Research ArticleMarine Ecology

Virus-mediated archaeal hecatomb in the deep seafloor

See allHide authors and affiliations

Science Advances  12 Oct 2016:
Vol. 2, no. 10, e1600492
DOI: 10.1126/sciadv.1600492


Viruses are the most abundant biological entities in the world’s oceans, and they play a crucial role in global biogeochemical cycles. In deep-sea ecosystems, archaea and bacteria drive major nutrient cycles, and viruses are largely responsible for their mortality, thereby exerting important controls on microbial dynamics. However, the relative impact of viruses on archaea compared to bacteria is unknown, limiting our understanding of the factors controlling the functioning of marine systems at a global scale. We evaluate the selectivity of viral infections by using several independent approaches, including an innovative molecular method based on the quantification of archaeal versus bacterial genes released by viral lysis. We provide evidence that, in all oceanic surface sediments (from 1000- to 10,000-m water depth), the impact of viral infection is higher on archaea than on bacteria. We also found that, within deep-sea benthic archaea, the impact of viruses was mainly directed at members of specific clades of Marine Group I Thaumarchaeota. Although archaea represent, on average, ~12% of the total cell abundance in the top 50 cm of sediment, virus-induced lysis of archaea accounts for up to one-third of the total microbial biomass killed, resulting in the release of ~0.3 to 0.5 gigatons of carbon per year globally. Our results indicate that viral infection represents a key mechanism controlling the turnover of archaea in surface deep-sea sediments. We conclude that interactions between archaea and their viruses might play a profound, previously underestimated role in the functioning of deep-sea ecosystems and in global biogeochemical cycles.

  • Deep-sea ecosystems
  • virus-host interactions
  • bacteria
  • archaea
  • microbial loop
  • biodiversity
  • ecosystem functioning


Viruses are the most abundant biological entities in the global ocean and are believed to infect all living organisms (15). By killing their hosts, viruses can manipulate marine environments, terminating phytoplankton blooms (6) and controlling the dynamics and biodiversity of their hosts, thereby playing key roles in carbon and nutrient cycling (particularly N and P) as well as ecosystem functions (14). Deep-sea ecosystems cover >65% of the world’s surface and represent >90% of the global biosphere volume, and microbial communities in the surface sediment (to a depth of 50 to 100 cm) are fundamentally important for nutrient regeneration and therefore vital for sustaining oceanic production (7). Viral infections in surface deep-sea sediments are responsible for the abatement of up to ~80% of the overall heterotrophic carbon production by bacteria and archaea (below 1000-m depth), causing the release of ~0.37 to 0.63 gigatons of carbon (GtC) year−1 on a global scale, suggesting that viruses can influence global biogeochemical cycles in fundamental ways (5).

Although bacteria tend to outnumber archaea in the world’s oceans, archaea make an important contribution to microbial biomass in deep waters (with abundances equivalent to those of bacteria at depths >1000 m) (8) and in marine subsurface sediments (that is, >1-m depth below the sediment surface) (9, 10). With a few exceptions (that is, hydrothermal vents, cold seeps, and anoxic ecosystems), archaea in surface deep-sea sediments are mainly represented by taxa belonging to the Thaumarchaeota (1113), which play important roles in biomass production and nutrient cycling (1416). Not unexpectedly, viruses infecting archaea have been identified from a wide range of environments, including marine ecosystems (1719). Despite increasing recognition of the crucial importance of archaea in biogeochemical and ecological processes (15, 16, 2022), the extent to which viral infection influences archaea in the oceans is unknown (23). This gap in knowledge limits understanding of the overall microbial dynamics and, hence, the ability to completely comprehend ecological processes and biogeochemical cycles occurring in the oceans.

Here, to discriminate between viral killing of archaea and bacteria, we applied and compared independent methods based on (i) the quantification, by real-time quantitative polymerase chain reaction (qPCR), of the number of 16S ribosomal RNA (rRNA) genes released from bacteria and archaea after viral lysis of the host cells, and (ii) the quantification of the number of 16S rRNA genes of bacteria and archaea after the selective inhibition of archaeal or both archaeal and bacterial metabolism. All determinations were carried out by multiple experiments performed both in mesocosms and in the field on deep-sea sediment samples from the Atlantic Ocean, Arctic Ocean, and Pacific Ocean and the Mediterranean Sea. Further high-throughput sequencing analyses were carried out on the DNA released following viral lysis and on the benthic viruses to identify the most affected archaeal taxa and the archaeal viruses causing their mortality.


Deep-sea sediment samples spanning depths from ca. 1000 to 10,000 m and covering a wide range of bottom-water temperatures and environmental conditions (table S1) were collected. Overall, we conducted more than 35 independent in situ and laboratory experiments and analyzed more than 480 sediment samples. To the best of our knowledge, the interactions between viruses, bacteria, and archaea in benthic deep-sea ecosystems have not previously been investigated using the high level of replication and geographic coverage of the present study.

In the top 50 cm of the sediments, fluorescence in situ hybridization analyses targeting rRNA (table S2) revealed that the number of archaea was lower than that of bacteria at all sampling sites (Fig. 1, A and B), representing 5 to 32% of the total microbial (archaeal and bacterial) abundance (on average 12%) and 5 to 23% of the total biomass (on average 11%; fig. S1). Massive sequencing of 16S rRNA genes of these archaea and bacteria (ca. 2,700,000 sequences obtained after cleaning; table S3) revealed that archaeal sequences were almost entirely affiliated with Marine Group I (MG-I) Thaumarchaeota, whereas bacterial sequences were mainly affiliated with unclassified proteobacterial groups (fig. S2). Rarefaction curves for both bacterial and archaeal 16S rRNA genes indicated that the sequencing depth was adequate to capture the diversity present in each sample (fig. S3). Notably, archaeal rarefaction curves reached less clear asymptotes, potentially indicating diversification levels for archaea (and especially, MG-I Thaumarchaeota) in deep-sea samples even higher than what was reported here.

Fig. 1 Bacteria, archaea, and viruses in deep-sea sediments.

Reported are bacterial (A) and archaeal (B) abundance obtained by catalyzed reporter deposition fluorescence in situ hybridization (CARD-FISH), and viral abundance (C) and production (D)along the vertical profiles at depth intervals of 0 to 1, 10 to 15, 20 to 30, and 40 to 50 cm in deep-sea sediments collected in the Arctic Ocean, Atlantic Ocean, and Pacific Ocean and Mediterranean Sea, with average values and SDs.

Irrespective of water column depth, viral abundance in the top 50 cm of surface sediments was high (range, 6.9 × 1012 to 36.4 × 1012 viruses m−2; Fig. 1C), and a range of archaeal viruses were present (that is, identified through metagenomic analyses of viromes; Fig. 2), suggesting that viruses can infect archaea inhabiting benthic deep-sea ecosystems. Viral infection of marine microorganisms (including archaea and bacteria) can cause lysis of host cells (14, 19). Rates of lytic infection (referred to here as production of viruses after cell lysis) were high in all surface sediments investigated, ranging from 3.6 × 1012 to 12.6 × 1012 viruses m−2 day−1 (Fig. 1D). The depth-integrated virus abundances divided by the respective viral production rates resulted in fast viral turnover times, averaging 2 to 3 days, consistent with previous evidence indicating that the deep-sea virome is a highly dynamic and active component of deep-sea ecosystems (5). Lytic viral infections release new viral progeny (virions) along with DNA (referred to here as extracellular DNA) (24) from the lysed host cells into the environment. The molecular method that we used exploits this property and, to the best of our knowledge, the first time permits quantification of and discrimination between the impact of viral infection on either archaea or bacteria, with a sensitivity much higher than that possible based on cell counts by fluorescence in situ hybridization. To do so, we determined, by real-time qPCR, the number of archaeal and bacterial 16S rRNA genes in the DNA released after cell lysis (Fig. 3), using primers and probes selected for consistency with previous studies investigating archaeal and bacterial dynamics in deep-sea sediments conducted worldwide (25, 26). We found that the higher the production rate of viruses, the higher the release of 16S rRNA gene copies in the extracellular DNA fraction (fig. S4).

Fig. 2 Archaeal viruses in the viromes.

Relative contribution (expressed as percentage) of the sequences belonging to each archaeal virotype to the total number of sequences of archaeal viruses identified in each virome from the different sampling sites.

Fig. 3 Release of bacterial and archaeal genes due to viral lysis.

The panel shows the release of viruses from killed archaea and bacteria (left column) and consequent shift of the 16S rRNA genes from bacteria (central column) and archaea (right column) to the extracellular DNA pools during time-course experiments conducted on deep-sea sediments from the Atlantic Ocean (first row), Mediterranean Sea (second row), Arctic Ocean (third row), and Pacific Ocean (bottom row). Reported are means ± SDs at the different time intervals, as well as the equation and R2 value of each line of best fit.

To test the assumption that factors other than viruses were negligible in causing cell lysis and in releasing 16S rRNA genes, additional experiments based on the use of selective inhibitors of bacterial and archaeal metabolism were conducted. These experiments demonstrated that the inhibition of viral replication blocked the release of archaeal or bacterial 16S rRNA genes (Fig. 4 and table S4). Moreover, the use of an archaeal-specific protein synthesis inhibitor (22) resulted in the complete cessation of virus-mediated archaeal lysis and of the accompanying release of archaeal 16S rRNA genes, with no effects on the viral lysis of bacteria (Fig. 5 and table S5). Overall, the experiments of inhibition of bacterial and/or archaeal metabolism indicated that nonviral causes of mortality were negligible during the short-time incubations we performed. Nonetheless, to exclude potential biases due to natural cell death over time, potentially causing release of 16S rRNA genes not due to viral infections, we recommend avoiding incubation lasting more than 12 hours when using cell metabolism inhibitors.

Fig. 4 Effect of the inhibition of archaeal and bacterial metabolism.

Reported are, from left to right, viral abundance, bacterial 16S rRNA gene copies in the extracellular DNA, and archaeal 16S rRNA gene copies in the extracellular DNA, during time-course incubations of untreated (control, blue bars) and treated (light blue bars, following addition of GC7 plus antibiotics to inhibit archaeal and bacterial metabolism) deep-sea sediment samples. Reported are average values and SDs. Asterisks indicate the significant increase (**P < 0.01) in the abundance of viruses and the number of bacterial or archaeal 16S rRNA gene copies after 12 hours of incubation in untreated (control) samples. ns, no significant differences occurring over time in treated samples; rDNA, ribosomal DNA.

Fig. 5 Effect of the selective inhibition of archaeal metabolism.

Reported are, from left to right, viral production, bacterial 16S rRNA gene copies in the extracellular DNA, and archaeal 16S rRNA gene copies in the extracellular DNA, during time-course incubations of untreated (controls, blue bars) and treated (light blue bars, following addition of GC7 to selectively inhibit archaeal metabolism) deep-sea sediment samples. Reported are average values and SDs. Asterisks (**P < 0.01; ***P < 0.001) indicate a significant decrease after 12 hours of incubation for viral production rates in treated samples, a significant increase in the number of both bacterial and archaeal 16S rRNA gene copies in control samples, or a significant increase in the number of bacterial only 16S rRNA gene copies in treated samples. ns, no significant increase in archaeal 16S rRNA gene copies occurring over time in treated samples.

In addition to the findings obtained from the study of the bacterial and archaeal cell metabolism, independent evidence of the importance of archaeal viruses in infecting and killing archaea was provided by the increase over time in the relative abundance of DNA sequences of archaeal viruses, including sequences affiliated to the known putative viruses infecting MG-I Thaumarchaeota (for example, Thaumarchaeota phage AAA160-J20 and thaumarchaeal putative virus Oxic1_7) (fig. S5 and Supplementary Methods) (27, 28). All of these results support the contention that viruses are a major cause of mortality for benthic deep-sea archaea and bacteria (5). However, we also note that adequate precautions should be taken and controls performed when analyzing natural systems using our newly developed approach, particularly when bacterial and archaeal mortality might be caused by other factors.

The accuracy of quantitative estimates of 16S rRNA gene copies of natural microbial assemblages could be affected by selective amplification of PCR products (29). To minimize the effect of primer selectivity, we used primer sets with the widest coverage for bacterial and archaeal taxa (table S2) (30). Moreover, the diversity of 16S rRNA genes contained within the intracellular and extracellular DNA was similar (figs. S2 and S6, A and B) as well as the diversity of the bacterial and archaeal sequences contained within these two metagenomes (that is, the intracellular and extracellular DNA pools; fig. S6, C and D). This indicated that we were assessing the same components of the intracellular and extracellular DNA pools and could make reliable estimates of the diversity and the number of 16S rRNA genes released by viral lysis. The presence of multiple genomes per cell (polyploidy), documented in certain bacteria (31) and archaea (Euryarchaeota) (32), should also be taken into account to accurately convert the number of 16S rRNA gene copies into cell abundance. However, in our samples Euryarchaeota had a negligible representation, thereby indicating that polyploidy did not have a significant effect on the estimates of virus-induced archaeal mortality.

We estimated that viral infections were responsible for the abatement of 1.0 to 2.2% day−1 (on average 1.6% day−1) of the bacterial abundance and 2.3 to 4.3% day−1 (on average 3.2% day−1) of the archaeal abundance in deep-sea sediments (Fig. 6). Therefore, despite the number of archaea in the sediments being much lower than that of bacteria, the impact of viruses on archaea was on average significantly higher. These findings were consistent along the vertical profile of the sediments down to 50-cm depth, across all oceanic regions investigated, and were not affected by pressure conditions (fig. S7). These values, once combined, are consistent with estimates of the overall virus-induced mortality rates determined at similar depths worldwide (5). The overall mortality rates of archaea plus bacteria (on average 2.6 ± 0.6 × 106 killed cells per gram per day) were very similar to those obtained by dividing the viral production by the average burst size reported for deep-sea sediments (that is, 45 viruses released per lysed cell, resulting in 2.1 ± 0.4 × 106 killed cells per gram per day) (5). Moreover, the viral production divided by the rates of lysis we derived from the release of 16S rRNA gene copies resulted in values of burst size (mean, 48; range, 17 to 187) almost identical to those previously reported by Danovaro et al. (5), providing further evidence of the reliability of the results presented here.

Fig. 6 Virus-induced mortality of bacteria and archaea.

Reported are average values and SDs of the virus-induced mortality of bacteria (A) and archaea (B) along vertical profiles at depth intervals of 0 to 1, 10 to 15, 20 to 30, and 40 to 50 cm, and the integrated values for the top 50 cm (C), of deep-sea sediments collected from Arctic Ocean, Atlantic Ocean, and Pacific Ocean and Mediterranean Sea. Asterisks indicate significantly higher (*P < 0.05) virus-induced mortality rates of archaea as found for all the oceanographic regions investigated.

To be sustainable over time, the rates of virus-induced mortality of bacteria and archaea should be balanced by cellular turnover. The turnover times of archaea and bacteria determined on the basis of C production rates (fig. S8 and table S6) were similar (11 to 23 days for archaea and 12 to 27 days for bacteria). These values were unaffected by pressure conditions (fig. S9) and were consistent with results available from deepwater masses (range, 20 to 100 days) (33) and deep-sea sediments (range, <2 to 67 days) (22). Notably, the turnover times of archaea, estimated by dividing the virus-induced cell mortality rates by the abundance of archaeal cells (12 to 22 days), were roughly equivalent to those determined by the radiolabeling experiments. In contrast, the turnover times of bacteria estimated from virus-induced mortality rates were significantly longer (31 to 52 days) than those calculated on the basis of radiolabeled substrate incorporation. This means that viral infection is responsible for the abatement of almost the entire biomass production of benthic archaea, whereas factors other than viruses (for example, predation or grazing by benthic consumers) may be responsible for the removal of a larger fraction of the biomass produced by bacteria than by archaea.

Our results, based on massive sequencing analyses of the 16S rRNA genes released by lysed bacteria and archaea, provided new insight on the role of viral infections in the microbial biodiversity and functioning of the deep oceans (5, 34). We found that, in all investigated deep-sea ecosystems, the viral impact occurred primarily on specific taxa belonging to MG-I Thaumarchaeota (fig. S10), suggesting that the viral impact is highest on the most represented archaeal group inhabiting surface deep-sea sediments (1113).

Our findings also indicated that C production of benthic bacterial assemblages was almost entirely due to heterotrophic metabolism, whereas more than 60% of the C production of benthic archaeal assemblages was due to chemosynthetic processes (fig. S8). In oxygenated surface deep-sea sediments, chemosynthesis is largely dependent on the oxidation of ammonia (1 mol of CO2 fixed to 10 mol of NH4+ oxidized) (20), which is supplied by microbial heterotrophic metabolism (22, 35). In line with this, we found a significant relationship between chemoautotrophic and heterotrophic C production (Fig. 7A). In turn, our results confirm previous evidence (5) that the labile organic C released by virus-induced cell lysis can stimulate the heterotrophic metabolism of uninfected deep-sea archaea and bacteria in the deep sea (Fig. 7B). These findings provide new evidence for a possible link between viral infections and chemoautotrophic production in the microbial food web. Viral-mediated mortality of bacteria and archaea may stimulate heterotrophic metabolism. This, in turn, may enhance nitrogen regeneration processes, supporting ammonia-dependent chemoautotrophic production. This biological feedback could have important consequences for C and N cycling. Assuming that the entire organic matter pool released by viral lysis is used by heterotrophic metabolism, and considering that 1 mol of ammonia is produced for every 4 mol of CO2 released (based on archaea and bacteria C/N ratio of 4:1), our results suggest that this process can provide 30 to 60% of the ammonia required to sustain chemoautotrophic C production of archaea. The relevance of this process indicates that viral infections can be one of the main drivers of chemoautotrophic production in deep-sea sediments, but it can also be hypothesized that a large fraction of ammonia, as well as other compounds supporting archaeal metabolism, derives from other heterotrophic processes (for example, from benthic fauna).

Fig. 7 Relationships between different microbial processes occurring in surface deep-sea sediments.

Reported are the relationship between total heterotrophic C production and (A) chemoautotrophic C production in surface deep-sea sediments (equations of the line of best fit, y = 0.011 + 0.019x7.67; R2 = 0.918; n = 18; P < 0.01) and (B) the amount of C released by viral lysis of host cells (equations of the line of best fit, y = 0.0988e1.58x; R2 = 0.831; n = 18; P < 0.01).

Despite the numerical dominance of bacteria, the C released by virus-killed archaea contributes 15 to 30% of the total C released by viral lysis in surface deep-sea sediments. Viral lysis of archaea and bacteria is estimated to release 0.37 to 0.63 GtC year−1 globally (5). The viral lysis of archaea alone is equivalent to ~0.08 to 0.14 GtC year−1 (in the top 1 cm), reaching ~0.3 to 0.5 GtC year−1 (in the top 50 cm of the sediments) when extrapolated to a global scale. We show here for the first time the crucial role of viruses in controlling archaeal dynamics and therefore the functioning of deep-sea ecosystems, and suggest that virus-archaea interactions play a central role in global biogeochemical cycles.


Study area and sampling strategy

Deep-sea sediment samples were collected in 2009–2011, during five independent oceanographic cruises conducted in the Atlantic Ocean, Arctic Ocean, and Pacific Ocean and the Mediterranean Sea (table S1). At each sampling station, surface (the top 1 cm) and subsuperficial sediment samples, down to 50-cm depth, were collected from independent cores (n = 3) using either multiple corers or a remotely operated vehicle (ROV). The experimental setup, which was used for all of the sampling areas investigated, consisted of 12-hour incubations of mesocosms maintained in the dark and in situ temperature conditions. Additional parallel incubations at in situ pressure were also carried out.

Abundance of bacteria, archaea, and viruses and viral production

The abundance of bacteria and archaea was determined by epifluorescence microscopy using the CARD-FISH technique (36), whereas viral abundances were determined after staining with SYBR Green I (37). The viral production in the deep-sea sediment samples was determined by the dilution technique (that is, a dilution-based procedure) through the measurement of the increase in viral abundance over time (5, 38, 39). This technique is the most widely used for benthic environments (thus enabling a comparison with available data), and it has the advantage of minimizing the impact of protozoa and/or fauna grazing on bacteria and archaea during the incubations (5, 38, 39). Details are reported in the Supplementary Materials.

Viral impact on archaea and bacteria

The same experimental mesocosms used for the determination of viral production were used to selectively quantify the number of archaea and/or bacteria killed by viruses, as follows. Sediment subsamples were incubated in the dark at the in situ temperature and collected at the beginning and after 1 to 3, 3 to 6, and 12 hours of incubation. Extracellular DNA was recovered from sediments collected at each time interval of incubation (40). Real-time qPCR analyses were used to quantify the number of 16S rRNA gene copies contained in the recovered extracellular DNA fraction for each time interval of incubation. The quantification of archaeal and bacterial 16S rRNA gene copies contained within the extracellular DNA was performed using primers and probes selectively targeting archaea (41) or bacteria (table S2 and Supplementary Materials) (42). The number of copies of 16S rRNA genes was calculated on the basis of the template length in base pairs (bp) and an average weight of 650 Da bp−1. The rates of release of 16S rRNA gene copies from archaea and bacteria killed by viruses were determined in the extracellular DNA pools from the linear regression analysis of the number of 16S rRNA gene copies over time. These rates were then converted into bacterial and archaeal cells killed by viruses assuming a value of 1.7 copies of 16S rRNA genes per cell, both for bacteria and archaea. This conversion factor represented the average number of 16S rRNA gene copies per archaeal cell provided by the Ribosomal RNA operon copy number Database (rrnDB) ( (43). This value was lower than the 16S rRNA gene copy number of bacteria (on average 4.2 copies per cell considering all bacterial genomes in the rrnDB) and slightly higher than that of all currently known Thaumarchaeota, including those inhabiting deep-sea sediments (1 copy per cell) (16, 43, 44). Thus, our estimates of the mortality rates of archaea were conservative, because the relative impact of viruses on archaea could be even higher using one copy per cell as a conversion factor.

To test the hypothesis that the 16S rRNA gene release into the extracellular DNA pool was related to the possibility that metabolically active bacterial and archaeal cells were infected and lysed by viruses, additional time-course experiments were conducted by incubating deep-sea sediment samples with GC7 (N1-guanyl-1,7-diaminoheptane; final concentration of 1 mM) to selectively inhibit archaeal metabolism (22, 4547) or with GC7 plus different antibiotics [chloramphenicol (0.3mg ml−1), tetracycline (0.05 mg ml−1), rifampicin (0.05 mg ml−1), and streptomycin (1 mg ml−1)] to also inhibit bacterial metabolism (4749). These concentrations were sufficient to inhibit both archaeal and bacterial metabolism, as revealed by synoptic incorporation experiments with radiolabeled substrates (both 3H-leucine and 14C-bicarbonate).

Bacterial and archaeal metabolism using radiolabeled substrates

For heterotrophic or chemoautotrophic C production determinations, sediment subsamples were added with 0.2-μm prefiltered seawater (collected at the sediment water interface of each station), containing 3H-leucine (68 Ci mmol−1; final concentration of 0.5 to 1.0 μM) or 14C-bicarbonate (58 mCi mmol−1; final concentration of 12 μCi ml−1), respectively (5, 22). The relative contribution of bacteria and archaea to the overall heterotrophic and chemoautotrophic C production was assessed by parallel experiments using selective inhibitors of microbial metabolism (see the Supplementary Materials).

Effects of pressure on the estimates of archaeal and bacterial mortality

Analyses of viral and archaeal plus bacterial production in deep-sea sediment samples collected from the Mediterranean Sea were conducted at both in situ and atmospheric (that is, 0.1 MPa) pressure to test for differences in the estimates of virus-induced mortality of archaea and bacteria. Moreover, we independently checked for the presence of possible artifacts due to the manipulation of the sediment during the sampling by determining the ratios between the numbers of copies of 16S rRNA genes recovered in the extracellular DNA pool and the total archaeal plus bacterial abundance for all sediment samples collected at the different depths.

Identification of archaea lysed by viruses and viruses responsible for their mortality

To identify the main taxa of archaea lysed by viruses at each sampling site, 16S rRNA gene sequences contained within the extracellular DNA pools at the beginning and after 12 hours of incubation were analyzed by high-throughput sequencing. The bacterial and archaeal 16S rRNA genes contained within the extracellular DNA pools were amplified (30) and sequenced on a 454 FLX Titanium platform. Metagenomic analyses conducted on the free viruses recovered from deep-sea sediment samples at the beginning of the time-course experiments and after 12hours (that is, after the viral infection events) were used to identify archaeal viruses causing cell mortality. Viruses were dislodged from the sediment using a physical-chemical procedure (50), and viral DNA, which was extracted and purified according to standard protocols, was sequenced on a 454 FLX Titanium platform (see the Supplementary Materials).


Supplementary material for this article is available at

Supplementary Methods

fig. S1. Bacterial and archaeal biomass in deep-sea sediments.

fig. S2. Diversity of bacterial and archaeal 16S rRNA genes in deep-sea sediments.

fig. S3. Rarefaction curves.

fig. S4. Rates of lytic infections versus 16S rRNA genes released from lysed hosts.

fig. S5. Viruses responsible for archaeal mortality identified by metagenomics.

fig. S6. Similarity between bacterial and archaeal assemblage composition in intracellular and extracellular DNA pools.

fig. S7. Effect of pressure on viral production, DNA release, and cell burst.

fig. S8. Heterotrophic and chemoautotrophic C production.

fig. S9. Effect of pressure on C production.

fig. S10. Impact of viruses on MG-I Thaumarchaeota in surface deep-sea sediments.

table S1. Details on the deep-sea sampling areas investigated.

table S2. Output of the in silico analyses dealing with the specificity and the coverage of oligonucleotides targeting 16S rRNA used in the present study.

table S3. Number of 16S rDNA sequences obtained before and after the cleaning process.

table S4. Output of the statistical tests for the experiments of inhibition of bacterial and archaeal metabolism.

table S5. Output of the statistical tests for the experiments of selective inhibition of archaeal metabolism.

table S6. Details on 14C analyses.

References (5171)

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We thank J. Middelburg and one anonymous reviewer for useful suggestions. Funding: This study was supported by the Flagship Project RITMARE (Italian Research for the Sea) coordinated by the Italian National Research Council and funded by the Italian Ministry of Education, Universities, and Research within the National Research Program 2011–2013. Further support was provided by the national project EXPLODIVE (A metagenomic approach for exploring microbial diversity and gene flows in deep-sea ecosystems) (FIRB 2008, contract no. I31J10000060001) and by the European Commission Seventh Framework Program Project MIDAS (Managing Impacts of Deep-seA reSource exploitation; grant no. 603418). We would like to thank the captain and crew of the R/V Kairei and operation team of ROV Automatic Bottom Inspection and Sampling Mobile onboard Japan Agency for Marine-Earth Science and Technology cruise KR11-11. The research performed by R.C. was supported by the Australian Research Council. Author contributions: R.D., A.D., C.C., and R.N. conceived the work and experimental approach. A.D., C.C., E.R., M.K., D.P., and T.N. contributed to the laboratory/bioinformatic analyses and experimental work. R.D., E.R., A.D., C.C., R.C., M.K., and D.P. contributed to data elaboration. All authors discussed and wrote the manuscript and commented on the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors. The sequences of the viromes and of the ribosomal genes analyzed in this paper have been deposited in MetaVir ( and in MG-RAST (, respectively, within the project EXPLODIVE (virome IDs: Med_Vir, Arct_Vir, Atl_Vir, Pac_Vir; ribosomal genes IDs: Med_1-6, Arct_1-6, Atl_1-6, Pac_1-6).
View Abstract

Stay Connected to Science Advances

Navigate This Article