Digging into the Soil Microbiome

Research into soil microbial activity can improve understanding of soil carbon turnover and persistence.

Soils provide over 98% of the calories consumed by humans and hold more organic carbon than the atmosphere and terrestrial biosphere combined, but much remains unknown about the complexity of soil systems and functions of its organic components. Yet, soil systems comprise the most biologically complex systems on Earth. As intensive agriculture and land-use changes are causing soil carbon and biodiversity loss, uncovering the functional relationships between soil carbon and microbial ecophysiology—how microbes live, die, and interact with their environment—is increasingly important.

Lawrence Livermore, within the “Microbes Persist: Systems Biology of the Soil Microbiome” Scientific Focus Area (SFA), has been conducting research to understand microbial carbon cycling since 2017, in collaboration with partners at the University of California (UC) at Berkeley; UC Davis; University of Minnesota (UM); Northern Arizona University (NAU); Lawrence Berkeley National Laboratory (LBNL); and Pacific Northwest National Laboratory (PNNL). The project seeks to uncover the soil microbiome’s role in carbon persistence—the length of time carbon remains in the soil before being released back into the atmosphere. This task is challenging in part because of soils’ spatial, temporal, and organismal variety and the complexity of the soil microbiome itself. Microbes dead and alive—including bacteria, archaea, fungi, viruses, and protists—comprise up to 50 percent of soil organic carbon compounds. However, determining which organisms, habitats, and environmental conditions lead to carbon’s persistence in soil is extremely challenging and has long been treated as a “black box” by soil system modelers. Microbial activity depends on a multitude of factors. In addition, microbial species process and use carbon differently, and their cells have distinct biochemistries. Distinct soil organic compounds may also persist for different periods. All of these factors can influence the residence time (inputs minus outputs) and transformations of soil organic compounds.

Using unique isotope probing techniques, the Microbes Persist team works to quantitatively determine rates of microbial growth, death, and consumption under various conditions, especially different soil moisture regimes. These quantitative traits can be combined with chemical and multivariate ‘omics data (DNA, RNA, proteins, metabolites) from sequencing and mass spectrometry to inform more accurate predictive models of organic matter turnover between soil microbes and their environment.

Experimental Exploration

Carbon’s path into the soil from the atmosphere begins with plant roots. As plants collect carbon dioxide (CO₂) from the atmosphere, they convert it into carbon-rich sugars via photosynthesis, which are used to grow their biomass. When plant roots grow, they release low-molecular-weight chemical compounds (called exudates) into the narrow region of soil nearby; this region, called the rhizosphere, is particularly critical for soil microbial carbon cycling. Exudates foster a rapid bloom of organisms with high growth rates and also predators. As plants grow, die, and then decompose, the roots’ organic carbon contents become substrates that microbes decompose and use for their own growth. Over time, the cells of these microbes accumulate alongside the original plant matter, with only a fraction of soil microbes left alive and active. “The organic matter that forms and starts to persist in soil partly looks like the original plant material, but up to half of it resembles molecules derived from bits and pieces of dead microorganisms,” says Jennifer Pett-Ridge, principal investigator of the SFA and lead of Livermore’s Carbon Initiative.

Once available to microbes in the soil, plant carbon’s fate varies greatly. Livermore scientists uses a suite of isotopic tracing and imaging techniques, some unique to the Laboratory, as well as various ‘omics techniques (methods used to analyze large amounts of biological data) to examine the fates of plant and microbial molecules and to determine the ecophysiological genetic traits—such as nutrient uptake, growth, and death rates—most predictive for carbon transformation.

Technologies such as quantitative stable isotope probing (qSIP), nanoscale secondary ion mass spectrometry (NanoSIMS), and novel radioisotope approaches allow SFA researchers to visualize and quantify nutrient cycling in the soil, an otherwise difficult endeavor due to the microbiome’s small scale and complex environment. A given soil sample’s DNA might contain 15 to 30 percent “relic” DNA that is no longer part of a live organism. Using qSIP, the Livermore team has shown that DNA reflecting active populations could be as low as 8 percent or as high as 50 to 60 percent. “Our team is known for using isotopes as a way to be quantitative about elemental flows in soil,” says Pett-Ridge. “We have a series of unique methodological tools that allow us to image soil matter, identify which organisms created it, and make longer-term, quantitative measurements of the turnover time of carbon.”

The qSIP technique is a quantitative version of widely used stable isotope probing (SIP), in which a stable, nonradioactive isotope is introduced to a sample—for example, via water with a heavy isotope of oxygen, such as oxygen-18 (¹⁸O) or CO₂ with the heavy isotope carbon-13 (¹³C)—and incorporated into organisms as they grow through the process of DNA replication. Sequencing this new, isotopically “heavy” DNA offers a significantly more detailed glimpse into which microbes are active in the sample and at what rates, a capability that the original SIP approach lacked. Steven Blazewicz, a Livermore staff scientist and technical co-lead of the SFA, played a key role in developing qSIP in collaboration with NAU. In this process, the enriched microbial DNA (with the stable isotope incorporated) is separated and sequenced, and a value of isotope enrichment—how much of the isotope has been incorporated—is calculated. Using the open-source R coding platform, these values are then used to inform estimates of growth, death, or nutrient consumption rates on a taxon-specific level through a code written by Livermore staff scientist Jeff Kimbrel. “At the moment, qSIP is our best way to identify active microbes in soil and quantitatively evaluate their rates of activity, be it carbon consumption or microbe growth and death rates,” says Blazewicz. Livermore has also developed unique high-throughput stable isotope probing (HT-SIP) capabilities for qSIP, using robotics to enable the rapid, automated, accurate fractionation of DNA. With these capabilities, the Laboratory can conduct well-replicated studies using thousands of samples, something that would have been impossible without HT-SIP.

In addition to quantifying carbon uptake by microbes, imaging carbon’s path from plants to soil and beyond is crucial to determine how it cycles through (or persists in) soil. NanoSIMS, an ion imaging technology, leverages stable isotope tagging and enables the team to see the spatial composition of isotopes in a sample by scanning an ion beam across its surface and generating an image at a spatial resolution of 200 nanometers or better. According to Peter Weber, a Laboratory staff scientist and Livermore’s NanoSIMS facility lead, the technology makes visible the destination of tagged elements, whether inside a cell or elsewhere in the soil ecosystem, including imaging of viruses. “NanoSIMS enables us to study processes occurring in soil with minimal disturbance,” says Weber. “For the SFA, we’re interested in which soil microbes are consuming carbon and what’s happening to that carbon after they’ve consumed it. We’re tracing it further down the chain to visualize its fate.”

In tandem with these isotope tracing and imaging techniques, the team employs both gene-based approaches (using all the DNA present in an environment, known as the metagenome) and metagenome-assembled genomes (MAGs)—individual microbial genomes assembled from individual sequences found in the broader metagenome—to link genomic data to environmental carbon cycle processes. To generate MAGs that are specific to the most active microbes in a sample, the team combines their qSIP approach (where they label active microbes in a complex community with isotopes and separate out active DNA), with whole metagenome sequencing. Looking for ecophysiological traits in MAGs that reflect how a microbe lives and dies (for example, growth rate, mortality rate, or carbon use efficiency) can help the team upscale their results from individual microbial populations to communities to the entire ecosystem. Linnea Honeker, a postdoctoral researcher for the SFA, is examining metagenomics in tandem with other types of ‘omics data, including metatranscriptomics (gene expression), lipidomics, and metabolomics, to make sense of patterns amongst these datasets—these data types have historically been researched separately.

Exploring the Carbon Pool

In the context of soil organic matter, recalcitrance is the idea that the thermodynamic properties of specific molecules can lead to differences in how long carbon persists. For example, a thermodynamically stable molecule of a fungal cell wall chitin may take longer to break down, and therefore persist longer in the soil, than a root exudate compound. “The soil carbon pool contains many different classes of carbon molecules from proteins to lipids to polysaccharides to aromatics,” says Pett-Ridge. “We sometimes oversimplify soil carbon by talking about it as if it’s one homogeneous pool, but soil carbon is really many different things.”

Historically, molecular recalcitrance was the most prominent theory explaining what led organic carbon to accumulate in soil. While research has since revealed other factors crucial in carbon’s persistence and disproven recalcitrance as an absolute determinant, some compounds do persist differently based on their class—a set of compounds with a shared structural feature. Earth scientist Karis McFarlane, who works in Livermore’s Center for Accelerator Mass Spectrometry, led a Laboratory Directed Research and Development (LDRD) project in which the team applied compound class extraction techniques to soil at different depths to explore the age of distinct carbon compounds. Samples were taken from from UC’s Hopland Research and Extension Center field site, the SFA’s main site in Northern California. The team extracted total lipids using solvents, amino acids via hydrolysis, and water-extractable organic carbon with water, leaving behind any acid-insoluble organic matter in a remnant phase.

Using radiocarbon (carbon-14) analysis, the team found that amino acids had a relatively young signature, indicating frequent turnover and less persistence. Analysis of the lipid fraction revealed an age gradient across soil depths; soil lipids were older in deeper samples but still relatively young overall. On the other hand, the acid-insoluble, refractory (dissolved) organic matter had a much older carbon signature. “This resistant carbon didn’t look as though it was very actively cycling. If we want to prioritize putting new carbon into soil pools in which it will last for a long time, deeper is better, and refractory microbial pools are ideal,” says McFarlane.

To further explore the persistence of distinct carbon molecules, McFarlane and her team are now performing even more specific amino acid and lipid extractions. This work will help identify where the molecules originated (for example, plants or microbes), whether microbes have processed them, and whether that causes them to last longer than others that were not. SFA researchers are also applying techniques from McFarlane’s LDRD in experiments connecting compound class and biogeochemistry to genomics, providing more insight into the puzzle of what causes soil carbon persistence.

Going Viral

By infecting bacteria and fungi, soil viruses play a critical role in influencing soil organic matter turnover. Livermore works frequently with collaborators at UC Davis to explore viral ecology and virus–microbe interactions. Viruses contribute in complex ways to the soil microbiome. They are dynamic, numerous, and varied in the organisms they target, and their roles and impacts are variable. For example, some viruses lyse, or break open, heterotrophic bacteria that break down organic carbon molecules, halting their process of CO₂ release. However, in doing so, they also release carbon-, nitrogen-, and phosphorus-rich material from within the bacteria’s biomass, feeding other microbes and contributing to soil nutrient stocks. The team is exploring whether cell lysis helps carbon persist longer, either via consumption by other organisms or becoming associated with minerals. “Last year, we discovered that the viruses that lyse their hosts could be contributing 40 to 45 percent of the turnover of microbial cells in soil. Simply understanding which viruses are active and when has been a big achievement for our project,” says Pett-Ridge.

Soil viruses can also move genes between organisms and create new capabilities for their hosts. Still others alter host behavior without incorporating new genetic code, leading to upregulation or redirection of activities such as metabolism. In both cases, new host activities may lead to changes in expected carbon turnover. Whether viruses are harmful or beneficial, their ability to target bacteria (phages), fungi, or protists, may lead to different effects on soil carbon cycling.

The qSIP technique provides a unique look into specific viral activity. Livermore staff scientist Gary Trubl had previously used viromics—a subset of metagenomics—to examine viruses through the genetic makeup of the community, but this approach lacks specificity. “While metagenomics is a powerful approach, we realized we couldn’t understand activity, such as which microbe was doing what and when, or link identity to function,” says Trubl. “With qSIP, we can see which viruses are active and isolate those interactions with their hosts.” In 2021, the team collaborated with the U.S. Geological Survey and PNNL and used heavy water to examine activity in arctic peat soil viruses during the winter. The researchers found that a large portion of the viruses present were viable and active, even in anoxic and subfreezing temperatures, and many carried auxiliary metabolic genes with functions that benefit carbon utilization, potentially providing their hosts with higher fitness for acquiring nutrients and, therefore, increasing the amount of carbon uptake in the local soil system. However, viruses’ net effect on carbon is complex. Viruses both increased carbon release through higher host respiration and increased soil carbon persistence through microbial death. “Viruses are key players in carbon cycling and major drivers of microbial community restructuring. In this case, viruses aided their bacterial hosts in nutrient scavenging, supporting their persistence and survival until lysis ultimately released carbon back into the system,” says Trubl. Expanding their viral explorations to other locations, the SFA is providing more data for soil carbon turnover modeling, improving understanding and predictions with every new finding.

Viruses also contribute to soil phosphorus cycling, which is essential for plant growth and broader soil nutrient cycles. As lead of another LDRD project, Trubl investigated phosphorus and viruses in Hopland site soils using radioisotopes since phosphorus does not have any stable isotopes. Trubl collaborated with Richard Bibby and Forrest Tolman at Lawrence Livermore’s Environmental Radioanalytical Monitoring Laboratory to leverage Livermore’s expertise in radioisotope tracing. He also examined viral activity in environments with added phosphate using SIP. The data suggest that phosphorus amendment leads to a bloom of RNA phages. In the future, the SFA plans to continue exploring and building on these results, seeking a broader understanding of nutrient cycling in soil. “We think viruses may play an outsized role in phosphorus cycling,” says Trubl. “When viruses infect and lyse their hosts, they disproportionately use phosphorus when making new viruses, more so than any other types of microorganisms. Therefore, we think viruses may be disproportionately concentrating phosphorus from the soil system.”

A Product of Its Habitat

Distinct soil habitats may also play a role in carbon persistence, but microbial habitats occur at scales that are so small they are difficult to directly measure. “We typically think of habitats as different biomes: the arctic biome, a tropical forest, a temperate forest, a semiarid grassland,” says Noah Sokol, a staff scientist in the Nuclear and Chemical Sciences Division. “Those biomes are relevant to larger animals and tree species. Soil microbes, however, experience habitat differences at a much tinier length scale.”

Soil without actively growing or decaying roots, what the SFA calls “bulk soil,” has less organic matter and lower microbial activity due to its distance from sources of carbon and nutrients. In contrast, the rhizosphere has a high abundance of resources, a far greater density and activity of microbial communities, and a larger number of ecological interactions between soil organisms. Strikingly, the rhizosphere and bulk soil can be just a few millimeters apart. Other soil habitats such as the detritusphere (the soil around decaying roots) and the hyphosphere (the soil surrounding fungal hyphae) also have distinct microbial inhabitants and activity relative to bulk soil. Explicitly measuring processes in different soil habitats can provide more accurate representations of carbon cycling in soil and better inform soil biogeochemical models.

The SFA team has shown that the microbial habitat where soil carbon is transformed can influence the association of organic matter with minerals—a major process in the soil carbon cycle. Carbon that is no longer part of a living microbe or plant root can become associated with soil minerals such as clays, joining a critical pool of soil organic matter that often persists on long timescales. “The mineral-associated organic carbon pool is the biggest pool of organic carbon in the entire terrestrial biosphere,” says Sokol. “An estimated 1,500 petagrams of soil carbon are mineral associated. Mineral-associated organic matter also tends to cycle much more slowly, on a decadal to centennial to even millennial scale.”

Mineral association of carbon usually results from one of two pathways: in the first, microbial necromass (dead cells) is decomposed and becomes bound to minerals, and in the second, simple compounds exuded from roots bound directly to minerals after entering the soil. “When microbes die, they leave byproducts that stick to minerals and create aggregates of physically protected organic matter, which are hard for a new cell or enzyme to access,” says McFarlane.

Livermore team members and collaborators examined the effects of soil moisture and different microbial traits—identified and quantified with qSIP—on the creation of mineral-associated soil organic matter in different habitats using soil from the Hopland site and Avena barbata, an annual grass common in Mediterranean climates. To detect how new carbon became incorporated into soil, the team tracked isotopically enriched ¹³C that was added to soil either as decaying ¹³C root detritus or via the living roots of plants growing in an atmosphere of rhizotron carbon-13 dioxide. The plants photosynthesized using this heavier CO₂, incorporated ¹³C into their tissues, and eventually released it into the soil via exudation. Once in the soil, any newly added ¹³C that living microbes had consumed and utilized could be detected and quantified with qSIP. Other ¹³C-enriched molecules became associated with minerals, which the team separated from particulate organic matter using a density-based fractionation procedure.

Using an elemental analyzer coupled to an isotope ratio mass spectrometer (EA-IRMS), the team inspected the mineral-associated fraction and determined how much ¹³C had been incorporated. By comparing the ¹³C signal in the mineral-associated soil carbon fraction to the ¹³C values in the plant input and the native soil organic matter, the team determined the ratio of the ¹³C label that ended up associated with minerals, which allows them to measure rates of mineral association. In both the rhizosphere and detritusphere, increased soil moisture led to higher amounts of mineral-associated carbon. Microbial traits that were positive predictors of mineral-associated carbon formation in the rhizosphere included the microbial community’s overall growth rate, total soil microbial biomass, and the quantity of extracellular secretions they produced. In contrast, microbial growth rate was a negative predictor of mineral-associated carbon in the detritusphere, and microbial biomass and extracellular polymeric substances had no significant effect. These findings suggest there is no one-size-fits-all microbial trait that best predicts the formation of mineral-associated soil carbon in soils’ different habitats.

While the team found stark differences in the microbial factors that drive mineral association of carbon between soil habitats, they also observed a commonality: drought conditions consistently led to decreased mineral association. As mineral-associated organic carbon is known to be the largest and longest-lasting soil carbon pool, this discovery can be used to improve biogeochemical models of mineral carbon cycling.

A Sprouting Field

Soil science research at Livermore has blossomed from just two researchers in 2005 (including Pett-Ridge) into the Soil SFA and several related SFAs, LDRD projects, and other initiatives, which release dozens of publications annually. SFA collaborators at UM specialize in fungi, NAU offers scaling insights and a specialty in protists, and UC Berkeley provides expertise in environmental metagenomics and biogeochemistry. “Projects like the SFA allow us to engage in long-term, collaborative research across multiple institutions and disciplines. This partnership provides us with the stability needed to explore ideas both deeply and broadly,” says Blazewicz.

Collaborators at LBNL are essential to incorporating data into computational models and developing microbial trait-based models for data from the project. LBNL’s modeling approach incorporates new microbial trait data into preexisting carbon stability models to improve accuracy. Livermore’s Katerina Georgiou also contributes to the SFA’s modeling efforts and has analyzed how mineral associations drive soil carbon cycling at the global scale. The SFA envisions that quantitative microbial trait data from different environmental conditions, locations, and interactions, combined with ‘omics data, can inform better models of soil microbial guilds and global carbon cycling. The team is currently working to identify the most effective combination of traits to make strong predictions for carbon changes.

The SFA will continue exploring soil’s fundamental genomic and molecular mysteries and building a framework for scaling from single cells, roots, and minerals to ecosystem and multiyear timescales to better predict biogeochemical cycles. Livermore is constructing a new Organic Matter Research Laboratory that will support the SFA and other related projects. “We’re excited to put our team’s expertise and new tools to work across several Department of Energy (DOE) mission areas—from sustainable food production to agricultural security,” says Pett-Ridge. “DOE has invested heavily in the field of genome sequencing, and genomics-informed research can tell us how biological processes are controlling the flows of materials. This information is critical to growing a robust bioeconomy.”

—Lilly Ackerman

For further information contact Jennifer Pett-Ridge (925) 424-2882 (pettridge2 [at] llnl.gov (pettridge2[at]llnl[dot]gov)) or Steven Blazewicz (925) 423-1506 (blazewicz1 [at] llnl.gov (blazewicz1[at]llnl[dot]gov)).