Transitions between saline and fresh waters have been shown to be infrequent for microorganisms. Based on host-specific interactions, the presence of specific clades among hosts suggests the existence of freshwater-specific viral clades. Yet, little is known about the composition and diversity of the temperate freshwater viral communities, and even if freshwater lakes and marine waters harbor distinct clades for particular viral sub-families, this distinction remains to be demonstrated on a community scale.To help identify the characteristics and potential specificities of freshwater viral communities, such communities from two lakes differing by their ecological parameters were studied through metagenomics. Both the cluster richness and the species richness of the Lake Bourget virome were significantly higher that those of the Lake Pavin, highlighting a trend similar to the one observed for microorganisms (i.e. the specie richness observed in mesotrophic lakes is greater than the one observed in oligotrophic lakes). Using 29 previously published viromes, the cluster richness was shown to vary between different environment types and appeared significantly higher in marine ecosystems than in other biomes. Furthermore, significant genetic similarity between viral communities of related environments was highlighted as freshwater, marine and hypersaline environments were separated from each other despite the vast geographical distances between sample locations within each of these biomes. An automated phylogeny procedure was then applied to marker genes of the major families of single-stranded (Microviridae, Circoviridae, Nanoviridae) and double-stranded (Caudovirales) DNA viruses. These phylogenetic analyses all spotlighted a very broad diversity and previously unknown clades undetectable by PCR analysis, clades that gathered sequences from the two lakes. Thus, the two freshwater viromes appear closely related, despite the significant ecological differences between the two lakes. Furthermore, freshwater viral communities appear genetically distinct from other aquatic ecosystems, demonstrating the specificity of freshwater viruses at a community scale for the first time.

1. We performed three, 1-week in situ experiments in March-April (expt 1), May (expt 2) and August (expt 3) 2003 in order to assess protozoan and virus-induced mortality of heterotrophic bacteria in a French lake. Viral and bacterial abundances were obtained using flow cytometry (FCM) while protozoa were counted using epifluorescence microscopy (EFM). 2. A dilution approach, applied to pretreated grazer-free samples, allowed us to estimate that viral lysis could be responsible for 60% (expt 1), 35% (expt 2) and 52% (expt 3) of daily heterotrophic bacterial mortality. Flagellate (both mixotrophic and heterotrophic) grazing in untreated samples, was responsible for 56% (expt 1), 63% (expt 2) and 18% (expt 3) of daily heterotrophic bacteria removal. 3. These results therefore suggest that both viral lysis and flagellate grazing had a strong impact on bacterial mortality, and this impact varied seasonally. 4. From parallel transmission electron microscopy (TEM) analysis, we found that the burst size (i.e. the number of viruses potentially released per lysed cell) ranged from nine to 25 (expt 1), 10 to 35 (expt 2) and eight to 25 (expt 3). The percentage of infected heterotrophic bacteria was 5.7% (expt 1), 3.4% (expt 2) and 5.7% (expt 3) so that the calculated percentage of bacterial mortality induced by viruses was 6.3% (expt 1), 3.7% (expt 2) and 6.3% (expt 3). 5. It is clear that the dilution-FCM and TEM methods yielded different estimates of viral impact, although both methods revealed an increased impact of viruses during summer.