It is likely that all living organisms can be infected by one or more viruses. One of the latest higher taxa to be converted from ‘no characterized viruses’ to ‘well characterized viruses’ are the diatoms (Bacillariophyceae, Heterokontophyta) with the recent publication of three papers characterizing an ssRNA and a ssDNA virus from two genera (Chaetoceros and Rhizosolenia) of marine planktonic diatom (Nagasaki et al., 2004, 2005; Bettarel et al., 2005). It would have been strange if viruses had not been able to exploit the dominant, in terms of global primary production, photosynthetic organisms in the ocean (assimilating perhaps as much as 20 Pg inorganic C into organic C per year), despite the less than completely convincing arguments assembled by Raven & Waite (2004) as to possible anti-viral defences unique to diatoms.

Viruses are the most abundant members of marine ecosystems and play an enormous role in ocean processes through their interactions with all types of marine organisms. This short review provides examples of the dramatic increase in our knowledge of the diversity of marine viruses as pathogens of bacteria, protists, molluscs, crustaceans, cnidaria, reptiles, fish and mammals. Several examples are provided showing evidence of evolution of new strains, changes in virulence, and transfer of viruses between ecosystems. The natural and anthropogenic causes of these shifts are discussed. Despite considerable advances in recent years, knowledge of the importance of viruses in many important groups of marine organisms is lacking or incomplete. Suggestions for future investigations necessary to understand the dynamics of biogeochemical processes and the impacts of disease in our oceans are proposed.

Dinoflagellates (Dinophyceae) are considered to be one of the most abundant and diverse groups of phytoplankton; however, the viral impact on dinoflagellates was not studied until recently. This review shows the present information concerning the viruses infecting dinoflagellates and the ecology relationships between the host and the virus. So far, two viruses have been isolated and characterized: a large DNA virus (HcV: Heterocapsa circularisquama virus) and a small RNA virus (HcRNAV: H. circularisquama RNA virus); both of which are infectious to the harmful bloom-forming dinoflagellate H. circularisquama.

In the present review, we mainly discuss the relationship between HcRNAV and H. circularisquama from the viewpoint of physiology, ecology and genetics. It will help clarify the viral impact on dinoflagellate populations in marine environments to understand the host/parasite ecology.

We report the first complete genome sequence of the marine diatom-infecting, positive-sense single-stranded RNA (ssRNA) virus, Rhizosolenia setigera RNA virus (RsRNAV). The genome is 8877 nucleotides (nt), polyadenylated, lacking a cap structure, and has two large open reading frames (ORFs): ORF-1 (4818 nt), a polyprotein gene coding for replicases, e.g. RNA helicase, RNA-dependent RNA polymerase (RdRp); and ORF-2 (2883 nt), a polyprotein gene coding for structural proteins. The ORFs are separated by a 323 nt intergenic region (IGR), flanked by a 624 nt 5′-untranslated region (UTR) and a 229 nt 3′-UTR. The deduced amino acid sequences for ORF-1 and ORF-2 respectively show considerable similarities to the non-structural and structural proteins of a marine raphidophyte-infecting virus HaRNAV (Heterosigma akashiwo RNA virus). Phylogenetic analyses of concatenated amino acid sequences of RNA helicase and RdRp domains supported the monophyly of RsRNAV, HaRNAV and a marine protist-infecting virus SssRNAV (Schizochytrium single-stranded RNA virus) with moderate bootstrap values of 79–83%, but not at the family level, whilst their monophyly was only weakly supported (50–55%) in the phylogenetic tree based on RdRp whole domain. As a result, comparison of the genome organization and sequence suggests RsRNAV is not a member of any currently defined virus family. In the RdRp tree, the positive-sense ssRNA viruses infecting Stramenopiles (RsRNAV, HaRNAV and SssRNAV) and Alveolata (HcRNAV (Heterocapsa circularisquama RNA virus)) were categorized into phylogenetically distant clades, which suggests a host/virus coevolution. Our study supports the hypothesis that a diverse array of ssRNA viruses exists in marine environments.

A new virus that infects the harmful algal bloom-forming microalga Phaeocystis globosa was isolated from surface water in the English Channel off the coast of Plymouth, UK, in May 2001. Phylogenetic analysis of the DNA polymerase gene revealed the virus isolate, designated PgV-102P, belongs to the family Phycodnaviridae, a group of large double-stranded DNA viruses known to infect algae. Basic characterization of PgV-102P revealed it was a lytic virus with a relatively slow culture lysis period of 10-days. The genome size (176 kbp) and capsid diameter (98 nm) of PgV-102P fall at the bottom end of the range expected for phycodnaviruses. Interestingly, PgV-102P did not cluster with other P. globosa viruses; instead, it was more closely related to other prymnesioviruses that infect the marine prymnesiophyte Chrysochromulina brevifilum. We discuss the effectiveness of DNA polymerase as a diagnostic marker. Although it is ideal for determining what family or even genus an algal virus belongs to, it is clear that the DNA polymerase gene does not have sufficient resolution when looking for relationships within algal virus genera.

Phytoplankton:virus interactions are important factors in aquatic nutrient cycling and community succession. The number of viral progeny resulting from an infection of a cell critically influences the propagation of infection and concomitantly the dynamics of phytoplankton populations. Host nucleotide content may be the resource limiting viral particle assembly. We present evidence for a strong linear correlation between measured viral burst sizes and viral burst sizes predicted from the host DNA content divided by the viral genome size, across a diversity of phytoplankton:viral pairs. An analysis of genome sizes therefore supports predictions of taxon-specific phytoplankton population density thresholds beyond which viral proliferation can trim populations or terminate phytoplankton blooms. We present corollaries showing that host:virus interactions may place evolutionary pressure towards genome reduction of both phytoplankton hosts and their viruses.

Cyanophage abundance has been shown to fluctuate over long timescales and with depth, but little is known about how it varies over short timescales. Previous short-term studies have relied on counting total virus numbers and therefore the phages which infect cyanobacteria cannot be distinguished from the total count.

In this study, an isolation-based approach was used to determine cyanophage abundance from water samples collected over a depth profile for a 24 h period from the Indian Ocean. Samples were used to infect Synechococcus sp. WH7803 and the number of plaque forming units (pfu) at each time point and depth were counted. At 10 m phage numbers were similar for most time-points, but there was a distinct peak in abundance at 0100 hours. Phage numbers were lower at 25 m and 50 m and did not show such strong temporal variation. No phages were found below this depth. Therefore, we conclude that only the abundance of phages in surface waters showed a clear temporal pattern over a short timescale. Fifty phages from a range of depths and time points were isolated and purified. The molecular diversity of these phages was estimated using a section of the phage-encoded psbD gene and the results from a phylogenetic analysis do not suggest that phages from the deeper waters form a distinct subgroup.

The aim of this study was to determine the abundance of cyanophages over an annual cycle in the Red Sea from the period April 1999 to December 1999 at a range of depths. Cyanophage numbers from 71 water samples were determined by the use of plaque assays using four different Synechococcus strains. The results indicate that cyanophage are found throughout the water column from surface waters to depths of 150 m, with a discrete maximum in the number of cyanophages in the summer months of July, August and September at a depth of 30 m. Eighty-seven cyanophages were isolated and characterized in terms of host range, genome size and possession of a myoviral portal vertex gene. Cyanophages were found to infect multiple strains of Synechococcus from different phylogenetic clades. The genome sizes of cyanophages were also found to be bigger than previously estimated.

The unicellular cyanobacterial species Synechococcus and Prochlorococcus are known to be vital components of marine ecosystems, especially in the vast oligotrophic areas. Lytic cyanophages infecting unicellular phytoplankton are prevalent and have been demonstrated to act as important constraints on community composition contributing to the seasonal succession in genotypes. Lysogeny in Synechococcus has been documented experimentally in natural environments by prophage induction. At this time it is completely unknown how prevalent lysogeny is among Synechococcus populations. This study was performed to document important features such as size, morphology and the incidence of the T4-like capsid portal protein gene (g20) in a group of lytic Synechococcus cyanophages (35 isolates) isolated from the Gulf of Mexico. A group of Synechococcus isolates (24 strains) were isolated concurrently to investigate the virulence and cross-infectivity of the lytic cyanophages and to determine the frequency of lysogeny by detection of inducible prophage. The host range of the cyanophages toward these Synechococcus strains ranged from 1 of 25 (host of isolation only) to 17 of 25 (68%). Of the 35 cyanophage isolates the large majority were myoviruses (94%) and only two (6%) were of the podovirus type. The expected polymerase chain reaction product for g20 was detected in 20 of the phages (63%). The presence of a detectable g20 was associated with low-infectivity cyanophages at the 90% confidence interval. The Synechococcus strains varied in their resistance to lytic infection from 11% to resistance to all of the phage isolates utilized in testing. The prevalence of inducible prophage-like particles was determined in the Synechococcus strains using mitomycin C and enumerating viruses by epifluorescence microscopy. A statistically significant increase in viruses was detected in 11 of the strains (46%) in response to mitomycin C. There was no observed relationship between the occurrence of prophage induction in the Synechococcus isolates and their resistance to lytic infection. One putative lysogen was induced by continuous high light and contained a prophage-like particle with a single-stranded DNA (ssDNA) genome. Such a prophage-like particle is unlike any prophage described to date, implying that the process of lysogeny is unique in certain marine Synechococcus strains.

A collection of 17 cyanophage isolates able to infect the heterocystous, filamentous cyanobacterium Nodularia spumigena has been established from the Baltic Sea. These cyanophages have been characterized based on their morphology, cross infectivity and genetic structure. Short fragments (450 bp) of the gene encoding the major capsid protein (g23) were amplified and sequenced from several isolates, and the encoded protein was found to be 99% identical across all the N. spumigena-specific cyanophages tested. These results suggest that the Nodularia-specific cyanophages are very closely related. However, these cyanophages were found to be diverse in terms of their morphology and host range. Cyanophages belonging to two families within the order Caudovirales, Myoviridae and Siphoviridae, were included in the collection of isolates. The cyanophage particles are large in comparison with cyanophages previously isolated from the marine environment, with the largest capsid measuring 127×122×888 nm. Host ranges of the cyanophage isolates varied, some being able to infect up to five genotypically distinct strains of Nodularia spumigena, while others were very specific, infecting only one strain. We conclude that Nodularia-specific cyanophages form a diverse community in surface waters during summer and autumn months and that they may play a role both in the transfer of genetic information between Nodularia lineages and in promoting changes in the genetic structure of the host population.

Laboratory experiments with whole water-columns from shallow, eutrophic lakes repeatedly showed collapse of the predominant filamentous cyanobacteria. The collapse could be due to viral activity, from the evidence of electron microscopy of infected cyanobacterial cells and observed dynamics of virus-like particles. Burst-size effects on single-host single-virus dynamics was modelled for nutrient-replete growth of the cyanobacteria and fixed viral decay rate in the water column. The model combined previously published equations for nutrient-replete cyanobacterial growth and virus–host relationship. According to the model results, burst sizes greater than 200 to 400 virions per cell would result in host extinction, whereas lower numbers would allow coexistence, and even stable population densities of host and virus. High-nutrient status of the host cells might accommodate a large burst size. The ecological implication could be that burst-size increase accompanying a transition from phosphorus to light-limited cyanobacterial growth might destabilize the virus–host interaction and result in the population collapse observed in the experiments.

Seasonal investigation of virus dynamics by flow cytometry was conducted in the Yangtze river estuarine area in April, August, November 2002 and February 2003, and a supplemental investigation in the inner estuary and downstream of the river was conducted in October 2005. The majority of the total viral abundance was bacteriophage and only 5.4% of the total was algal virus. Total viral abundance varied with season and location, ranging from 6.75×105–1.68×107 particles/ml, and the virus:bacterium ratio (VBR) ranged from 1.52 to 72.02 with a mean of 8.7. In the present study, viral abundance peaked in both the summer and the winter, unlike the typical seasonal pattern reported in the literature, in which viral abundance peaks in the summer when bacterial hosts are also at their most abundant. However, the driving forces for the two peaks reported here were totally different, the summer viral abundance peak coupled with the development of bacterial hosts which were controlled largely by temperature year-round and by trophic state occasionally, while the winter one seemed to be multi-factor controlled. The host-phage interaction was no longer predominant in control of the winter viral abundance as bacterial abundance was lowest in this season. The winter low temperature would help maintain a high viral abundance as high temperatures might increase viral inactivation and viral decay; the VBR peak values actually occurred in the winter. More importantly, the high virus-containing freshwater discharge in winter due to a higher proportion of anthropogenic sewage relative to low natural flooding in winter run-off, turned out to be the first factor contributing to the high winter viral abundance and VBR values. In addition, the variation of intrusion of warm and relatively oligotrophic water from oceanic currents played a role alternating the distribution patterns of temperature, salinity and trophic conditions and consequently the distribution patterns of virus and bacteria seasonally and spatially. Dynamics of virus in the Yangtze river estuarine area is thus characterized by distinct seasonal and spatial variations due to natural forcing and by pronounced alternation of the regular patterns due to anthropogenic impacts.

The microscale spatial distributions of viruses were investigated in three contrasting environments including oligotrophic open ocean, eutrophic coastal and estuarine habitats. The abundances of two discrete populations of both viruses and heterotrophic bacteria were measured at spatial resolutions of between 1 and 5 cm using purpose-designed microscale sampling equipment and flow cytometric sample analysis. Within open water samples, virus distributions were characterized by non-normal distributions and by ‘hotspots’ in abundance where concentrations varied by up to 17-fold. In contrast to patterns generally observed at larger spatiotemporal scales, there was no correlation between bacterial and viral abundance or correspondence between bacteria and virus hotspots within these samples. Consequently, strong hotspots and gradients in the virus:bacteria ratio (VBR) were also apparent within samples. Within vertical profiles taken from above the sediment–water interface within a temperate mangrove estuary, distributions of planktonic viruses were characterized by gradients in abundance, with highest concentrations observed within the 1–2 cm immediately above the sediment surface, and virus distributions were correlated to bacterial abundance (P<0.01). The patterns observed in these contrasting habitats indicate that microscale patchiness of virus abundance may be a common feature of the marine environment. This form of heterogeneity may have important implications for virus–host dynamics and subsequently influence microbial trophodynamics and nutrient cycling in the ocean.

Using flow cytometry, two distinct populations of virus-like particles (VLP) and heterotrophic bacteria were defined within the 12 cm water layer immediately overlying healthy, diseased and dead acroporid corals. Bacterial abundances were similar in overlying water for all coral types, however, VLP were 30% higher above diseased corals than healthy or dead corals. Mean virus to bacteria ratios (VBR) were up to 30% higher above diseased corals than above healthy or dead coral or in distant water. Concomitant with increasing VLP concentrations within 5 cm of coral surfaces, VBR distributions were generally highest above healthy and diseased coral and depressed above dead coral. These results suggest fundamental shifts in the VLP and bacterial community in water associated with diseased corals.

Plasmids and phages influence bacterial phenotype and may serve as vectors for transferring genes between bacteria. In the present study, we examined 130 marine bacterioplankton isolates for the presence of plasmids and prophages. Samples were obtained in spring, summer and autumn in the Baltic Sea proper. Plasmids and inducible prophages were found in 19% and 28% of the isolates, respectively. During spring, plasmids and prophages were 41–55% and 30% more common compared to the summer and autumn measurements and prevalence varied up to five-fold between bacterial phylogenetic groups, with the highest plasmid prevalence found in Bacteriodetes (41%), and lysogeny being common in α-, β-, and γ-Proteobacteria (32–50%). Plasmid genome sizes ranged from 1.5–15 kb with most in the 2.1–4.0 kb size-range. No plasmids showed identity to the broad-host-range incompatibility groups N and P. Phage genomes ranged in size from 8–87 kb, with 57% being 35–45 kb in size. Strain typing of phages with similar genome sizes by means of DP-RAPD (degenerated primer randomly amplified polymorphic DNA) showed that all were different (except two that were not resolved). In PFGE (pulsed-field gel electrophoresis) 34% of the lysates produced multiple bands. Transmission electron microscopy suggested that these originated from several phage morphotypes indicating that polylysogeny is common. The widespread distribution of small cryptic plasmids as well as of lysogeny and polylysogeny in Baltic Sea bacterioplankton may have important implications for bacterial phenotype and for lateral gene transfer; hence, the ecological significance of these vectors in marine environments requires further study.

This study examined the relationship between viral infection and the richness, diversity and composition of bacterial assemblages in the water column. Viruses were enriched by ultrafiltration, added to water column incubation experiments at 15 locations in the North Atlantic, North Pacific, Gulf of Mexico and Southern California. In a separate experiment, viruses were removed from bacterioplankton by diafiltration at the San Pedro Ocean Time Series Station. Bacterial assemblage composition was observed using a high throughput and sensitive molecular fingerprinting analysis, automated rRNA intergenic spacer analysis (ARISA). Diazotrophs were used as a model functional group to represent rare organisms hypothesized to benefit from viral activity, and their richness and diversity was determined by terminal restriction fragment length polymorphism of a nitrogenase gene fragment (nifH). The enrichment and removal experiments demonstrated mixed impacts of viral pressure upon bacterial communities, and we observed significant effects of viruses on several microbial parameters in all but two experiments. However, there was no consistent response of viral enrichment on total bacterial and diazotroph assemblages at stations with similar environmental conditions, suggesting that untested variables, small spatial scale factors, or stochastic processes influence the outcome of viral activities. Across all experiments, the relative abundance of the more common operational taxonomic units (OTUs) in fingerprints were not significantly impacted compared to the abundance of rare OTUs. These data indicate that viruses may have significant influence upon community structure of bacterioplankton; however, effects were not consistent between sampling locations nor water masses.

Viruses are hypothesized to cause enhanced diversity in bacterial communities by regulating the outcome of intertaxon competition. However, concomitant documentation of viral and bacterial assemblage composition in oligotrophic waters are rare, particularly in situ over time, and there is almost no information on the temporal variability in virioplankton assemblage composition in oligotrophic water masses. Assemblage composition of viruses (via pulsed-field gel electrophoresis, PFGE) and bacteria (via automated rRNA intergenic spacer analysis, ARISA) was compared during surface lagrangian drifter deployments in the oligotrophic Gulf of Mexico during summer 2001, 2002, and 2003. In vertical profile, viruses and bacteria both had maximum abundances in surface waters, which decreased with depth; however, the richness of their assemblages was not significantly different between depths, suggesting independence of biomass and diversity. Viral assemblages changed rapidly (0.17–0.32 Jaccard index d−1), which was similar to the rate of change in bacterial assemblages reported in surface waters. Patterns of viral and bacterial assemblage composition were significantly related (P<0.001, r=0.58 between node ranks), and both assemblages clustered primarily by year and then by depth. These cultivation-independent observations demonstrate relationships between viral and bacterial assemblages, which are dynamic in patches of open ocean water. Even at the relatively low phylogenetic resolution of the ARISA and PFGE methods, the results support the idea that viruses may influence the species composition of host assemblages.

Viral infection of bacteria causes release of dissolved organic matter (DOM), which is available for bacterial uptake. In aquatic environments, this virus-mediated transformation of living cells into dissolved and colloidal organic matter may be a quantitatively important process in the pelagic recycling of carbon and nutrients, but little is known about the amount, composition, or bioavailability of viral lysates. By using a model system of a marine bacterium (Cellulophaga sp.) and a virus specific to this bacterium, the present study provides a first quantification of the input of dissolved free and combined amino acids (DFAA and DCAA) and bacterial cell wall compounds following viral lysis. The DCAA constituted 51–86% of the total virus-mediated organic carbon release of 1087–1825 μg C l−1 (estimated biomass of the lysed bacteria), whereas DFAA and glucosamine each accounted for 2–3% of total lysate-C. The viral particles themselves constituted 4–6% of the released organic carbon, and altogether, the applied analyses thus identified 53–92% of the released lysates. Approximately 12% of the identified compounds were derived from bacterial cell wall peptidoglycan, including various D-isomers of DFAA and DCAA, glucosamine and diaminopimelic acid (DAPA). Although a portion of this cell wall material may have entered the pool of refractory material, a significant fraction of some peptidoglycan-derived components, e.g. 83% of the released D-DFAA, were removed from the dissolved phase during the last part of the incubations, suggesting that part of the cell wall material were utilized by the developing virus-resistant Cellulophaga population. Therefore, we suggest that virus-mediated DOM is a source of a variety of organic compounds, which contribute significantly to the pool of rapidly recycling material in the ocean.

Viral burst size (BS), i.e. the number of viruses released during cell lysis, is a critical parameter for assessing the ecological and biogeochemical role of viruses in aquatic systems. Burst size is typically estimated by enumerating the viral particles in bacteria using transmission electron microscopy. Here, we review the average BS reported for different aquatic systems, present several hypotheses on the control of the BS and evaluate whether there are relationships between BS and bacterial activity parameters across systems. Based on reports from a variety of different aquatic environments, we calculated a mean BS of 24 and 34 for marine and freshwater environments, respectively. Generally, the BS increased with the trophic status of the environment and with the percentage of infected cells in marine populations. When diel dynamics were investigated or averages from large-scale environments were used, BS was positively related to bacterial production but no trend was detectable across systems. The across systems' finding that BS was significantly related to the frequency of infected cells (FIC) could be due to co-infection or superinfection. At any given site, BS seems to be influenced by a number of factors such as the size of the host cell and the viruses, the metabolic activity of the host and phage and host diversity. Thus, based on the available data collected over the past two decades on a variety of aquatic systems, some relations between BS and bacterial variables were detectable.