Unearthing Soil Viruses That May Help Bacterial Hosts

* Updated Feb. 14, 2023

Beneath our feet, the soil is teeming with viruses. What function viruses play in the soil ecosystem, however, has been relatively unexplored. To better understand this question, a multi-institutional team of researchers looked at the genomic makeup of DNA viruses in soil. Their findings, published in Nature Communications, suggest that viruses may help bacterial hosts by supporting decomposition of key nutrients in the soil ecosystem. 

As DNA sequencing technologies have advanced, it has become relatively cheap and fast to extract and sequence genomic materials from samples. This has resulted in an explosion of information available to researchers. The Joint Genome Institute (JGI), a Department of Energy (DOE) Office of Science user facility, maintains an enormous database of microbial genomes that served as the starting point for researchers’ quest to better understand soil viruses.  

The team of scientists pinpointed genes that belong to viruses and identified genes that don’t directly help a virus reproduce, known as auxiliary metabolic genes (AMGs). What exactly these non-essential genes do remains a bit of a mystery. Many AMGs have been assigned assumed functions based on their similarity to genes in other organisms, but these functions have not been validated. 

As John Cort, a chemist at Pacific Northwest National Laboratory (PNNL) and coauthor of the study, explains, “it’s complicated, because there can be pseudogenes that look like a functioning gene but don’t work or genes that appear to have a certain function but turn out not to.” 

Garry Buchko, also a chemist at PNNL and coauthor of the paper, concurs. He is part of the Soil Microbiome Science Focus Area that led this research. Automated tools can sequence the DNA and provide annotations, but until tested “the function is really the mystery.” Researchers therefore designed a study that would test whether an annotated function was correct.  

Finding enzymes that may support nutrient cycling  

Insect exoskeletons and fungal cell walls get their structure from the same natural building block, a carbon polymer called chitin. It is estimated that 1 billion tons of chitin, and the closely related chitosan, are produced on Earth yearly. As these abundant materials are broken down, they provide an important source of carbon and nitrogen for organisms in a variety of ecosystems, including soil. Researchers decided to examine soil viral AMGs suspected of encoding enzymes that help metabolize chitosan, called chitosanases. 

They scanned multiple databases for genetic sequences likely to encode chitosanases and applied stringent screening and inspection criteria to select genomic content for further analyses. They determined that the viruses carrying chitosanase AMGs were bacteriophages. In the process of phylogenetic analysis, they also compared viral chitosanases to other microbial chitosanases, such as bacterial and fungal, and discovered that viral chitosanases probably originated from bacteria.  

JGI then cloned the selected genes and sent plasmids to the Environmental Molecular Sciences Laboratory (EMSL), a DOE Office of Science user facility, where the encoded proteins were expressed. Then the products (the proteins suspected of being chitosanases) were tested for enzymatic activity to confirm that they were functional. This experimental validation was an important step in verifying that viruses could have a broader impact on nutrient cycling in the soil ecosystem. 

Determining molecular structure in 3-D 

At EMSL, one protein also was crystallized and sent to the Stanford Synchrotron Radiation Lightsource (SSRL), a national user facility operated by Stanford University on behalf of the DOE Basic Energy Sciences program. Here, cutting-edge X-ray crystallography was able to determine the molecular structure of the protein in astonishing detail.  

The results were groundbreaking because the structure had never been determined for this class of enzyme. Cort elaborated, “we wanted to find enzyme activity, and the structure was icing on the cake.” 

Proteins are composed of amino acid chains that translate into three-dimensional structures, like when a flat piece of paper is crumpled into a ball. This crumpling is called folding, and scientists have described and classified thousands of unique protein folds. In this study, researchers were amazed to find that the enzyme under review contained a fold that had never been seen before. 

As Cort described it, “it’s like finding a new planet. Twenty years ago, we used to find new folds all the time, but now it’s really rare.”  

Structure is tied to function, so seeing the molecular structure with such clarity allowed researchers to better understand how and where the chemical reaction that breaks down chitin occurs on the protein. They also used an artificial-intelligence-based software called AlphaFold to predict the structure of this protein, as well as a number of other chitosanases. This allowed researchers to perform a comparative structure analysis of various chitosanases.  

Determining the structure of one protein via both AlphaFold and X-ray crystallography also allowed researchers to test how well the software was able to predict structure. The two methods produced remarkably similar results, and AlphaFold was even able to predict the novel protein fold. 

According to Cort, “this doesn’t mean AlphaFold is this accurate in all circumstances, but it does show how well the software can work.”  

He emphasized that the software has offered an order of magnitude change in ability to predict protein structures, but that experimental validation is incredibly important. 

“The activity of enzymes can’t be predicted, so it was critical to use both,” he said. 

Cort said the software predictions and experimental validation complement each other well. One of the most unique aspects of the study was the opportunity to combine tools and resources from multiple institutions for a more holistic picture. Collaboration across institutions with different resources enabled the team to compare results across methodologies and test hypotheses in unique ways. 

Buchko emphasized the importance of this collaboration, noting, “that was a key strength of our study.” 

Ecological implications 

By harnessing the power of these diverse tools, not only did researchers provide the first structure of this previously uncharacterized protein, but they also confirmed the annotated function for the protein. They established that soil viruses carried at least one gene for a functional protein that could help decompose chitosan if it were to be translated. In marine environments, some viruses have been shown to carry genes that help support photosynthesis in their hosts. Similarly, soil viruses may play a central role in supporting their bacterial hosts, but more research is needed. 

According to Ruonan Wu, a computational scientist at PNNL and coauthor of the paper, the next step in understanding the function of viruses in the soil is to see the enzymes working in natural systems. 

The study’s findings support the hypothesis that soil viruses help their hosts by decomposing important sources of nutrients, but a number of unknowns remain. For instance, it is not known which organisms specifically use the nutrients made available by chitosan degradation.  

Additionally, it is not understood when the genes are translated; it may be that they only become active under certain environmental conditions like drought. Thus, while the study provides an exciting first step in understanding the role of soil viruses in nutrient cycling, more mysteries about the world beneath our feet remain to be unearthed. 

 A portion of this work was performed on a project award under the FICUS program and used resources at JGI and EMSL.