2012). Seasonal Synechococcus and Thaumarchaeal population dynamics examined with high resolution with remote in situ instrumentation. ISME J, 6, 513–523. Rocap, G., Distel, D.L., Waterbury, J.B. & Chisholm, S.W. (2002). Resolution of Prochlorococcus and Synechococcus ecotypes by using 16S-23S ribosomal DNA internal transcribed spacer sequences. Appl. Environ. Microbiol., 68, 1180–1191. Rocap, G., Larimer, F.W., Lamerdin, J., Malfatti, S., Chain, P., Ahlgren, N.A., et al. (2003). Genome divergence in two Prochlorococcus ecotypes reflects oceanic niche differentiation. Nature, 424, 1042–1047. Rodriguez-Brito, B., Li, L., Wegley, L., Furlan, M., Angly, F.E., Breitbart, M., et al. (2010). Viral and microbial community dynamics in four aquatic environments. ISME J, 4, 739–751. Rohwer, F. & Edwards, R.A. (2002). The Phage Proteomic Tree: a genome-based taxonomy for phage. J. Bacteriol., 184, 4529–4535. Rohwer, F. & Thurber, R.V. (2009). Viruses manipulate the marine environment. Nature, 459, 207–212. Roux, S., Enault, F., Robin, A., Ravet, V., Personnic, S., Theil, S., et al. (2012). Assessing the Diversity and Specificity of Two Freshwater Viral Communities through Metagenomics. PLoS ONE, 7, e33641. Ruan, Q., Dutta, D., Schwalbach, M.S., Steele, J.A., Fuhrman, J.A. & Sun, F. (2006a). Local similarity analysis reveals unique associations among marine bacterioplankton species and environmental factors. Bioinformatics, 22, 2532–2538. Ruan, Q., Steele, J.A., Schwalbach, M.S., Fuhrman, J.A. & Sun, F. (2006b). A dynamic programming algorithm for binning microbial community profiles. Bioinformatics, 22, 1508– 1514. Sambrook, M., Fritsch, E.F. & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Press, New York, USA. Sandaa, R.-A. (2008). Burden or benefit? Virus–host interactions in the marine environment. Res. Microbiol., 159, 374–381. Sandaa, R.-A. & Larsen, A. (2006). Seasonal Variations in Virus-Host Populations in Norwegian Coastal Waters: Focusing on the Cyanophage Community Infecting Marine Synechococcus spp. Appl. Environ. Microbiol., 72, 4610–4618. Sandaa, R.-A., Clokie, M.R.J. & Mann, N.H. (2008). Photosynthetic genes in viral populations with a large genomic size range from Norwegian coastal waters. FEMS Microbiol Ecol, 63, 2–11. Sandaa, R.-A., Gómez-Consarnau, L., Pinhassi, J., Riemann, L., Malits, A., Weinbauer, M.G., et al. (2009). Viral control of bacterial biodiversity - evidence from a nutrient-enriched marine mesocosm experiment. Environ Microbiol, 11, 2585–2597. 169 Sandaa, R.-A., Short, S.M. & Schroeder, D.C. (2010). Fingerprinting aquatic virus communities. MAVE, 9–18. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., et al. (2009). Introducing mothur: Open-Source, Platform-Independent, Community-Supported Software for Describing and Comparing Microbial Communities. Appl. Environ. Microbiol., 75, 7537– 7541. Schoenfeld, T., Patterson, M., Richardson, P.M., Wommack, K.E., Young, M. & Mead, D. (2008). Assembly of Viral Metagenomes from Yellowstone Hot Springs. Appl. Environ. Microbiol., 74, 4164–4174. Schwalbach, M.S., Hewson, I. & Fuhrman, J.A. (2004). Viral effects on bacterial community composition in marine plankton microcosms. Aquat. Microb. Ecol., 34, 117–127. Shade, A. & Handelsman, J. (2012). Beyond the Venn diagram: the hunt for a core microbiome. Environ Microbiol, 14, 4–12. Shannon, P. (2003). Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Research, 13, 2498–2504. Sharon, I., Battchikova, N., Aro, E.-M., Giglione, C., Meinnel, T., Glaser, F., et al. (2011). Comparative metagenomics of microbial traits within oceanic viral communities. ISME J, 5, 1178–1190. Sharon, I., Tzahor, S., Williamson, S.J., Shmoish, M., Man-Aharonovich, D., Rusch, D.B., et al. (2007). Viral photosynthetic reaction center genes and transcripts in the marine environment. ISME J, 1, 492–501. Sher, D., Thompson, J.W., Kashtan, N., Croal, L. & Chisholm, S.W. (2011). Response of Prochlorococcus ecotypes to co-culture with diverse marine bacteria. ISME J, 5, 1125–1132. Sherr, E. & Sherr, B.F. (1988). Role of microbes in pelagic food webs: a revised concept. Limnol Oceanogr., 33, 1225–1227. Sherr, E. & Sherr, B.F. (2002). Significance of predation by protists in aquatic microbial food webs. Antonie Leeuwenhoek, 81, 293–308. Short, C.M. & Suttle, C.A. (2005). Nearly identical bacteriophage structural gene sequences are widely distributed in both marine and freshwater environments. Appl. Environ. Microbiol., 71, 480–486. Short, C.M., Rusanova, O. & Short, S.M. (2011). Quantification of virus genes provides evidence for seed-bank populations of phycodnaviruses in Lake Ontario, Canada. ISME J, 5, 810–821. Short, S.M. & Short, C.M. (2009). Quantitative PCR reveals transient and persistent algal viruses in Lake Ontario, Canada. Environ Microbiol, 11, 2639–2648. Short, S.M. & Suttle, C.A. (2002). Sequence analysis of marine virus communities reveals that groups of related algal viruses are widely distributed in nature. Appl. Environ. Microbiol., 68, 1290–1296. 170 Short, S.M. & Suttle, C.A. (2003). Temporal dynamics of natural communities of marine algal viruses and eukaryotes. Aquat. Microb. Ecol., 32, 107–119. Simek, K., Pernthaler, J., Weinbauer, M., Hornák, K., Dolan, J.R., Nedoma, J., et al. (2001). Changes in Bacterial Community Composition and Dynamics and Viral Mortality Rates Associated with Enhanced Flagellate Grazing in a Mesoeutrophic Reservoir. Appl. Environ. Microbiol., 67, 2723–2733. Simpson, G.L. (2009). cocorresp: Co-correspondence analysis ordination methods (R package version 0.1-9). R package version 0.1-9. Smoot, M.E., Ono, K., Ruscheinski, J., Wang, P.-L. & Ideker, T. (2011). Cytoscape 2.8: new features for data integration and network visualization. Bioinformatics, 27, 431–432. Sole, R.V. & Montoya, M. (2001). Complexity and fragility in ecological networks. Proc R Soc Lond B, 268, 2039–2045. Staniewski, M.A., Short, C.M. & Short, S.M. (2012). Contrasting Community versus Population- Based Estimates of Grazing and Virus-Induced Mortality of Phytoplankton. Microb Ecol. Steele, J.A., Countway, P.D., Xia, L., Vigil, P.D., Beman, J.M., Kim, D.Y., et al. (2011). Marine bacterial, archaeal and protistan association networks reveal ecological linkages. ISME J, 5, 1414–1425. Storey, J.D. (2002). A direct approach to false discovery rates. J Roy Stat Soc B Met, 64, 479– 498. Steward, G.F. & Culley, A.I. (2010). Extraction and purification of nucleic acids from viruses. MAVE, 154–165. Steward, G.F. & Preston, C.M. (2011). Analysis of a viral metagenomic library from 200 m depth in Monterey Bay, California constructed by direct shotgun cloning. Virol J, 8, 287. Steward, G.F., Montiel, J.L. & Azam, F. (2000). Genome size distributions indicate variability and similarities among marine viral assemblages from diverse environments. Limnol Oceanogr., 45, 1697–1706. Sullivan, M.B., Coleman, M.L., Quinlivan, V., Rosenkrantz, J.E., DeFrancesco, A.S., Tan, G., et al. (2008). Portal protein diversity and phage ecology. Environ Microbiol, 10, 2810–2823. Sullivan, M.B., Huang, K.H., Ignacio-Espinoza, J.C., Berlin, A.M., Kelly, L., Weigele, P.R., et al. (2010). Genomic analysis of oceanic cyanobacterial myoviruses compared with T4-like myoviruses from diverse hosts and environments. Environ Microbiol, 12, 3035–3056. Sullivan, M.B., Lindell, D., Lee, J.A., Thompson, L.R., Bielawski, J.P. & Chisholm, S.W. (2006). Prevalence and Evolution of Core Photosystem II Genes in Marine Cyanobacterial Viruses and Their Hosts. Plos Biol, 4, e234. Sullivan, M.B., Waterbury, J.B. & Chisholm, S.W. (2003). Cyanophages infecting the oceanic cyanobacterium Prochlorococcus. Nature, 424, 1047–1051. 171 Suttle, C.A. (1994). Dynamics and Distribution of Cyanophages and Their Effect on Marine Synechococcus spp. Appl. Environ. Microbiol. Suttle, C.A. (2005). Viruses in the sea. Nature, 437, 356–361. Suttle, C.A. (2007). Marine viruses — major players in the global ecosystem. Nat Rev Micro, 5, 801–812. Tai, V. & Palenik, B. (2009). Temporal variation of Synechococcus clades at a coastal Pacific Ocean monitoring site. ISME J, 3, 903–915. Thingstad, T.F. (2000). Elements of a theory for the mechanisms controlling abundance, diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol Oceanogr., 1320–1328. Thingstad, T.F. & Cuevas, L. (2010). Nutrient pathways through the microbial food web: principles and predictability discussed, based on five different experiments. Aquat. Microb. Ecol., 61, 249251–249262. Thingstad, T.F. & Lignell, R. (1997). Theoretical models for the control of bacterial growth rate, abundance, diversity and carbon demand. Aquat. Microb. Ecol., 13, 19–27. Thurber, R.V. (2009). Current insights into phage biodiversity and biogeography. Curr. Opin. Microbiol, 12, 582–587. Thurber, R., Haynes, M., Breitbart, M. & Wegley, L. (2009). Laboratory procedures to generate viral metagenomes : Abstract : Nature Protocols. Nat Protoc. Treusch, A.H., Vergin, K.L., Finlay, L.A., Donatz, M.G., Burton, R.M., Carlson, C.A., et al. (2009). Seasonality and vertical structure of microbial communities in an ocean gyre. ISME J, 3, 1148–1163. Vigil, P., Countway, P., Rose, J., Lonsdale, D., Gobler, C. & Caron, D. (2009). Rapid shifts in dominant taxa among microbial eukaryotes in estuarine ecosystems. Aquat. Microb. Ecol., 54, 83–100. Wang, K. & Chen, F. (2004). Genetic diversity and population dynamics of cyanophage communities in the Chesapeake Bay. Aquat. Microb. Ecol., 34, 105–116. Wang, K., Wommack, K.E. & Chen, F. (2011). Abundance and Distribution of Synechococcus spp. and Cyanophages in the Chesapeake Bay. Appl. Environ. Microbiol., 77, 7459–7468. Waterbury, J.B. & Valois, F.W. (1993). Resistance to co-occurring phages enables marine Synechococcus communities to coexist with cyanophages abundant in seawater. Appl. Environ. Microbiol., 59, 3393–3399. Weigele, P.R., Pope, W.H., Pedulla, M.L., Houtz, J.M., Smith, A.L., Conway, J.F., et al. (2007). Genomic and structural analysis of Syn9, a cyanophage infecting marine Prochlorococcus and Synechococcus. Environ Microbiol, 9, 1675–1695. 172 Weinbauer, M.G. & Rassoulzadegan, F. (2003). Are viruses driving microbial diversification and diversity? Environ Microbiol, 6, 1–11. Weinbauer, M.G., Brettar, I. & Höfle, M.G. (2003). Lysogeny and virus-induced mortality of bacterioplankton in surface, deep, and anoxic marine waters. Limnol Oceanogr., 48, 1457– 1465. Weinbauer, M.G., Bonilla-Findji, O., Chan, A.M., Dolan, J.R., Short, S.M., Simek, K., et al. (2011). Synechococcus growth in the ocean may depend on the lysis of heterotrophic bacteria. J. Plankton Res., 33, 1465–1476. Weinbauer, M.G., Christaki, U., Nedoma, J. & Simek, K. (2003). Comparing the effects of resource enrichment and grazing on viral production in a meso-eutrophic reservoir. Aquat. Microb. Ecol., 31, 137–144. Weinbauer, M.G., Hornák, K., Jezbera, J., Nedoma, J., Dolan, J.R. & Šimek, K. (2007). Synergistic and antagonistic effects of viral lysis and protistan grazing on bacterial biomass, production and diversity. Environ Microbiol, 9, 777–788. Wilhelm, S.W. & Suttle, C.A. (1999). Viruses and nutrient cycles in the sea. Bioscience, 49, 781– 788. Williamson, S.J., Cary, S.C., Williamson, K.E., Helton, R.R., Bench, S.R., Winget, D., et al. (2008). Lysogenic virus–host interactions predominate at deep-sea diffuse-flow hydrothermal vents. ISME J, 2, 1112–1121. Williamson, S.J., Rusch, D.B., Yooseph, S., Halpern, A.L., Heidelberg, K.B., Glass, J.I., et al. (2008b). The Sorcerer II Global Ocean Sampling Expedition: Metagenomic Characterization of Viruses within Aquatic Microbial Samples. PLoS ONE, 3, e1456. Wilson, W.H., Fuller, N.J., Joint, I.R. & Mann, N. (2000). Analysis of cyanophage diversity in the marine environment using denaturing gradient gel electrophoresis. Microbial Biosystems: New Frontiers Proceedings of the 8th International Symposium on Microbial Ecology. Atlantic Canada Society for Microbial Ecology, 565–570. Winter, C., Smit, A., Herndl, G.J. & Weinbauer, M.G. (2004). Impact of virioplankton on archaeal and bacterial community richness as assessed in seawater batch cultures. Appl. Environ. Microbiol., 70, 804–813. Wommack, K.E. & Colwell, R.R. (2000). Virioplankton: viruses in aquatic ecosystems. Microbiol. Mol. Biol. Rev., 64, 69–114. Wommack, K.E., Hill, R., Kessel, M., Russek-Cohen, E. & Colwell, R.R. (1992). Distribution of viruses in the Chesapeake Bay. Appl. Environ. Microbiol., 58, 2965–2970. Wommack, K.E., Ravel, J., Hill, R., Chun, J. & Colwell, R.R. (1999). Population dynamics of chesapeake bay virioplankton: total-community analysis by pulsed-field gel electrophoresis. Appl. Environ. Microbiol., 65, 231–240. 173 Worden, A.Z. & Binder, B.J. (2003). Application of dilution experiments for measuring growth and mortality rates among Prochlorococcus and Synechococcus populations in oligotrophic environments. Aquat. Microb. Ecol., 30, 159–174. Xia, L.C., Steele, J.A., Cram, J.A., Cardon, Z.G., Simmons, S.L., Vallino, J.J., et al. (2011). Extended local similarity analysis (eLSA) of microbial community and other time series data with replicates. BMC Syst. Biol., 5, S15. Yoon, J.S. & Jung, W.H. (2011) "A GPU-accelerated bioinformatics application for large-scale protein interaction networks", APBC poster presentation. Yooseph, S., Sutton, G., Rusch, D.B., Halpern, A.L., Williamson, S.J., Remington, K., et al. (2007). The Sorcerer II Global Ocean Sampling Expedition: Expanding the Universe of Protein Families. Plos Biol, 5, e16. Yilmaz, P., Iversen, M.H., Hankeln, W., Kottmann, R., Quast, C. & Glöckner, F.O. (2012). Ecological structuring of bacterial and archaeal taxa in surface ocean waters. FEMS Microbiol Ecol, n/a–n/a. Zhang, R., Weinbauer, M.G. & Qian, P.-Y. (2007). Viruses and flagellates sustain apparent richness and reduce biomass accumulation of bacterioplankton in coastal marine waters. Environ Microbiol, 9, 3008–3018. Zhang, W.J., Aldrich, P.R., Tacutu, R., Budovsky, A., Yanai, H., Wolfson, M., et al. (2011). Constructing ecological interaction networks by correlation analysis: hints from community sampling. Network Biology, 1, 81–98. Zhong, Y., Chen, F., Wilhelm, S.W., Poorvin, L. & Hodson, R.E. (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, 1576–1584. Zinger, L., Amaral-Zettler, L.A., Fuhrman, J.A., Horner-Devine, M.C., Huse, S.M., Welch, D.B.M., et al. (2011). Global Patterns of Bacterial Beta-Diversity in Seafloor and Seawater Ecosystems. PLoS ONE, 6, e24570. Zinser, E.R., Lindell, D., Johnson, Z.I., Futschik, M.E., Steglich, C., Coleman, M.L., et al. (2009). Choreography of the Transcriptome, Photophysiology, and Cell Cycle of a Minimal Photoautotroph, Prochlorococcus. PLoS ONE, 4, e5135.
Abstract (if available)
Abstract
Our ability to define and quantify individual microbe-microbe interactions with traditional community ecology principles is still in development yet these relationships are key to understanding microbial roles in the ocean. Application of molecular methods, such as community fingerprinting and metagenomics have added significant insight into marine viral and microbial ecology in general over the last decade. Great progress has been made in identifying what organisms or taxonomic groups are present, their abundances and their genomic context. Metagenomic and genomic studies have highlighted the phylogenetic and functional diversity of viruses in the ocean. However, aside from work on the limited number of cultured organisms, our understanding of microbe-microbe relationships is often broad and does not typically lead to systematic identification of multiple relationships from the environment at the same time. The ease and feasibility of these molecular tools are increasingly an essential part of the microbial ecologists’ toolbox. Continuing research with microbial association networks can lead to a better understanding of how individual members of a diverse microbial community relate to one another and will ultimately facilitate predictions on how resilient or susceptible the microbial community may be to their future ocean climate. ❧ The microbial community at the San Pedro Ocean Time-series was characterized in great detail – with a specific focus on viral and bacterial communities and inter-microbe relationships as part of the USC Microbial Observatory program. First, ten-years of observations of bacterial communities in the euphotic zone revealed seasonal and annual trends in the surface ocean (Chapter 2). There were only a few dissimilarities between the surface and deep chlorophyll maximum (DCM) depths in terms of the microbial constituents
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Patterns of molecular microbial activity across time and biomes
PDF
Changes in the community composition of marine microbial eukaryotes across multiple temporal scales of measurement
PDF
Dynamics of marine bacterial communities from surface to bottom and the factors controling them
PDF
Spatial and temporal dynamics of marine microbial communities and their diazotrophs in the Southern California Bight
PDF
Genetic characterization of microbial eukaryotic diversity and metabolic potential
PDF
Ecological implications of daily-to-weekly dynamics of marine bacteria, archaea, viruses, and phytoplankton
PDF
Annual pattern and response of the bacterial and microbial eukaryotic communities in an aquatic ecosystem restructured by disturbance
PDF
Marine bacterioplankton biogeography over short to medium spatio-temporal scales
PDF
Microbial ecology in the deep terrestrial biosphere: a geochemical, metagenomic and culture-based approach
PDF
Temporal variability of marine archaea across the water column at SPOT
PDF
Plant-microbial interactions in mangrove sediments under different nutrient conditions
PDF
Spatial and temporal investigations of protistan grazing impact on microbial communities in marine ecosystems
PDF
Ecological patterns of free-living and particle-associated prokaryotes, protists, and viruses at the San Pedro Ocean Time-series between 2005 and 2018
PDF
Developing statistical and algorithmic methods for shotgun metagenomics and time series analysis
PDF
Characterizing protistan diversity and quantifying protistan grazing in the North Pacific Subtropical Gyre
PDF
Using sequencing techniques to explore the microbial communities associated with ferromanganese nodules and sediment from the South Pacific gyre
PDF
Enhancing recovery of understudied and uncultured lineages from metagenomes
PDF
Unexplored microbial communities in marine sediment porewater
PDF
Marine protistan diversity, spatiotemporal dynamics, and physiological responses to environmental cues
PDF
The influence of diet on the gut microbial community in ring-tailed lemurs and Verreaux’s sifaka
Asset Metadata
Creator
Chow, Cheryl-Emiliane T.
(author)
Core Title
Microbe to microbe: monthly microbial community dynamics and interactions at the San Pedro Ocean Time-series
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Biology
Publication Date
07/26/2013
Defense Date
05/24/2012
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
association networks,bacteria,marine microbiology,microbial ecology,OAI-PMH Harvest,time-series,virus
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Fuhrman, Jed A. (
committee chair
), Caron, David A. (
committee member
), Heidelberg, John F. (
committee member
), Roberts, Richard W. (
committee member
), Webb, Eric A. (
committee member
)
Creator Email
cherylec@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-69109
Unique identifier
UC11288745
Identifier
usctheses-c3-69109 (legacy record id)
Legacy Identifier
etd-ChowCheryl-1027.pdf
Dmrecord
69109
Document Type
Dissertation
Rights
Chow, Cheryl-Emiliane T.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
association networks
bacteria
marine microbiology
microbial ecology
time-series
virus