The twilight zone, formally known as the mesopelagic zone, occupies the depth range between the base of the euphotic layer and roughly one thousand meters. It is a critical waypoint in the biological carbon pump, the suite of processes by which carbon fixed by photosynthesis in the surface ocean is transported to depth, where a portion of it can be sequestered away from the atmosphere for decades to centuries. Most of that vertical transport happens on aggregates of particulate organic matter, flakes and pellets and colloidal clusters that vary enormously in size, density and composition. Microbes colonize these particles and remineralize their carbon, and viruses, by infecting and lysing those microbes, can accelerate the breakdown and reshape which carbon makes it to the deep ocean. Yet the viral ecology of sinking particles has remained one of the least charted corners of marine microbiology, because particles are hard to sample intact and viruses are hard to assign to hosts.

To overcome these obstacles, the team collected seawater from three oceanic regions and separated the particulate matter into multiple size fractions, then sequenced the metagenomes of samples taken at different depths from the surface down through the twilight zone and into deeper waters. From this material they recovered a total of 5,016 viral operational taxonomic units, or vOTUs, a level of diversity that underscores how rich and poorly described the particle-associated virosphere really is. By analyzing how these viral populations were distributed across particle sizes and water layers, the researchers could ask a question that has rarely been addressed at this scale: do viruses simply float passively in the water, or do they travel with the particles that sink out of the sunlit zone?

The answer turned out to depend strongly on both depth and particle size. Small particles near the surface harbored proportionally more lytic viruses, the kind that replicate aggressively inside a host cell and then burst it open, releasing new virions and dissolved organic matter in the process. Large particles at greater depths, by contrast, were enriched in lysogenic viruses, which integrate their genomes into those of their hosts and ride along with the dividing cells, replicating only when triggered. This pattern fits ecological theory: in productive surface waters, where microbial hosts are abundant and fast-growing, the kill-the-winner dynamics of lytic infection pay off, while in the resource-poor twilight zone, a dormant, integrated lifestyle is a safer bet for a virus whose potential hosts are scarce and scattered.

Host assignment revealed the players behind these interactions. More than 1,000 of the recovered vOTUs could be linked to prokaryotic hosts, and the dominant hosts belonged to the phylum Pseudomonadota, a group that includes many of the most abundant heterotrophic bacteria in the ocean. Using co-occurrence patterns between viral and host genomes, together with metagenome-assembled genomes recovered from the same samples, the researchers built a network of putative virus-host relationships spanning the particle size spectrum. The virus-to-microbe ratios and correlation analyses further suggested that the abundance of hosts shapes not only how many viruses are present but also the balance between lytic and lysogenic strategies in the community.

One of the most striking findings concerns what these viruses can actually do. Metabolic annotation of the viral genomes identified 178 auxiliary metabolic genes, or AMGs, which are host-like genes carried by viruses that can augment the metabolism of the cells they infect. The study also detected 12 families of carbohydrate-active enzymes, known as CAZymes, including glycoside hydrolases, glycosyltransferases and polysaccharide lyases. These enzymes are capable of degrading or remodeling the complex polysaccharides that make up the structural scaffolding of marine particles. Notably, the AMGs and CAZymes were found primarily in lysogenic members of the class Caudoviricetes, the tailed viruses that dominate marine viral communities. In practical terms, this means that viruses integrated into particle-dwelling bacteria may be supplying their hosts with extra tools to break down particulate organic carbon, thereby influencing how fast sinking particles are degraded and how much carbon escapes remineralization on the way down.

The functional complementation implied by these genes has direct consequences for carbon cycling. If a lysogenic virus equips its host with a glycoside hydrolase that can cleave a polysaccharide the host could not otherwise digest, the host gains access to a new energy source, and the particle loses structural material more quickly. Multiply that interaction across thousands of viral types and the countless particles sinking through the water column, and viral genomes emerge as a significant, previously underappreciated lever on the efficiency of the biological carbon pump. The study’s authors frame their results as new insight into the roles of viruses in carbon cycling and vertical connectivity in the ocean, and the metabolic evidence gives that framing concrete biochemical substance.

Equally important is the study’s evidence for connectivity between water layers. Co-occurrence analysis revealed extensive overlap of vOTUs across depths, and marker viruses identified with statistical comparisons of community composition showed that viruses originating in the surface ocean are transported via particles into the twilight zone and beyond, even reaching bathypelagic depths. In other words, the sinking particle highway does not carry carbon alone; it carries an entire mobile virome. Some viral types were significantly enriched at the upper or lower boundaries of the twilight zone in the different study regions, suggesting that particle-associated viruses can serve as biological tracers of vertical export, linking the ecology of the euphotic zone to the microbial processing of carbon far below.

The three-region comparative design strengthens these conclusions. By sampling across distinct oceanographic settings, the researchers could distinguish region-specific patterns from general ones, and the consistency of the size-dependent shift from lytic to lysogenic dominance suggests a robust ecological rule rather than a local curiosity. The use of tools such as vConTACT2 for viral clustering, non-metric multidimensional scaling for community ordination, and STAMP for statistical profiling of marker viruses reflects the maturing analytical toolkit of viral metagenomics, which now allows oceanographers to treat viruses as quantifiable components of biogeochemical models rather than as an unstructured background of biological noise.

The broader implications reach into climate science. Models of the biological carbon pump currently struggle to predict how efficiently carbon is transferred from the surface to the deep ocean, and microbial degradation on sinking particles is one of the largest sources of uncertainty. If particle-associated viruses systematically accelerate the solubilization and remineralization of particulate organic carbon, and if their lytic versus lysogenic balance shifts predictably with depth and particle size, then viral activity belongs explicitly in carbon export models. This study provides the first large-scale, size-resolved baseline for doing exactly that, cataloguing thousands of viral populations, their hosts and their metabolic genes along the particle size spectrum. As sequencing of twilight-zone particles continues, the viral ferry system described here is likely to prove a fundamental feature of ocean biogeochemistry, one that quietly shapes how much carbon the sea sends to the deep and how long it stays there.

Subject of Research: Viral communities associated with particles of different sizes in the ocean twilight zone and their roles in carbon cycling and vertical connectivity

Article Title: Connectivity and potential functions of viral communities along the particle size spectrum in the twilight zone

Article References: Zhang, Y., Wang, P., Liu, H., & Jing, H. (2026). Connectivity and potential functions of viral communities along the particle size spectrum in the twilight zone. BMC Biology. https://doi.org/10.1186/s12915-026-02750-0

Image Credits: AI Generated

DOI: 10.1186/s12915-026-02750-0

Keywords: ocean twilight zone, marine viruses, viral metagenomics, particle-associated viruses, biological carbon pump, auxiliary metabolic genes, carbohydrate-active enzymes, lysogeny, Pseudomonadota, carbon export, vOTUs, vertical connectivity

The research, led by Jia Yao and Ming Nie of Fudan University in Shanghai, together with colleagues, tackled a question with direct implications for climate modeling. Soil microbial respiration releases carbon dioxide as microbes metabolize organic matter, and the rate of that release accelerates as temperatures rise. The magnitude of this acceleration is captured by the temperature sensitivity coefficient, or Q10, which describes how much respiration increases for every ten degrees Celsius of warming. Whether Q10 varies systematically across climates, and what biological mechanisms drive that variation, remains one of the more contested issues in terrestrial carbon cycle science, with direct consequences for how strongly soils are expected to feed back into atmospheric warming.

To probe the role of lysogenic viruses, the team collected soils from 13 forest sites distributed along a 3,600-kilometer latitudinal climate gradient across China, from cool northern forests to warm subtropical ones. In the laboratory, they incubated each soil at three temperatures, 15, 25, and 35 degrees Celsius, and measured microbial respiration. Crucially, they also employed a chemical trick that has become a standard tool in viral ecology: treatment with mitomycin C, a compound that damages bacterial DNA and thereby provokes temperate prophages harbored within bacterial genomes to enter the lytic cycle. The response of a soil community to mitomycin C induction serves as an operational proxy for the abundance and activity of lysogenic viral processes in that community.

The design allowed the researchers to construct three distinct estimates of thermal sensitivity. Soils treated with mitomycin C exhibited higher Q10 values than untreated controls, indicating that chemically forcing prophages out of their lysogenic state intensified the temperature response of respiration. More intriguing was a third quantity: the Q10 calculated from the contrast between induced and uninduced soils, which the authors interpret as an SLV-associated Q10, reflecting the thermal behavior of respiration linked to lysogenic viral processes. This proxy-derived sensitivity was the lowest of the three, suggesting that processes tied to lysogenic viruses are associated with a dampened apparent response of respiration to warming.

Across the 13 sites, the variation in this SLV-associated temperature sensitivity was most strongly associated with two characteristics of the bacterial communities: their abundance and their life-history strategies. Soil bacteria are commonly sorted along a spectrum from r-strategists, which grow rapidly and opportunistically when resources are plentiful, to K-strategists, which invest in persistence, resource efficiency, and stress tolerance under more competitive conditions. Previous work has shown that the dominance of one strategy or the other shapes how microbial respiration responds to temperature, and the new findings indicate that the viral induction response is entangled with this same axis of microbial life history. Communities where lysogenic processes left a distinct imprint on respiration were also communities whose compositional and functional makeup reflected particular positions along the r-to-K spectrum.

Climate itself did not disappear from the picture, but its influence was largely indirect. The researchers used structural equation modeling, a statistical framework that allows direct and indirect pathways among variables to be separated, to trace how climatic variables propagate to the viral-linked temperature sensitivity. The analysis revealed that climate was connected to the SLV-associated Q10 through mediating factors, chiefly soil pH and the physical properties of the soil, such as texture-related characteristics. This pattern is consistent with a growing body of viral ecology showing that soil pH is a powerful determinant of viral community structure at scales from local plots to the globe. In other words, climate appears to set the stage, through the chemical and physical environment it helps create, on which lysogenic viral processes and bacterial life-history traits together govern how temperature-sensitive respiration becomes.

The geographic pattern embedded in these results is especially noteworthy. The association between the mitomycin C response and reduced apparent thermal sensitivity was strongest in warmer forest soils, implying that in the very ecosystems where microbial respiration could potentially release the most additional carbon under continued warming, lysogenic viral processes may be exerting a moderating influence. The authors are careful with their language throughout, framing the findings in terms of association rather than proof of causation. Mitomycin C induction is a blunt instrument: the compound is a DNA-damaging agent, and the observed differences between treated and untreated soils could in principle reflect the combined consequences of prophage induction, bacterial mortality, altered community composition, and the release of cellular contents. The interpretation of the induced contrast as an SLV-associated signal is therefore an operational one, a proxy rather than a direct measurement of viral behavior.

Even so, the study lands at a moment when the viral dimension of soil carbon cycling is moving rapidly from the margins to the mainstream. A global atlas of soil viruses published in 2024 catalogued an enormous and previously uncharted viral diversity and flagged potential biogeochemical impacts, while other recent work has demonstrated that viral lysis can alleviate microbial nutrient limitation and accumulate chemically recalcitrant dissolved organic matter in soils. Theoretical treatments have argued that viral infections likely mediate microbial controls on ecosystem responses to warming, but empirical tests along real climate gradients have been scarce. By combining a standardized induction assay with controlled incubations across thousands of kilometers, the new study offers one of the more systematic empirical links to date between a specific viral strategy and a keystone carbon-cycle parameter.

The mechanistic possibilities behind the observed pattern remain open. Lysogeny can benefit bacterial hosts directly: cryptic prophages can confer tolerance to environmental stress, and integrating viruses may alter host metabolism in ways that change growth rates and resource use. If lysogeny is more prevalent or more consequential in warm, low-latitude soils, and if it favors or co-occurs with K-strategist lineages whose respiration is inherently less temperature responsive, then the dampened SLV-associated Q10 would follow as an emergent property of the host-virus ecosystem. Alternatively, the induction treatment itself may reveal the scale of the lysogenic reservoir, with soils harboring more temperate phages showing different post-induction respiratory trajectories. Distinguishing among these mechanisms will require experiments that go beyond induction proxies, perhaps tracking viral and host population dynamics through time under warming.

For climate modelers, the message is both cautionary and constructive. The findings suggest that bacterial responses associated with lysogenic viral processes deserve a place in future assessments of how forest soil carbon responds to warming, particularly when evaluating microbial controls on carbon turnover. If lysogeny-associated processes genuinely reduce the thermal sensitivity of respiration in warmer forests, then current projections that ignore viral regulation may overestimate carbon losses from those systems. But because the evidence is associative and proxy-based, the authors’ conclusions point less toward immediate model revision and more toward an agenda: combining induction experiments with viromics, microbial trait measurements, and gradient sampling to pin down when, where, and how the quiet passengers in bacterial genomes bend the temperature curve of one of Earth’s largest carbon fluxes. The forest floor, it turns out, is not merely a chemical reactor warmed by the atmosphere above; it is also an arena where viruses, hosts, and climate quietly negotiate the pace of carbon’s return to the sky.

Subject of Research: Association between lysogenic soil viruses and the temperature sensitivity of forest soil microbial respiration

Article Title: Lysogenic viral processes are associated with reduced temperature sensitivity of forest soil microbial respiration

Article References: Yao, J., Xu, J., Xu, X., Liu, M., Chen, C., Bao, Y., Li, J., & Nie, M. (2026). Lysogenic viral processes are associated with reduced temperature sensitivity of forest soil microbial respiration. Plant and Soil. https://doi.org/10.1007/s11104-026-09126-x

Image Credits: AI Generated

DOI: 10.1007/s11104-026-09126-x

Keywords: lysogeny, soil viruses, bacteriophages, forest soils, microbial respiration, temperature sensitivity, Q10, carbon cycle, climate gradient, mitomycin C, bacterial life-history strategies, climate warming