Seasonal dynamics and diversity of Antarctic marine viruses reveal a novel viral seascape

doi:10.21203/rs.3.rs-3778832/v1

Download PDF

Article

Seasonal dynamics and diversity of Antarctic marine viruses reveal a novel viral seascape

https://doi.org/10.21203/rs.3.rs-3778832/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 24 Oct, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

The Southern Ocean microbial ecosystem, with its pronounced seasonal shifts, is vulnerable to the impacts of climate change. Since viruses are key modulators of microbial abundance, diversity, and evolution, we need a better understanding of the effects of seasonality on the viruses in this region. Our comprehensive exploration of viral diversity in the Southern Ocean revealed a unique and largely uncharted viral landscape, of which 75% was previously unidentified in other oceanic areas. We uncovered novel viral taxa at high taxonomic ranks, expanding our understanding of crassphage, polinton-like virus, and virophage diversity. Our results indicate that Nucleocytoviricota viruses are an important and diverse group of Antarctic viruses, which regulate phytoplankton population dynamics. Our temporal analysis reveals complex seasonal patterns in viral populations, highlighting the interplay between bacteriophages, eukaryotic viruses, and their microbial hosts, whilst deepening our understanding of their roles in the world's most sensitive and rapidly changing ecosystem.

Biological sciences/Microbiology/Microbial communities/Metagenomics

Earth and environmental sciences/Ecology/Biooceanography/Microbial biooceanography

Biological sciences/Microbiology/Bacteriophages

Biological sciences/Microbiology/Virology/Metagenomics

Earth and environmental sciences/Planetary science/Cryospheric science

Viruses are prevalent and numerically dominant agents of microbial mortality, as well as drivers of biodiversity^1–3. They play important roles in shaping microbial communities and biogeochemical cycles in the oceans, where two-thirds of living biomass is microbial⁴. Understanding the ecological significance of viruses requires a comprehensive insight into their diversity. Viral metagenomics has revolutionized marine viral ecology^5–7, revealing viral community structure, biogeochemical roles, evolutionary trajectories, and taxonomic diversity^8–13. Still, key oceanic regions, such as the Southern Ocean, remain understudied¹⁴.

The Southern Ocean, pivotal in global ocean circulation¹⁵, absorbs over 40% of anthropogenic carbon dioxide emissions and hence, plays a crucial role in moderating climate change¹⁶. This process is significantly driven by microbial activity¹⁷, emphasizing the need for studying the impact of viruses on these microbial hosts. The short polar productive season results in pronounced temporal dynamics in (micro)biological activity and community composition^18–21. Substantial seasonal variation was also reported for viral lysis of Antarctic marine microorganisms^22–26. Yet, our understanding of the diversity and dynamics of the virus communities remains limited. The few studies that addressed their metagenomic characterization provide a restricted view due to a limited sample number and narrow sampling period^23,27–30.

The present study delves into the intricate seasonal variations in Antarctic marine dsDNA virus diversity. It was conducted with high temporal resolution (24 sampling days) over the full productive season in Marguerite Bay, located in the Western Antarctic Peninsula. We advanced and expanded the catalog of marine viral diversity by combining multiple virus detection and classification approaches tailored to specific viral taxa. Moreover, an improved understanding of seasonal variations in virus-host dynamics highlights the ecological importance of viruses in the Southern Ocean as drivers of diversity and succession. Therefore, this study establishes an essential foundation for understanding the dynamics of this region, which is particularly sensitive to global climate change.

Antarctic viral communities harbor a diverse range of taxa

To explore the dynamics of Antarctic marine viruses and interactions with their hosts, we generated paired metagenomes from viral (< 0.22µm and free-DNA digestion) and cellular (> 0.22µm) size fractions. Viral metagenomes, totaling 887 million reads, were co-assembled into 161,651 viral scaffolds (on average 35% of the reads per sample), of which 8,045 were mid- to high-quality (> 10 kb or > 70% CheckV³¹ completeness). The viral fraction metagenomes displayed clear enrichment in viral sequences and depletion in cellular markers (Supplementary Fig. 1). When comparing the mid- to high-quality viral scaffolds to the Global and Southern Ocean Viromes^8,29 (GOV and SOV), we found 75% of viral sequences to be novel at the species rank.

The most frequent and abundant viruses were the tailed dsDNA bacteriophages of the class Caudoviricetes (Fig. 1a-c), outnumbering other viruses by 53-fold and making up 75% of all scaffolds (Fig. 1b-c). We further categorized Caudoviricetes sequences to family rank using PhaGCN2³², except the order Crassvirales which was characterized using a targeted approach. Seasonal patterns in viral abundance could be discerned for various Caudoviricetes taxa, even more pronounced in the cellular fraction (Fig. 1c-d). For example, Autographiviridae family sequences were the most abundant (on average 11% viral fraction and 3.5% cellular fraction reads), and displayed increased virus productivity in summer, reflected by the higher abundance in the cellular fraction from mid-January to February (Fig. 1c-d). Furthermore, kyanoviruses illustrate different temporal occurrences in the viral and cellular metagenomes, with less reads in the virus fraction than in the cellular metagenome. This could result from seasonal dynamics in lytic and lysogenic viral infection, as the viral reads from the cellular fraction can reflect both actively replicating and lysogenic viral infection while the virus fraction is composed of free viral particles. Few viruses (0.36% reads or 279 scaffolds) were identified as archaeal viruses, which showed higher abundance in November and March (Fig. 1e) and were mostly classified as haloviruses (112 scaffolds).

Abundant eukaryotic viruses belonged to the phylum Nucleocytoviricota (originally nucleocytoplasmic large DNA viruses), the classes Polintoviricetes (polinton-like viruses or PLVs) and Maveriviricetes (virophages), and the family Circoviridae (ssDNA; Fig. 1b-c,f). The Nucleocytoviricota viruses (NCVs) are the most abundant eukaryotic viruses, accounting for 12.4% and 1.6% of the cellular and viral fraction reads, respectively. NCV abundance in the cellular fraction spiked in March and December 2018, accounting for up to 27% and 77% of viral reads (respectively), implying heightened NCV production.

PLVs form an important fraction of the viral diversity, accounting for 1,678 scaffolds, 243 of which exceeded 10 kb. These scaffolds form near-complete genomes, as their length is typically around 20 kb^33,34. PLVs also constitute a significant proportion of reads, averaging 3.5% and 1.2% of cellular and viral fractions respectively (Fig. 1d,f). A maximum of 27% of the viral reads in the cellular fraction during January 2019 reflects an increase in the cellular host abundance rather than lytic viral infection, as these are prevalent endogenous viral elements of eukaryotes³³. Meanwhile, their high abundance in the viral fraction likely indicates a shift to an active infectious state.

Though our library preparation methods are biased against ssDNA viruses³⁵, we identified five scaffolds assigned to the Microviridae family (non-enveloped bacteriophages), two belonging to the filamentous Inoviridae family (both infecting bacteria), and 16 to the eukaryotic specific Circoviridae family³⁶. We do not expect ssDNA viruses to be a significant portion of the viral communities, as previous surveys using methods biased towards ssDNA viruses still found dsDNA viruses to dominate²⁸.

Antarctic virus diversity responds to seasonal changes

Viral communities changed over the sampling period and could be delineated into four distinct clusters (Fig. 2a-b), each representing specific periods in the productive season (November to March) with distinct environmental conditions. The early summer (cluster D) is characterized by low water temperature, high salinity, high dissolved inorganic nutrient concentrations, low chlorophyll a (Chl-a) concentration, and low bacterial and viral abundances. During mid-summer (cluster A), increased temperature led to sea-ice melting and subsequently lower salinity and stronger stratification (shallower mixed layer depth). Accordingly, higher photosynthetic active radiation (PAR) coincided with peak Chl-a concentrations and nutrient depletion, while bacterial and viral abundances increased. Late-summer (Cluster C) maintained high Chl-a concentrations and microbial abundances but at increased temperature, diminished sea-ice cover and relatively low nutrients. By the onset of autumn (cluster B), higher wind activity and lowered temperature led to increased mixing and sea-ice formation, (with replenished nutrient concentrations and increased salinity) and declining microbial productivity (low Chl-a and microbial abundances)³⁷. February 2018 and 2019 came in two separate clusters (C and A respectively), indicating some interannual variability.

We evaluated environmental determinants of virus community composition using a multivariate regression model. Temperature and salinity explained 31% of the community variance (p-value = 0.001 and 0.004, respectively), while another 23% was explained by inorganic phosphate (p-value = 0.001), nitrogen-to-phosphorus ratio (p-value 0.016), and PAR (p-value 0.077). The correlation between viral and microbial community compositions (Supplementary Fig. 2) was examined by comparing the pairwise Aitchison distances. Microbial community composition alone explained 72% of the variation in the viral community (mantel r-statistic = 0.7227, p-value = 0.001).

The abundances and temporal dynamics of the dominant viruses mirrored the total virus count by flow cytometry (Fig. 2c; R² = 0.82 and p-value = 9.5 x 10^− 7), and followed the bacterial (bloom) dynamics (Fig. 3a). Viral communities in late summer were directed by a higher contribution (and number) of the dominant viruses (96% Caudoviricetes; Supplementary Table 1) than in November till mid-January (Fig. 2c). Consequently, during periods of high viral abundance, the viral communities were heavily skewed towards this small subset of viruses, leading to a decrease in alpha diversity (Fig. 2d and Supplementary Fig. 3). Overall, these results highlight distinct viral communities during different periods in the season (Fig. 2c).

Abundant Antarctic bacteriophages infect co-abundant bacterial taxa

We predicted the hosts of 2,412 viral scaffolds using the integrated bacteriophage host prediction tool iPHoP³⁸. These scaffolds covered 48% of the reads mapping to mid to high-quality bacteriophage genomes (Fig. 3d) and were consistent with the vContact2 gene-sharing network analysis (Supplementary Fig. 4). Most bacteriophages targeted one of the three predominant bacterial classes (Fig. 3): Gammaproteobacteria (n = 769), Flavobacteria (n = 545), and Alphaproteobacteria (n = 348), which are important components of coastal Antarctic bacterial communities^39–41. iPHoP predicted 59 viruses to infect Archaea which peaked in February, contrasting with the earlier seasonal abundance of Archaea and archaeal cluster viruses (Fig. 3a-c and 1f).

Gammaproteobacteria-infecting bacteriophages represented 25% of the dominant phages (Supplementary Table 1), suggesting a significant role in seasonal virus production. In December, there was a noticeable rise in Gammaproteobacteria, predominantly Oceanospirillales, ahead of their bacteriophages (Fig. 3b-d). In late-summer (March), Gammaproteobacteria bacteriophage abundances were relatively high and increasing, while host abundances were not particularly high (Fig. 3d). Alternatively, Flavobacteria-infecting bacteriophages increased steadily with their host during December-January (mostly Polaribacter). Alphaproteobacteria-infecting bacteriophages were most abundant at the end of the season (April) and were still relatively high at the start of the next season (November), coinciding with the highest abundances of their host (mostly SAR11). Thereafter from December to January, the host population declined more rapidly than that of the bacteriophages.

Temperate bacteriophage reactivation with bacterial bloom onset

To elucidate the seasonal dynamics of temperate bacteriophages, we monitored the abundance of viral scaffolds containing genes involved in lysogenization and prophage reactivation (Fig. 3e and Supplementary Fig. 5), specifically integrase (n = 2,502), excisionase (n = 40), and bacteriophage repressor proteins (n = 134). All displayed similar seasonal patterns in the cellular and viral fractions. With an average vertical coverage (read abundance normalized by scaffold length) of 25,157 bp bp^-1, the scaffolds containing integrases were most abundant (2.4% of total viral vertical coverage). These viruses were most abundant during mid and late-summer (Supplementary Fig. 5), coinciding with elevated bacterial counts (Fig. 3a).

Antarctic waters harbor novel and diverse Crassvirales bacteriophages

Our target analysis identified 62 novel crassphage sequences (order Crassvirales). The teminase large subunit (TerL) phylogeny (Fig. 4a) validates the phylogenetic placement of these sequences within the Crassvirales order and their further classification to the family rank. Most scaffolds clustered into novel clades within the known Crassvirales diversity, labeled as C1 through C4 (Fig. 4a-b) and prospective new crassphage families. Additionally, 20 scaffolds belong to the Steigviridae family (Clade St, Fig. 4), which encompasses the Cellulophaga phage phi14:2 that infects marine free-living Flavobacteria^42,43. We further confirmed the family placement with the phylogeny of the portal protein (Portal, n = 51) and major capsid protein (MCP, n = 37, Supplementary Fig. 6a-b). These crassphages presented a typical gene content and organization for members of the Crassvirales (Fig. 3c), with all the 30 core genes⁴⁴ identified in at least one of the Antarctic Crassvirales scaffolds (Supplementary Fig. 6c). Clade C2 displayed the highest diversity and prevalence, with several viruses persisting throughout the season and others only detectable during late and early summer.

Nucleocytoviricota viruses are a diverse and abundant group of Antarctic eukaryotic viruses

Considering the large genomes and particle sizes of many NCVs⁴⁵, we characterized their diversity using metagenomic binning in both the viral and cellular fractions. This resulted in 94 bins (Fig. 5a), that were phylogenetically placed in the orders Imitervirales (n = 51), Algavirales (n = 15), candidate Pandoravirales⁴⁶ (n = 4) and Pimascovirales (n = 4), and candidate family Mamonoviridae⁴⁷ (n = 15). Additionally, our analysis recovered five bins belonging to the candidate phylum Mirusviricota⁴⁸.

Most Imitervirales bins belonged to the Mesomimiviridae family (n = 25), which includes viruses infecting the haptophytes Phaeocystis globosa and Chrysochromulina ericina^49,50. The December peak of three prevalent mesomimiviruses (Fig. 5b-c) coincided with a decline in P. antarctica (Supplementary Fig. 2c), the most abundant haptophyte. Of these, bin 2-643264.cc.b128 was most abundant of all NCVs and closely related to P. globosa viruses. Twenty of the remaining Imitervirales-related bins belonged to the Schizomimiviridae family, with their peak abundance during late summer (Fig. 5b), particularly in the viral fraction (Supplementary Fig. 7). The remaining six Imitervirales viruses belonged to the Allomimiviridae family and occurred in late-summer (Feb-March; Fig. 5b-c).

The Algavirales viruses belonged to the genus Prasinovirus (n = 10) and Raphidovirus (n = 4), predominantly abundant in the cellular fraction during early and late season (Fig. 5b-c; Supplementary Fig. 8). Prasinovirus bins vRhyme.bin.4783 and vRhyme.bin.4827 were persistent throughout the productive season in the viral fraction (Supplementary Fig. 7).

Fifteen large bins (average 11.2 Mb) were recovered from the cellular fraction and classified as Mamonoviridae-like⁴⁷, with medusavirus⁵¹ as the closest relative (Fig. 5a). Their large sizes suggest these to be endogenous viral elements as confirmed by 14 having some degree of Eukaryotic contamination and 10 having scaffolds containing both NCVs phylogenetic markers and eukaryotic signal (Supplementary Table 2). Additionally, five bins in the new phylum Mirusviricota⁴⁸ were identified in the cellular fraction (Fig. 5b-c). Lastly, we detected eight NCVs of which four belonged to the order of the Pimascovirales (which includes the Pithovirus and Marseillevirus) and four to the proposed Pandoravirales⁴⁶.

Antarctic virophages and polinton-like viruses expand the known diversity of these groups

By targeting their distinct yet diverse⁵² MCP phylogeny, we identified 1,678 PLV and 57 virophage scaffolds, and delineated them at the family rank. Most PLVs and virophages possessed virus hallmarks other than MCP, e.g., packaging ATPase or one replication related gene such as polymerase B, primase-helicase or helicase (Fig. 6c-d).

For the virophages, we applied the recently updated taxonomic framework⁵² (all candidate families). Most sequences belong to the families Sputniviroviridae (n = 13) and Omnilimnoviroviridae (n = 12). The most abundant virophage (NODE_23824) belongs to the latter family and is closely related to the Antarctic organic lake virophage⁵³. Notably, a distinct clade of 30 virophages was identified, likely forming a novel virophage family (Fig. 6a). Within this clade, the virophage sequence NODE_3533 has a presumably complete 23.9 kb genome with core genes and direct terminal repeats (Fig. 6c). This sequence occurred mostly in January, together with NODE_23824, whereas NODE_6367 and _4538 were most prevalent in March. Only NODE_29197 dominated at the start and end of the season.

We identified putative NCVs associated with these virophages by matching the promoter motif of the virophages with the late transcription promoter motifs of the co-occurring NCVs (Fig. 6c), as these are often shared^54,55. Antarctic virophages from the Sputniviroviridae matched with two mesomimiviruses and a schizomimivirus. In the Omnilimnoviroviridae, one matched with mesomimivirus and 2 with prasinoviruses. The novel clade matched mostly with prasinoviruses and mesomimiviruses with one match to a schizomimivirus, a mirusvirus and a pandoravirus.

We classified PLVs based on previous surveys on this group^34,56, as most clades have no formal taxonomic classification. Remarkably, Antarctic PLVs populated most phylogenetic groups (Fig. 6b), except for clades A and B which mostly contain MCPs endogenous to Opisthokonta genomes and animal adintoviruses (Fig. 6b). Another three separate groups (X1, X2, and X3) could be defined, with reference viruses from the previous clade X⁴⁰ and cluster TriMCP⁵⁶. These groups seem to arise from single PLVs exhibiting a highly divergent MCP gene triplication (Fig. 6b and Supplementary Fig. 8a). Shorter sequences with less than three MCPs are likely incomplete (Supplementary Fig. 8). Many of these PLVs MCP (n = 1387) make-up novel clades within these three groups. Most of the 291 remaining Antarctic PLVs belong to the Chi, TSV, and PgVV groups, with several forming distinct Antarctic clades (Fig. 6b). Of the 30 most abundant PLVs (> 10kb), about half were prevalent during February and the others during the early and late season (Fig. 6d). We investigated the possibility of these viruses parasitizing on NCVs by matching their promoter motif with the NCV early transcription promoter motif ‘WWWWWTGWWWWW’ as it was found for Gezel-14T⁵⁷. This resulted in matches for 19% of the clade PGVV, 11% of the clades X1-3 and Chi and 6% of the clade TSV longer than 5kb.

This study provides an in-depth characterization of highly unique marine viral diversity in Antarctic surface waters, where 75% of viral species were not present in the GOV⁸ and SOV²⁹ datasets. The limited overlap with SOV indicates a high degree of variability within the Antarctic virome. Overall, these findings underscore the Southern Ocean as an underexplored region regarding viral diversity, emphasizing the need for further comprehensive studies in this region. Beyond identifying a high proportion of new viral species, our comprehensive approach uncovered novel diversity at higher taxonomic ranks. The discovery of four novel Crassvirales families, C2 and C4 with many representatives greatly expands our understanding of crassphages diversity and their environmental range, beyond the primate gut microbiome^43,58. These viruses might similarly⁵⁹ infect Antarctic Bacteroidetes, a significant component of Marguerite Bay bacterial communities (classes Flavobacteria, Sphingobacteriia, Saprospiria and Cytophagia). Similarly, we discovered a novel virophage family and successfully recovered a complete genome of this family. Virophages replicate with co-infecting NCVs and may offer their host population-level protection against NCVs^60,61. In our analysis, virophages were matched to co-occurring NCVs, some to prasinoviruses, suggesting virophages to parasitize NCVs beyond the Imitervirales order⁵².

Our findings reveal a remarkably high diversity of PLVs, a highly understudied group^34,56 that was recently found prevalent in Eukaryotic genomes³³. The notable prominence of PLVs in our dataset (1,678 sequences) raises the question of whether this is Antarctic-specific or more widespread in the ocean, as has been found for freshwater environments⁵⁶. We found 154 PgVV-clade PLVs³⁴, with the one known isolate displaying a virophage-like lifestyle (Gezel-14T⁵⁷). These PLVs had the highest proportion of matches to the NCV early transcription promoter motif, suggesting that, similarly to Gezel-14T, they may also depend on NCVs. While five sequences related with Gezel-14T have been previously reported in Antarctic virome samples²⁷, our study substantially broadens the known diversity of Antarctic PLVs beyond the PgVV group. Additionally, 360 PLVs belonged to the TsV group that includes PLV Tetraselmis viridis virus S1 that does not require an NCV co-infection⁶². Notably, most PLVs (1,387) were related to those in group X³⁴ which our analysis shows to form three MCP clades, many representing one virus with duplicate and triplicate MCPs. The high divergence between the MCPs suggests this to be an early evolutionary event. In summary, these findings paint a picture of PLVs as an active and diverse group of viruses infecting Antarctic eukaryotes.

Another highlight of this study is the improved insight into the intricate relationship between host and virus. The NCVs were the dominant eukaryotic viruses and most belonged to the families Allo-, Meso-, and Schizomimiviridae, as well as the order Algavirales, of which all isolated members infect photosynthetic protists. The prevalence and temporal dynamics of these NCVs confirms their active role in regulating Antarctic phytoplankton population dynamics^24,26, ranging from preventive to reductive viral control⁶³. Indeed, some of the prasinoviruses, for which all isolates are known to infect picophytoplankton in the order Mamiellales⁶⁴, and capable of maintaining stable coexistence with their hosts⁶⁵, were remarkably persistent suggesting a more continuous lysis pressure. PgV-like viruses from the Mesomimiviridae have previously been identified as an important member of Antarctic viral communities^27,66. However, this study’s temporal resolution can establish their role in the co-occurring P. antarctica host bloom termination (reductive control), aligning with recent findings of viral lysis a major mortality factor of this phytoplankter²⁴. The steep increase in PgV-like viruses at the end of a P. antarctica bloom in December corresponded with the increase of Gammaproteobacteria. This bacterial group has been reported to quickly respond to viral infection of Phaeocystis⁶⁷, reinforcing the importance of viral lysis in fueling the microbial loop.

The comprehensive seasonal coverage and enhanced sampling resolution of this study also offers a deeper insight into prokaryotic bacteriophage-host interactions, with group-specific responses. For instance, Gammaproteobacteria displayed a typical predator-prey relationship illustrated by the decoupling between the host and bacteriophages abundance during the low sea-ice period. This pattern, characterized by high viral activity and low host abundance, was also found by viral activity assays in the ice-free areas of the Weddell Sea²². Such observations gain significance in the context of global warming, which is leading to the expansion of ice-free zones and longer ice-free periods around the Antarctic Peninsula⁶⁸. Moreover, we found closely coupled dynamics of Flavobacteria with their bacteriophages, particularly during the phytoplankton bloom in January (cluster A). The increasing abundances of Flavobacteria bacteriophages coincided with the bacterial bloom, whereas Gammaproteobacteria bacteriophages were most dominant after the bloom (cluster C). Such successional pattern can be explained by the association of Antarctic Flavobacteria with Chl-a fluorescence, and their degradation of diverse complex organic compounds^69,70. They make labile compounds available to other bacteria such as the Gammaproteobacteria and Alphaproteobacteria⁶⁹. In our study, Alphaproteobacteria and their bacteriophages became more prominent during the colder periods, at the end of the season (cluster B) and pre-bloom phase (cluster D), consistent with this group affinity with low temperature⁷¹.

The early summer marks an important transition for viral communities, where the onset of microbial production leads to increased bacteriophage production and abundance²⁵. Our analysis shows this period is characterized by the emergence of a few dominant bacteriophages that account for a substantial portion of viral activity. A pronounced increase in bacteriophage activity during these months in the same study area led to a shift in carbon flow towards the microbial loop rather than higher trophic levels²⁵. Concurrently, there is an apparent activation of temperate bacteriophages during the bacterial bloom, which is indicative of a refugium-to-activation dynamics⁷², and aligns with earlier research focusing on Antarctic prophage induction^22,23,25.

This study reports for the first time the strong seasonal variations in diversity and dynamics for the Antarctic viral community. Our phylogeny-informed characterization of viral diversity underscores the distinctness of these communities. Finally, the time-resolved and co-sampling of host communities offers a unique perspective, shedding light on the seasonality and abundance of Antarctic marine viruses and providing a more complete understanding of their diversity in relation to microbial hosts.

Data availability

Metagenomic raw reads have been deposited in the European Nucleotide Archive under the Bioproject TBD. The sampling metadata, viral scaffold nucleotide sequences, NCV bins, taxonomic and host annotations, viral marker gene protein alignments and phylogenies are available through the NIOZ dataverse upon request.

Acknowledgements

We are grateful to British Antarctic Survey and MMB-NIOZ technical staff for the logistical and technical support, as well as Nisma Abdelmalik and Monika Krolikowski for field assistance and Gosse Metz for the Sterivex DNA extractions. We are grateful to Alejandro Abdala, Julia Engelmann and Nina Dombrowsky for providing the functional annotation pipeline. C.L. was supported by the Deutsche Forschungsgemeinschaft (DFG) under Germany’s Excellence Strategy [EXC2051-Proj.390713860] and a junior researcher fellowship from the Fonds voor Wetenschappelijk Onderzoek (FWO) [12D8623N].

M.V.F. was supported by the VIROINF, EU-financed Marie Skłodowska-Curie Action (MSCA) Innovative Training Network (H2020-MSCA-ITN-2020). This work was funded by the POMVIDDY project (grant ALWPP.2016.019 awarded to C.P.D.B.) which was supported by the Dutch Research Council (NWO), within the Netherlands Polar Programme.

Contributions

G.J.P, C.P.D.B, and B.E.D. designed the study. E.M.W. and G.J.P collected all samples. G.J.P. processed and analyzed the data. T.E.G.B. analyzed the environmental metadata. G.J.P, M.E.S, N.Y. and M.G.F. designed and performed the virophage and polinton-like virus analysis. C.L and M.V.F designed and performed the crassphage analysis. G.J.P and L.W. designed and performed the phage host prediction analysis. G.J.P took the lead in writing. C.P.D.B. conceived the study and helped shape the research, overall direction and writing of the manuscript. B.E.D. helped shape the analyses and provide critical feedback. H.B. and L.W. helped revise the manuscript. All authors provided feedback on the results and manuscript.

Study site and sample collection

Samples were collected from 15 m depth in Marguerite Bay at the site of the Rothera Time-Series⁷³ (RaTS, latitude 67.572°S; longitude 68.231°W). Sampling conditions in the Antarctic are highly variable and not always favorable for sampling, still we were able to collect samples on 24 days over two consecutive productive seasons (7 months period, from February until April 2018 and from November 2018 until February 2019). The water column properties were characterized using a SeaBird SBE 19 + CTD that measured conductivity, temperature, depth, salinity, chlorophyll autofluorescence, and photosynthetically available radiation. Mix layer depth (MLD) was defined as the depth with a water density difference of 0.05 kg m^− 3 relative to the surface¹⁹. Water samples were collected by performing sequential casts with Niskin bottles. These were emptied by gentle siphoning (using clean PC tube) into 20 L pre-rinsed polycarbonate carboys. All carboys and tubing used for sampling were decontaminated by rinsing with Contrad 70™ 2% v/v, followed by MilliQ ultrapure water (UP). Prior to sample collection, carboys were rinsed thrice with sample water. The carboys were protected from direct sunlight using opaque bags. Sampling was performed in the morning between 9:30 and 11:00 AM. Samples were further processed within 4 h from the start of sampling. The lab was kept at in situ temperature (with a minimum of 0.1°C), and if needed,samples were kept on ice.

For dissolved inorganic nutrient subsampling, seawater was filtered through a 0.2 µm pore-size Supor membrane Acrodisc® filter (Pall, Port Washington, NY 11050, USA) and collected in 3 times sample-rinsed high-density polyethylene pony vials (PerkinElmer, USA). The subsamples for dissolved inorganic nitrogen and phosphorus were stored at -20°C until analysis, and samples for dissolved inorganic silicate were stored in the dark at 4°C. Samples for chlorophyll-a (Chl-a) were collected by filtering 1 to 8.5 L of unfiltered (total Chl-a) and 20 µm reverse-sieved (< 20 µm fraction) onto GF/F filters (Whatman, UK), after which the filters were each wrapped in aluminum-foil, placed in a plastic bag and flash-frozen in liquid nitrogen and stored at − 80°C until further analysis. For microbial abundances, 1 mL of seawater was fixed with glutaraldehyde to 0.5% v/v final concentration (25%, EM-grade, Sigma Aldrich, USA) for 30 min at 1°C (in the dark), after which it was flash-frozen in liquid nitrogen and stored at − 80°C until analysis in the NIOZ laboratory (The Netherlands). Phytoplankton cells were counted on fresh samples (kept on ice) directly upon arrival in the Rothera lab.

Prior to handling the genomic samples, all bench surfaces were cleaned with Contrad 70™ 2% v/v (Decon Labs, East Sussex, UK) and rinsed with UP water. The cellular fraction metagenomic samples were collected by filtering 3 to 15 L (1h filtering time at in situ temperature) of whole water through a 0.22 µm Sterivex™ (MilliporeSigma, Massachusetts, US). The filters were sealed with parafilm and stored at -80°C until extraction. The viral fraction metagenomic samples were collected from 20 L seawater using the iron flocculation method according to Hurwitz et al. 2013 (ref⁷⁴). Briefly, particles bigger than 0.22 µm were removed by two subsequential filtrations using a A/E glass fiber filter (142 mm, Whatman, GE Healthcare, UK) followed by a 0.22 µm PES filter (142 mm, MilliporeSigma, Massachusetts, US). The viruses were flocculated with 1 mg L^-1 of FeCl3 (1h incubation) and collected onto a 1 µm PC filter (142 mm, Whatman, GE Healthcare, UK). Filters were stored at 4°C in the dark until further analysis.

Microbial abundances, chlorophyll-a, and nutrient concentrations

Viruses and bacteria were enumerated according to Brussaard et al. (2010) with modification to the TE-buffer (pH 8.2, according to ref⁷⁶). Upon dilution in TE-buffer and staining with SYBRGreen-I (Sigma Aldrich, USA), the samples were enumerated using a FACSCalibur flow cytometer (BD Biosciences, USA) equipped with an air-cooled Argon laser with an excitation wavelength of 488 nm with the trigger on green fluorescence. Total phytoplankton abundances were obtained using a benchtop FACS Celesta flow cytometer (BD Biosciences, USA) with the trigger on red chlorophyll autofluorescence⁷⁷.

Chlorophyll-a concentration (Chl-a) was determined using HPLC (according to ref¹⁸). Briefly, filters were freeze dried (48 h) and pigments were extracted using 90% acetone (48 h at 4°C in the dark)⁷⁸. Pigments were separated by high performance liquid chromatography (HPLC; Waters model 2690), equipped with a 3.5 µm particle size Zorbax Eclipse XDB-C8 column with a cooled autosampler (4°C) according to Van Heukelem and Thomas (2001). Pigment detection and quantification was based on comparison with DHI LAB standards at 346 nm (Waters 996 HPLC Photodiode Array Detector).

The concentrations of dissolved inorganic macro-nutrients were obtained using a TRAACS 800 autoanalyzer (Bran + Luebbe, Germany), following Grasshoff et al. (1983) for nitrate and nitrite, of Murphy and Riley (1962) for phosphate, and of Strickland and Parsons (1968) for silicate (refs^80–82). The calibration standards were diluted in low nutrient seawater (LNSW) with salinity levels similar to the samples (35‰). The detection limit was 0.007 µM for nitrate + nitrite, 0.001 µM for nitrite, 0.007 µM for phosphate, and 0.03 µM for silicate.

Viral and cellular fraction metagenomes

The DNA of the cellular fraction (Sterivex filters) was extracted using the Qiagen DNeasy Powersoil kit (Qiagen, Carlsbad, USA), according to the manufacturer’s instructions. The filters were opened and cut into approximately 3 x 3 mm pieces prior to 2x 10 sec bead beating at 3.55 m s^-1 with a 30 sec dwell on a Bead Ruptor Elite (Omni Internationals, USA). Viruses from the viral fraction (iron flocculation method) were resuspended at 4°C overnight in a EDTA (0.1 M) -MgCl₂ (0. 2M) - ascorbate (0.2 M) buffer with pH of 6.0 (1 mL per liter of sampled seawater). The resuspended sample was gently filtered through a 0.45 µm SFCA syringe filter (Sartorius, Göttingen, Germany) and treated with 1 U µL^-1 of DNase I (Roche, Basel, Switzerland) for 18h at 4°C. DNase I was deactivated by adding EDTA and EGTA (0.1 M final concentration). The viruses were concentrated to a volume of 0.8 to 1 mL using a 100 kDa Amicon (MilliporeSigma, Massachusetts, US). DNA was extracted using Wizard® PCR Preps DNA Purification kit (Promega, Wisconsin, United States) according to manufacturer’s instructions. The DNA was eluted twice in 50 µL of 80°C molecular grade low TE buffer (pH 8.0, 10 mM Tris and 0.1 mM EDTA).

The libraries were prepared using the TruSeq Nano DNA Kit (Illumina, San Diego, CA) and sequenced with the Illumina platform NovaSeq6000 S2 (2x 150 bp), according to the manufacturer’s instructions, on average 80M reads per sample (58 to 123M). The read quality was assessed before and after quality trimming using MultiQC v1.8⁸³. PhiX reads were removed using bbduk v38.07 (https://sourceforge.net/projects/bbmap/), low complexity (dust filter threshold of 20) and PCR duplicates were removed with prinseq-lite v0.20.4⁸⁴. Illumina adaptors were clipped with trimmomatic v0.36⁸⁵. To assess the degree of enrichment of our viromes, the abundance of microbial genetic markers was assessed using ViromeQC v1.0.1 in all metagenomes⁸⁶.

PhyloFlash v3.4⁸⁷ was used to characterize microbial communities by target assembly and classification of the 16/18S rRNA gene. Reads that mapped to plastid 16S rRNA genes were removed.

Virus genome assembly, taxonomy, and host prediction

All reads from the viral fraction metagenomes were co-assembled into scaffolds using metaSpades v3.14.1⁸⁸. Scaffolds of at least 1500 bp long were checked for viral signals using PPR-Meta v1.1⁸⁹, which was found to be the most sensitive virus prediction tool with relatively few false positives in a benchmark work that included some of the samples from this study⁹⁰. Sequences identified as viral but with no CheckV v0.7.0 viral hallmark genes and cellular hallmark genes³¹ and classified as non-viral by Cenote-Taker2⁹¹ were removed. The remaining co-assembled scaffolds are equivalent to viral species-level operational taxonomic units (vOTUS) as clustering at 95% identity over 80% of the sequences⁸ resulted in the clustering of only 0.039% of the sequences.

We quantified the overlap between Rothera viral sequences and the Global Ocean Virome⁸ and Southern Ocean Virome²⁹ datasets by performing an all versus all blastn which were then clustered at 95% average nucleotide identity over 80% of the sequence using the CheckV virus clustering.

For viral sequences longer than 5,000 bp or with > = 70% completeness (CheckV completeness), coding regions were predicted using prodigal⁹² (v2.6.3 -p meta). The viral sequences were clustered based on their coding regions using vContact2⁹³ using the RefSeq prokaryotic viruses (--db ProkaryoticViralRefSeq94-Merged) together with a costume collection of protist DNA viruses (Supplementary Table 3). The clusters were visualized using Cytoscape v3.8.2 with a perfuse force-directed layout⁹⁴. Cenote-taker2 taxonomy was overlayed. The Caudoviricetes (the single class in the Uroviricota Phylum), Nucleocytoviricota, Preplasmiviricota, and archaeal virus groups were manually defined (Supplementary Fig. 9) which always corresponded with the taxonomy of reference genomes within those groups and largely with the Cenote-taker2 taxonomy for the metagenomic viral sequences. For the unknown sequences within the above defined groups were attributed the group phylum rank taxonomy. Scaffolds annotated as Caudoviricetes by Cenote-taker2 and all sequences within the Caudoviricetes mega cluster were further classified into family rank taxonomy using PhaGCN2³². For Crassvirales, diverse and abundant, Maveriviricetes (virophages) and Polintoviricetes (polinton-like viruses) sequences, we performed a targeted identification and classification (see sections below).

The host genus was predicted for high and mid quality bacteriophage sequences (longer than 10kb or at least 70% complete) using iPHoP v0.9³⁸ with the standard cut-off for genus rank prediction and a 0.2 false discovery rate cut-off for class and phylum rank prediction (using All_combined_scores.csv). Conflicting host predictions were removed. The standard genus level output resulted in 744 predictions while the lower prediction threshold resulted in 1,668 predictions at the class level.

All viral sequences, excluding those classified as eukaryotic viruses, were inspected for the presence of the lysogeny genes integrase, excisionase, and bacteriophage repressor proteins genes. The coding regions were predicted using prodigal⁹² (v2.6.3 -p meta) which were matched by hmmsearch (e-value 1e-5, HMMER v3.1b2) to HMM profiles from the PFAM (34.0), KO (97.0), ncbi COGs (COGS2020) and by diamond search (e-value 1e-10, v0.9.22) to the nrNCBI (downloaded in March 2022). For integrase were considered matches to K03733, K14059, K04763, K21039, COG0582, COG1518, PF00589, PF13495, PF00665, PF13683, PF13102, PF18697, PF09003, PF02899, PF12835, PF02022, PF14659, PF18644, PF13976, PF13333, PF14882, PF12834, PF17921 and those containing the string “integrase” in the diamond top hit excluding those containing the string “CRISPR” and “Transposase”. For excisionase were considered matches to PF06806 and PF0782, and those containing the string “excisionase” in the diamond top hit. For the phage repressor protein were considered matched to PF00717, K22300, K07727, K18830 and COG2932.

Reads from the viral and cellular metagenomes were mapped to all scaffolds using bwa v0.7.17 mem⁹⁵ Reads mapped with less than 95% identity over 80% of the read length were excluded with CoverM v0.6.1⁹⁶, and the number of reads mapped to each scaffold was calculated with samtools coverage v1.11 ⁹⁷. When scaffolds had less than 70% horizontal coverage, their abundance was set to 0. The read relative abundances were calculated by dividing the read counts by the total number of viral reads in each sample. The depth corrected scaffold vertical coverage was calculated by dividing the relative read abundance by the total scaffold length multiplied by the average number of viral read bases.

Antarctic Crassvirales bacteriophages

To identify Crassvirales scaffolds, we used a targeted approach that does not depend on protein predictions as Crassvirales are known to include alternative genetic codes⁴⁴. First, we used transeq⁹⁸ to obtain the six-frames translations of the DNA scaffolds (options: -frame 6 -table 11 -clean). HMM profiles of the Crassvirales phylogenetic marker genes (see ref⁴⁴), including the major capsid protein (MCP), the terminase large subunit (TerL), and the portal protein (portal), were used to search for homologs via hmmscan (HMMER v3.1b2). Hmmscan results were then filtered (e-value < 0.05, coverage > 0.3). Scaffolds over 10 kb in length and containing matches to any of these three proteins were put forward for further inspection as candidate Crassvirales (n = 466).

To place the candidate Crassvirales scaffolds within the larger context of Caudoviricetes families, we used the Virus Metadata Resource from the International Committee on Taxonomy of Viruses (VMR_21-221122_MSL37). We selected at random ten isolates (or all if < 10 available) for each of the 43 Caudoviricetes families excluding: i) members of the Crassvirales order (Feb 2023), and ii) isolates from the Helgolandviridae (n = 1), Graaviviridae (n = 2), and Duneviridae (n = 6) from which no sequences could be extracted using our approach. This resulted in 245 species/isolates for which we obtained genomes, encoded proteins, and gene annotations from the NCBI Nucleotide database.

To reconstruct the phylogeny, we retrieved the TerL, MCP, and portal proteins encoded in the reference genomes by extracting the proteins containing in their headers "terminase", excluding "terminase small", "major head", "major capsid", and "portal", using grep and seqtk subseq (https://github.com/lh3/seqtk). We were able to extract the phylogenetic markers from most of the genomes (243 TerL (99.2%), 234 MCP (95.5%) and 185 portal (75.5%) proteins from the 245 genomes). In addition, we collected 673 Crassvirales genomes and TerL, MCP and portal protein sequences (from refs^43,44). These sequences were used with those from our candidate Crassvirales to build separate phylogenetic trees of each protein. For the multiple sequence alignment, we used MAFFT⁹⁹ (v7.475, option: –auto), modeltest-ng¹⁰⁰ to determine the amino acids substitution model (option: -d aa), and IQ-TREE2 v2.1.1¹⁰¹ to compute the phylogeny (option: -m LG + G4 + F -B 1000). Candidate Crassvirales scaffolds for which all three proteins (TerL, MCP, and Portal) that fell within the clade formed by the ICTV-approved Crassvirales families were deemed novel Crassvirales sequences.

The ORFs of candidate scaffolds were detected with Prodigal v2.6.3⁹² using three modes to consider stop codon reassignments (options: standard, -TGA W, and -TAG Q). These were functionally annotated using the marker protein HMMs from Yutin et al. 2021 (ref⁴⁴). The trees and their annotation were created using the software suite ggtree v3.6.2¹⁰² and iTOL v5¹⁰³. The genomic maps were built via gggenomes v0.9.5.9¹⁰⁴.

Scripts used to obtain the results of the analysis of Crassvirales members can be found on github (https://github.com/MVFofanov/rothera_crassvirales).

Nucleocytoviricota viruses

To identify NCVs, we first binned all assembled scaffolds using vRhyme v1.1.0 for the viral fraction co-assembly and using CONCOCT v1.1.0 for the corresponding cellular fraction assemblies. Scaffolds were examined for Nucleocytoviricota markers by performing a hmmsearch on prodigal predicted proteins⁹² (v2.6.3 -p meta) against the nine giant virus orthologous groups HMM profiles using ncldv_markersearch⁴⁶. All bins containing hits to phylogenetic marker genes were further investigated. Bins were decontaminated from likely contaminating cellular scaffolds using viralrecall¹⁰⁵ (as per ref¹⁰⁶). The bins were then screened for the seven genetic markers⁴⁶. Only bins containing at least four out of seven markers were kept, except for those classified as Mirusviricota⁴⁸ and those closely related with medusavirus (Mamonoviridae⁴⁷) where all bins with at least two markers were kept. This is due to Mirusviricota not possessing all NCV marker genes⁴⁸ and because only three markers were detected in the medusa virus isolate despite reported previously⁵¹. Bins with more than one duplication of the genes SFII, VLTF3, A32, and PolB were removed to avoid bins with high strain heterogeneity (as per ref¹⁰⁶). We removed redundant bins by clustering bins at 95% average nucleotide identity or 100% marker gene amino acid identity and selected the best representative (most marker genes or highest N50). Eukaryotic contamination was identified by predicting (using Pfam reference) and classifying (using MMETSP_zenodo_3247846_uniclust90_2018) ORFs using MetaEuk v6.a5d39d9¹⁰⁷.

A reference set containing NCV isolates and metagenomic bins were created based on previous metagenomic surveys^{46,48,106,108} and additional RefSeq NCV isolates as of September 2022 (Supplementary Table 4).

The marker genes were detected in the reference and bins protein sequences and alignments constructed and concatenated using ncldv_markersearch. The alignment was trimmed of regions where > 20% of the sequences have a gap and with an entropy score bellow 0.55 using BMGE v1.12¹⁰⁹. The trimmed alignment was used to build a maximum likelihood tree using the best model finder option (-m MFP) with ultrafast bootstrap of 1,000 replicates. The best fit model was LG + F + R10. The tree was visualized using FigTree v1.4.4¹¹⁰.

Maveriviricetes and Polintonviricetes viruses

All scaffolds were inspected for virophages (class Maveriviricetes) and polinton-like viruses (PLVs, class Polintonviricetes) major capsid protein (MCP) using a collection HMM built using global, and clade restricted protein alignments. Rothera scaffolds ORFs were predicted and translated to amino acid using prodigal⁹² (v2.6.3 -p meta).

For the Maveriviricetes, we used a collection of non-redundant isolate and metagenomic MCP sequences⁵². The global and family rank MCP alignments were built with MAFFT⁹⁹ (v7.407, -auto), and the HMM profile built with hmmbuild¹¹¹ (HMMER v3.1b2). We used these profiles to identify putative virophage MCPs in the Rothera scaffolds’ amino acid sequences (> 200 aa) using hmmsearch (HMMER v3.1b2). The new putative virophage MCP were joined with the reference sequences and the Maveriviricetes phylogeny was built, and the sequences classified according to the newly proposed taxonomy (ref⁵²).

For PLVs, HMM profiles were constructed using the MCP genes from a collection of curated PLV sequences^34,56. PLV MCP identified in Yutin et al., 2015 (ref³⁴) were used as queries to search the non-redundant NCBI protein database. The hits were collected, clustered with UCLUST¹¹² with a similarity threshold of 0.5 and were iteratively aligned and clustered using the approach described by Wolf et al., 2018 (ref¹¹³). From the alignment of the cluster containing the query proteins, a tree was built using FastTree¹¹⁴. From the tree topology eight clades were manually inferred for which separate HMM profiles were built. Translated Rothera scaffolds were searched with these HMM profiles and the profiles provided by Bellas and Sommaruga, 2021 (ref⁵⁶). The hits were mixed with reference PLV MCPs (proteins of the HMM profiles); the resulting protein set was iteratively aligned and clustered using the abovementioned approach; the alignment of the protein cluster containing reference PLV MCPs was used to build a tree using FastTree. Tree clades were named according to the clades as defined in Yutin et al., 2015 and Bellas and Sommaruga, 2021 (refs^34,56).

For the functional annotation of the other virophage genes, all predicted proteins from putative virophage scaffolds were pooled with proteins predicted from virophage genomes available from Roux et al. 2023 (ref⁵²). Proteins were clustered using mmseqs easy-cluster¹¹⁵, and for each cluster, alignments and profiles were generated by MAFFT⁹⁹ (v7.505, - einsi), hhconsensus¹¹⁶ (v3.3.0) and finally hhmake¹¹⁶ (v3.3.0). All profiles were then searched against the Pfam (http://pfam.xfam.org/) and PDB (https://www.rcsb.org/) databases using hhsearch¹¹⁶ (v3.3.0, -E 1e-4). Matches with at least 90% probability, e-values below 1e-4, clear functional annotation as well as a score that was at least 90% of the best score per cluster were used for annotation. Annotation of PLV scaffolds was performed similarly to the virophages, with all predicted proteins from putative PLV scaffolds being pooled with proteins predicted from available PLV genomes^34,56. We additionally searched the profiles against a database consisting of the profiles of clusters from Bellas and Sommaruga 2021 (ref⁵⁶) and transferred annotations. The annotated genomes were visualized using gggenomes ¹⁰⁴ in combination with ggtree¹⁰².

For all NCV bins, predicted proteins were searched against the GVDB database using HMMER v3.3.2. Upstream regions (50bp) of genes homologous to Cafeteria roenbergensis virus ‘late’ genes were extracted from the genomes and searched for enriched motifs using the streme tool from the meme-suite v5.5.0 and the protein-coding regions as negative controls. Candidate motifs were filtered for non-AT consensus sites and scores below 1e-3. Similarly, upstream regions of all virophage genes were extracted and searched for enriched occurrences of these candidate motifs using the sea tool (--thresh 1) from the meme-suite, again using the protein-coding regions as negative controls. PLV scaffolds above 5,000 bps were analyzed similarly for occurrences of a consensus ‘early’ NCV promoter motif (WWWWWTGWWWWW).

Viral diversity and correlation with environmental parameters

We performed a multivariate analysis on viral sequences of high and mid quality (longer than 10 kb or at least 70% complete) to avoid highly fragmented genomes. Viral read abundance was normalized by sequence length and centered-log ratio transformation using microbiome R package v1.20.0 (https://github.com/microbiome/microbiome). We explored the seasonal dynamics of alpha diversity by calculating the species richness and the Shannon and inverse Simpson diversity indexes for each day using hillR v0.5.1¹¹⁷, which are expressed in effective number of species. Simpson evenness was calculated by dividing the inverse Simpson index by the species richness. Additionally, we tracked the number and abundance of dominant viruses, defined as viral scaffolds representing over 0.5% of the viral reads at any given time (n = 203).

We performed a hierarchical cluster analysis based of the Aitchison distance using hclust¹¹⁸ (method centroid), and the normalized abundances visualized in heatmap using the stats package in R.

Microbial abundances and Chl-a concentrations were log-transformed and other environmental metadata was visually checked for normality and z-scored or log transformed to achieve a distribution closest to normal. Pairwise Pearson correlation was computed and from strongly covarying environmental metadata pairs, one was removed. To find which combination of variables better correlated with changes in the viral community, we fitted a general multivariate regression model to our data using Adonis2 (vegan v2.6-4)¹¹⁹. Variables were sequentially permuted so that the variables were input to the model by order of highest sum of squares, non-significant variables were removed from the model.

Suttle, C. A. Marine viruses - Major players in the global ecosystem. Nat. Rev. Microbiol. 5, 801–812 (2007).
Fuhrman, J. A. Marine viruses and their biogeochemical and ecological effects. Nature 399, 541–8 (1999).
López-García, P. et al. Metagenome-derived virus-microbe ratios across ecosystems. ISME J. 2023 1710 17, 1552–1563 (2023).
Bar-On, Y. M. & Milo, R. The Biomass Composition of the Oceans: A Blueprint of Our Blue Planet. Cell 179, 1451–1454 (2019).
Brum, J. R. & Sullivan, M. B. Rising to the challenge: accelerated pace of discovery transforms marine virology. Nat. Rev. Microbiol. 13, 147–159 (2015).
Sullivan, M. B., Weitz, J. S. & Wilhelm, S. Viral ecology comes of age. Environ. Microbiol. Rep. 9, 33–35 (2017).
Breitbart, M. et al. Genomic analysis of uncultured marine viral communities. Proc. Natl. Acad. Sci. 99, 14250–14255 (2002).
Gregory, A. C. et al. Marine DNA viral macro- and microdiversity from Pole to Pole. Cell 177, 1109–1123 (2019).
Roux, S. et al. Ecogenomics and potential biogeochemical impacts of globally abundant ocean viruses. Nature 537, 689–693 (2016).
Luo, E., Leu, A. O., Eppley, J. M., Karl, D. M. & DeLong, E. F. Diversity and origins of bacterial and archaeal viruses on sinking particles reaching the abyssal ocean. ISME J. 2022 166 16, 1627–1635 (2022).
Ignacio-Espinoza, J. C., Ahlgren, N. A. & Fuhrman, J. A. Long-term stability and Red Queen-like strain dynamics in marine viruses. Nat. Microbiol. 2019 52 5, 265–271 (2019).
Coutinho, F. H. et al. Marine viruses discovered via metagenomics shed light on viral strategies throughout the oceans. Nat. Commun. 2017 81 8, 1–12 (2017).
Cassman, N. et al. Oxygen minimum zones harbour novel viral communities with low diversity. Environ. Microbiol. 14, 3043–3065 (2012).
Heinrichs, M. E. et al. Breaking the Ice: A Review of Phages in Polar Ecosystems. in 31–71 (Humana, New York, NY, 2024).
Marshall, J. & Speer, K. Closure of the meridional overturning circulation through Southern Ocean upwelling. Nat. Geosci. 2012 53 5, 171–180 (2012).
Frölicher, T. L. et al. Dominance of the Southern Ocean in Anthropogenic Carbon and Heat Uptake in CMIP5 Models. J. Clim. 28, 862–886 (2015).
Huang, Y., Fassbender, A. J. & Bushinsky, S. M. Biogenic carbon pool production maintains the Southern Ocean carbon sink. Proc. Natl. Acad. Sci. U. S. A. 120, e2217909120 (2023).
Biggs, T. E. G. et al. Antarctic phytoplankton community composition and size structure: importance of ice type and temperature as regulatory factors. Polar Biol. 42, 1997–2015 (2019).
Venables, H. J., Clarke, A. & Meredith, M. P. Wintertime controls on summer stratification and productivity at the western Antarctic Peninsula. Limnol. Oceanogr. 58, 1035–1047 (2013).
Luria, C. M. et al. Seasonal shifts in bacterial community responses to phytoplankton-derived dissolved organic matter in the Western Antarctic Peninsula. Front. Microbiol. 8, 2117 (2017).
Carvalho, F., Kohut, J., Oliver, M. J., Sherrell, R. M. & Schofield, O. Mixing and phytoplankton dynamics in a submarine canyon in the West Antarctic Peninsula. J. Geophys. Res. Ocean. 121, 5069–5083 (2016).
Evans, C. & Brussaard, C. P. D. Regional variation in lytic and lysogenic viral infection in the southern ocean and its contribution to biogeochemical cycling. Appl. Environ. Microbiol. 78, 6741–6748 (2012).
Brum, J. R., Hurwitz, B. L., Schofield, O., Ducklow, H. W. & Sullivan, M. B. Seasonal time bombs: Dominant temperate viruses affect Southern Ocean microbial dynamics. ISME J. 10, 437–449 (2016).
Biggs, T. E. G., Huisman, J. & Brussaard, C. P. D. Viral lysis modifies seasonal phytoplankton dynamics and carbon flow in the Southern Ocean. ISME J. 15, 3615–3622 (2021).
Evans, C. et al. Shift from Carbon Flow through the Microbial Loop to the Viral Shunt in Coastal Antarctic Waters during Austral Summer. Microorg. 2021, Vol. 9, Page 460 9, 460 (2021).
Eich, C. et al. Ecological Importance of Viral Lysis as a Loss Factor of Phytoplankton in the Amundsen Sea. Microorganisms 10, 1967 (2022).
Yang, Q. et al. Metagenomic Characterization of the Viral Community of the South Scotia Ridge. Viruses 2019, Vol. 11, Page 95 11, 95 (2019).
Gong, Z. et al. Viral Diversity and Its Relationship With Environmental Factors at the Surface and Deep Sea of Prydz Bay, Antarctica. Front. Microbiol. 9, 2981 (2018).
Alarcón-Schumacher, T., Guajardo-Leiva, S., Martinez-Garcia, M. & Díez, B. Ecogenomics and Adaptation Strategies of Southern Ocean Viral Communities. mSystems 6, (2021).
Lopez-Simon, J. et al. Viruses under the Antarctic Ice Shelf are active and potentially involved in global nutrient cycles. Nat. Commun. 2023 141 14, 1–10 (2023).
Nayfach, S. et al. CheckV assesses the quality and completeness of metagenome-assembled viral genomes. Nat. Biotechnol. 2020 395 39, 578–585 (2020).
Jiang, J.-Z. et al. Virus classification for viral genomic fragments using PhaGCN2. Brief. Bioinform. 24, (2023).
Bellas, C. et al. Large-scale invasion of unicellular eukaryotic genomes by integrating DNA viruses. Proc. Natl. Acad. Sci. U. S. A. 120, e2300465120 (2023).
Yutin, N., Shevchenko, S., Kapitonov, V., Krupovic, M. & Koonin, E. V. A novel group of diverse Polinton-like viruses discovered by metagenome analysis. BMC Biol. 13, 1–14 (2015).
Roux, S. et al. Towards quantitative viromics for both double-stranded and single-stranded DNA viruses. PeerJ 2016, e2777 (2016).
Breitbart, M., Delwart, E., Rosario, K., Segalés, J. & Varsani, A. ICTV virus taxonomy profile: Circoviridae. J. Gen. Virol. 98, 1997–1998 (2017).
Venables, H. et al. Sustained year-round oceanographic measurements from Rothera Research Station, Antarctica, 1997–2017. Sci. Data 2023 101 10, 1–13 (2023).
Roux, S. et al. iPHoP: An integrated machine learning framework to maximize host prediction for metagenome-derived viruses of archaea and bacteria. PLOS Biol. 21, e3002083 (2023).
Straza, T. R. A., Ducklow, H. W., Murray, A. E. & Kirchmana, D. L. Abundance and single-cell activity of bacterial groups in Antarctic coastal waters. Limnol. Oceanogr. 55, 2526–2536 (2010).
Sow, S. L. S. et al. Biogeography of Southern Ocean prokaryotes: a comparison of the Indian and Pacific sectors. Environ. Microbiol. 24, 2449–2466 (2022).
Luria, C. M., Amaral-Zettler, L. A., Ducklow, H. W. & Rich, J. J. Seasonal succession of free-living bacterial communities in coastal waters of the western antarctic peninsula. Front. Microbiol. 7, 1731 (2016).
Turner, D. et al. Abolishment of morphology-based taxa and change to binomial species names: 2022 taxonomy update of the ICTV bacterial viruses subcommittee. Arch. Virol. 168, (2023).
Yutin, N. et al. Discovery of an expansive bacteriophage family that includes the most abundant viruses from the human gut. Nat. Microbiol. 3, 38–46 (2018).
Yutin, N. et al. Analysis of metagenome-assembled viral genomes from the human gut reveals diverse putative CrAss-like phages with unique genomic features. Nat. Commun. 2021 121 12, 1–11 (2021).
Abergel, C., Legendre, M. & Claverie, J. M. The rapidly expanding universe of giant viruses: Mimivirus, Pandoravirus, Pithovirus and Mollivirus. FEMS Microbiol. Rev. 39, 779–796 (2015).
Aylward, F. O., Moniruzzaman, M., Ha, A. D. & Koonin, E. V. A phylogenomic framework for charting the diversity and evolution of giant viruses. PLOS Biol. 19, e3001430 (2021).
Zhang, R., Takemura, M., Murata, K. & Ogata, H. “Mamonoviridae”, a proposed new family of the phylum Nucleocytoviricota. Arch. Virol. 168, 1–6 (2023).
Gaïa, M. et al. Mirusviruses link herpesviruses to giant viruses. Nat. 2023 6167958 616, 783–789 (2023).
Gallot-Lavallée, L. et al. The 474-Kilobase-Pair Complete Genome Sequence of CeV-01B, a Virus Infecting Haptolina (Chrysochromulina) ericina (Prymnesiophyceae). Genome Announc. 3, (2015).
Santini, S. et al. Genome of Phaeocystis globosa virus PgV-16T highlights the common ancestry of the largest known DNA viruses infecting eukaryotes. Proc. Natl. Acad. Sci. U. S. A. 110, 10800–10805 (2013).
Yoshikawa, G. et al. Medusavirus, a Novel Large DNA Virus Discovered from Hot Spring Water. J. Virol. 93, 2130–2148 (2019).
Roux, S. et al. Updated Virophage Taxonomy and Distinction from Polinton-like Viruses. Biomol. 2023, Vol. 13, Page 204 13, 204 (2023).
Yau, S. et al. Virophage control of antarctic algal host-virus dynamics. Proc. Natl. Acad. Sci. U. S. A. 108, 6163–6168 (2011).
Fischer, M. G. & Suttle, C. A. A virophage at the origin of large DNA transposons. Science (80-. ). 332, 231–234 (2011).
Roux, S. et al. Ecogenomics of virophages and their giant virus hosts assessed through time series metagenomics. Nat. Commun. 2017 81 8, 1–12 (2017).
Bellas, C. M. & Sommaruga, R. Polinton-like viruses are abundant in aquatic ecosystems. Microbiome 9, 1–14 (2021).
Roitman, S. et al. Isolation and infection cycle of a polinton-like virus virophage in an abundant marine alga. Nat. Microbiol. 8, 332–346 (2023).
Dutilh, B. E. et al. A highly abundant bacteriophage discovered in the unknown sequences of human faecal metagenomes. Nat. Commun. 2014 51 5, 1–11 (2014).
Edwards, R. A. et al. Global phylogeography and ancient evolution of the widespread human gut virus crAssphage. Nat. Microbiol. 2019 410 4, 1727–1736 (2019).
Fischer, M. G. & Hackl, T. Host genome integration and giant virus-induced reactivation of the virophage mavirus. Nature 540, 288–291 (2016).
Koonin, E. V. & Krupovic, M. Polintons, virophages and transpovirons: a tangled web linking viruses, transposons and immunity. Curr Opin Virol. 25, 7–15 (2017).
Pagarete, A., Grébert, T., Stepanova, O., Sandaa, R. A. & Bratbak, G. Tsv-N1: A Novel DNA Algal Virus that Infects Tetraselmis striata. Viruses 7, 3937 (2015).
Brussaard, C. P. D. Viral Control of Phytoplankton Populations—a Review1. J. Eukaryot. Microbiol. 51, 125–138 (2004).
Moreau, H. et al. Marine Prasinovirus Genomes Show Low Evolutionary Divergence and Acquisition of Protein Metabolism Genes by Horizontal Gene Transfer. J. Virol. 84, 12555–12563 (2010).
Yau, S. et al. Virus-host coexistence in phytoplankton through the genomic lens. Sci. Adv. 6, (2020).
Alarcón-Schumacher, T., Guajardo-Leiva, S., Antón, J. & Díez, B. Elucidating viral communities during a phytoplankton bloom on the west Antarctic Peninsula. Front. Microbiol. 10, 1014 (2019).
Sheik, A. R. et al. Responses of the coastal bacterial community to viral infection of the algae Phaeocystis globosa. ISME J. 2014 81 8, 212–225 (2013).
Massom, R. A. et al. Antarctic ice shelf disintegration triggered by sea ice loss and ocean swell. Nat. 2018 5587710 558, 383–389 (2018).
Williams, T. J. et al. The role of planktonic Flavobacteria in processing algal organic matter in coastal East Antarctica revealed using metagenomics and metaproteomics. Environ. Microbiol. 15, 1302–1317 (2013).
Piontek, J., Meeske, C., Hassenrück, C., Engel, A. & Jürgens, K. Organic matter availability drives the spatial variation in the community composition and activity of Antarctic marine bacterioplankton. Environ. Microbiol. 24, 4030–4048 (2022).
Tada, Y., Makabe, R., Kasamatsu-Takazawa, N., Taniguchi, A. & Hamasaki, K. Growth and distribution patterns of Roseobacter/Rhodobacter, SAR11, and Bacteroidetes lineages in the Southern Ocean. Polar Biol. 36, 691–704 (2013).
Silveira, C. B., Luque, A. & Rohwer, F. The landscape of lysogeny across microbial community density, diversity and energetics. Environ. Microbiol. 23, 4098–4111 (2021).
Clarke, A., Meredith, M. P., Wallace, M. I., Brandon, M. A. & Thomas, D. N. Seasonal and interannual variability in temperature, chlorophyll and macronutrients in northern Marguerite Bay, Antarctica. Deep Sea Res. Part II Top. Stud. Oceanogr. 55, 1988–2006 (2008).
Hurwitz, B. L., Deng, L., Poulos, B. T. & Sullivan, M. B. Evaluation of methods to concentrate and purify ocean virus communities through comparative, replicated metagenomics. Environ. Microbiol. 15, 1428–1440 (2013).
Brussaard, C. P. D., Payet, J. P., Winter, C. & Weinbauer, M. G. Quantification of aquatic viruses by flow cytometry. in Manual of Aquatic Viral Ecology 102–109 (American Society of Limnology and Oceanography, 2010).
Mojica, K. D. A., Evans, C. & Brussaard, C. P. D. Flow cytometric enumeration of marine viral populations at low abundances. Aquat. Microb. Ecol. 71, 203–209 (2014).
Marie, D., Partensky, F., Vaulot, D. & Brussaard, C. Enumeration of Phytoplankton, Bacteria, and Viruses in Marine Samples. Curr. Protoc. Cytom. 10, 11.11.1-11.11.15 (1999).
van Leeuwe, M. A., Villerius, L. A., Roggeveld, J., Visser, R. J. W. & Stefels, J. An optimized method for automated analysis of algal pigments by HPLC. Mar. Chem. 102, 267–275 (2006).
Van Heukelem, L. & Thomas, C. S. Computer-assisted high-performance liquid chromatography method development with applications to the isolation and analysis of phytoplankton pigments. J. Chromatogr. A 910, 31–49 (2001).
Strickland, J. D. H. & Parsons, T. R. A Practical Handbook of Seawater Analysis. Ottawa: Fisheries Research Board of Canada, Bulletin 167. Internationale Revue der gesamten Hydrobiologie und Hydrographie vol. 55 (1972).
Murphy, J. & Riley, J. P. A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta 27, 31–36 (1962).
Grasshoff, K., Kremling, K. & Ehrhardt, M. Methods of seawater analysis. (1983).
Ewels, P., Magnusson, M., Lundin, S. & Käller, M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics 32, 3047–3048 (2016).
Schmieder, R. & Edwards, R. Quality control and preprocessing of metagenomic datasets. Bioinformatics 27, 863–864 (2011).
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Zolfo, M. et al. Detecting contamination in viromes using ViromeQC. Nat. Biotechnol. 2019 3712 37, 1408–1412 (2019).
Gruber-Vodicka, H. R., Seah, B. K. B. & Pruesse, E. phyloFlash: Rapid Small-Subunit rRNA Profiling and Targeted Assembly from Metagenomes. mSystems 5, (2020).
Nurk, S., Meleshko, D., Korobeynikov, A. & Pevzner, P. A. metaSPAdes: a new versatile metagenomic assembler. Genome Res. 27, 824–834 (2017).
Fang, Z. et al. PPR-Meta: a tool for identifying phages and plasmids from metagenomic fragments using deep learning. Gigascience 8, 1–14 (2019).
Wu, L.-Y. et al. Benchmarking Bioinformatic Virus Identification Tools Using Real-World Metagenomic Data across Biomes. Preprint at https://doi.org/10.1101/2023.04.26.538077 (2023).
Tisza, M. J., Belford, A. K., Dominguez-Huerta, G., Bolduc, B. & Buck, C. B. Cenote-Taker 2 democratizes virus discovery and sequence annotation. Virus Evol. 7, (2021).
Hyatt, D. et al. Prodigal: Prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics 11, 1–11 (2010).
Bin Jang, H. et al. Taxonomic assignment of uncultivated prokaryotic virus genomes is enabled by gene-sharing networks. Nat. Biotechnol. 2019 376 37, 632–639 (2019).
Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003).
Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. Preprint at https://doi.org/10.48550/arXiv.1303.3997 (2013).
Robbins, S. J. et al. A genomic view of the reef-building coral Porites lutea and its microbial symbionts. Nat. Microbiol. 2019 412 4, 2090–2100 (2019).
Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Rice, P., Longden, L. & Bleasby, A. EMBOSS: The European Molecular Biology Open Software Suite. Trends Genet. 16, 276–277 (2000).
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Darriba, Di. et al. ModelTest-NG: A New and Scalable Tool for the Selection of DNA and Protein Evolutionary Models. Mol. Biol. Evol. 37, 291–294 (2020).
Minh, B. Q. et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol. Biol. Evol. 37, 1530–1534 (2020).
Yu, G., Smith, D. K., Zhu, H., Guan, Y. & Lam, T. T. Y. ggtree: an r package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods Ecol. Evol. 8, 28–36 (2017).
Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v5: an online tool for phylogenetic tree display and annotation. Nucleic Acids Res. 49, W293–W296 (2021).
Hackl, T., Duponchel, S., Barenhoff, K., Weinmann, A. & Fischer, M. G. Virophages and retrotransposons colonize the genomes of a heterotrophic flagellate. Elife 10, (2021).
Aylward, F. O. & Moniruzzaman, M. ViralRecall—A Flexible Command-Line Tool for the Detection of Giant Virus Signatures in ‘Omic Data. Viruses 2021, Vol. 13, Page 150 13, 150 (2021).
Moniruzzaman, M., Martinez-Gutierrez, C. A., Weinheimer, A. R. & Aylward, F. O. Dynamic genome evolution and complex virocell metabolism of globally-distributed giant viruses. Nat. Commun. 2020 111 11, 1–11 (2020).
Levy Karin, E., Mirdita, M. & Söding, J. MetaEuk-sensitive, high-throughput gene discovery, and annotation for large-scale eukaryotic metagenomics. Microbiome 8, 1–15 (2020).
Schulz, F. et al. Giant virus diversity and host interactions through global metagenomics. Nature 578, 432–436 (2020).
Criscuolo, A. & Gribaldo, S. BMGE (Block Mapping and Gathering with Entropy): A new software for selection of phylogenetic informative regions from multiple sequence alignments. BMC Evol. Biol. 10, 1–21 (2010).
Rambaut, A. FigTree v1.4.4. (2018).
Eddy, S. R. Accelerated profile HMM searches. PLoS Comput. Biol. 7, e1002195 (2011).
Edgar, R. C. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461 (2010).
Wolf, Y. I. et al. Origins and Evolution of the Global RNA Virome. MBio 9, (2018).
Price, M. N., Dehal, P. S. & Arkin, A. P. FastTree 2 – Approximately Maximum-Likelihood Trees for Large Alignments. PLoS One 5, e9490 (2010).
Steinegger, M. & Söding, J. MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nat. Biotechnol. 2017 3511 35, 1026–1028 (2017).
Steinegger, M. et al. HH-suite3 for fast remote homology detection and deep protein annotation. BMC Bioinformatics 20, 1–15 (2019).
Li, D. hillR: taxonomic, functional, and phylogenetic diversity and similarity through Hill Numbers. J. Open Source Softw. 3, 1041 (2018).
Murtagh, F. & Legendre, P. Ward’s Hierarchical Agglomerative Clustering Method: Which Algorithms Implement Ward’s Criterion? J. Classif. 2014 313 31, 274–295 (2014).
McArdle, B. H. & Anderson, M. J. Fitting multivariate models to community data: A comment on distance-based redundancy analysis. Ecology 82, 290–297 (2001).

There is NO Competing Interest.

SupplementaryTable1.csv
supplementary Table1
SupplementaryTable2.xlsx
supplementary Table2
SupplementaryTable3.csv
supplementary Table3
SupplementaryTable4.csv
supplementary Table4
SupplementaryFiguresAntarcticvirusPiedade.docx
supplementary figures

Download PDF

Journal Publication

published 24 Oct, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Seasonal dynamics and diversity of Antarctic marine viruses reveal a novel viral seascape

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Antarctic viral communities harbor a diverse range of taxa

Antarctic virus diversity responds to seasonal changes

Abundant Antarctic bacteriophages infect co-abundant bacterial taxa

Temperate bacteriophage reactivation with bacterial bloom onset

Antarctic waters harbor novel and diverse Crassvirales bacteriophages

Nucleocytoviricota viruses are a diverse and abundant group of Antarctic eukaryotic viruses

Antarctic virophages and polinton-like viruses expand the known diversity of these groups

Discussion

Declarations

Methods

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1