When a microbe dies, it does not immediately disappear. Decomposition, even the decomposition of a micron-sized cell, takes time. Some components of cells, including parts of its genome, may persist long after a microbe has died. We know this is the case for humans, as any fan of CSI or any other forensics show on television would be quick to tell you. Even human bodies that have been buried for extended periods of time can yield sufficient DNA to permit identification of the corpse. With this paper, we wanted to know how much of the DNA in soil comes from the corpses of microbial cells and how this DNA may influence our ability to accurately characterize the soil microbiome. Just as human corpses buried in a cemetery should not be counted in population censuses, we would ideally not include DNA from dead cells in any census of soil microbial populations.
Scientists have long known that naked DNA can be remarkably stable in soil. Numerous studies investigating the fate of DNA ‘spiked’ into soil have shown that DNA can persist in soil for weeks to decades, even in the presence of ubiquitous DNA-degrading enzymes, by sorbing onto mineral surfaces or complexing with organic matter. However, one question had been left largely unanswered: If DNA is so stable in soil, how much of the signal from DNA-based molecular techniques is derived from dead microbes? In other words, by analyzing soil microbial communities using DNA based methods, are we inadvertently taking a ‘census of the dead’ and capturing not only live and intact microbes, but also the DNA from microbes that died and left their DNA behind?
The project started in the summer of 2014 with Patrick Marsden, an undergraduate at the Metropolitan State University of Denver, under supervision from Tess Brewer, a PhD student in the Fierer Lab at the University of Colorado. Pat and Tess were interested in optimizing methods to assess microbial viability in soils. One commonly-used method to quantify microbial viability is to treat samples with propidium monoazide (PMA), a photoreactive PCR inhibitor that is excluded by cells with intact membranes. Pat carefully figured out how to optimize this approach to work with soil. For example, he found that if you mix soil with a buffer containing PMA and that slurry was too dense, there would be insufficient light penetration to activate PMA. Conversely, if that slurry was too dilute, DNA yields would be too low. Optimizing the procedure to work with a wide range of soils was painstaking work, but Pat generated enough preliminary data to suggest that relic DNA was reasonably abundant in soil and worthy of further study.
In the summer of 2015 I joined the Fierer lab after making the decision to switch my primary research foci from marine to terrestrial microbiology for the next phase in my career. After a brief discussion of the initial results with Noah, I took the lead on this project. That August, several Fierer lab members, including Noah, Pat and myself collected a range of soils from local field sites and treated them with PMA to remove relic DNA. The data we obtained surprised us. In some samples, most of the DNA was removed by PMA treatment, indicating DNA from dead cells was more abundant than DNA from living cells in those samples. Importantly, we did not find relic DNA in all samples, and some soils had only moderate amounts of relic DNA, suggesting there is natural variation in the amounts of relic DNA present in soil. Using Noah’s network of colleagues, we obtained additional samples from around the United States to determine whether other soils showed the same patterns, to correlate the amounts of relic DNA to soil characteristics, and to determine whether removal relic DNA alters the apparent diversity and composition of soil bacterial and fungal communities.
There are two important implications from our study. First, the impact of relic DNA on the characterization of microbial communities is dependent on the questions that are being asked, and it might not be necessary to account for relic DNA in every study of soil microbial ecology. For example, soils from very distinct ecosystem types will likely harbor distinct microbial communities regardless of whether relic DNA is removed. However, when studying a single soil over time or across experimental treatments (e.g. artificial warming), removal of relic DNA is likely to be very important. We just completed sampling for a year-long study in the foothills of the Rocky Mountains where we are explicitly testing whether relic DNA enhances our ability to detect temporal patterns in soil microbial communities. While we do not yet have the complete dataset, analysis of first six months indicates that removing relic DNA does ‘unmask’ seasonal patterns in soil communities by removing the legacy of DNA left behind by the corpses of microbial cells that are no longer alive.
To some degree, this work generated more questions than answers. Such questions include: What is the average age of relic DNA in soil, and how long does this relic DNA persist in soil? Can relic DNA be used as a sort of ‘fossil record’ in soils or sediments to investigate what microbes were once living at a given site? And, to what extent does relic DNA persist in other environments? Now that DNA-based approaches are routinely used to investigate microbial communities in a wide range of systems, if we do not remove relic DNA, are we biasing our results by including the corpses of microbes when we are instead trying to assess living microbial populations?
Click here for the paper.