Evaluating Fungal Diversity
with Environmental Sampling

© Brian A. Perry
Original publication: Mycena News, November 2008

In one of her numerous contributions to this column, Else Vellinga made the timely observation earlier this year that mycologists will quite soon be moving beyond the traditional methods used to assess the fungal diversity present in our forests and other habitats. Instead of exploring such environments with basket and favorite collecting knife in hand, soon we will likely carry an array of small tubes and plastic bags in which to place plant tissue samples, leaf litter, and soil cores, which will then be taken back to the lab and liberated of the fungal DNA they harbor. At least when we are not just collecting for the table, that is. Similar methods of sampling environmental DNA have greatly expanded our understanding of the biology and diversity of bacteria and Achaea, and are beginning to do the same for our knowledge of Fungi.

Estimates of fungal diversity suggest that we have described less than 5% (Hawksworth, 1991, 2001), and in some estimates less than 2% (O’Brien et al 2005), of the fungi present on our planet. When you stop to think about these numbers, they are quite staggering, especially when one considers that we have been discovering and naming species of fungi for well over 200 years. How is it that we are missing so many of these fungi? Well, as all of you know, not all fungi produce the macroscopic, complex sexual structures (i.e., mushrooms) most of us wander the woods in search of. In essence, only a fraction of the fungi do this. A vast majority of fungi are microscopic, living in the soil, litter, water, and in close association with plants, animals, and other organisms. Due to the nature of these fungi, they are typically overlooked using traditional survey methods. Many of these fungi also have very specific substrate requirements and/ or very slow rates of growth, and are therefore not recovered in studies attempting to isolate them from their substrate or host/symbiont using culturing methods. For these reasons, and others I will touch on later, a method of assessing fungal diversity that does not rely on the production of macroscopic structures, substrate requirements, or growth rate (such as environmental DNA sampling) is very attractive to many researchers.

The technology for sampling DNA from the environment has been around for a number of years, and the constant advancements in DNA sequencing continue to make faster and cheaper the processing of large volumes of material for researchers. In a very simplified view of the process, DNA is isolated from environmental samples of soil, plant tissues, etc., and the portion of the genome of interest (i.e., gene) is amplified (many thousands of copies are made). Because there is likely DNA from many different organisms present in the environmental samples, fungal specific primers (small fragments of DNA that recognize and bind to the region of the genome desired for amplification) are typically used. Next, because DNA from multiple species of fungi is typically present in the samples and amplified together in the same tube, individual copies of the gene amplified are isolated via cloning. Now, we’re not making sheep here folks; we are just taking a single fragment of DNA, forcing it into the genome of the bacterium Escherichia coli, and allowing these bacterial cells to make many copies for us as they replicate themselves. Simple, huh? Right. Once these cloning reactions have been screened to confirm the presence of the fungal DNA of interest, it is sequenced. As you can imagine, this process typically results in numerous cloning reactions and, potentially, numerous molecular DNA sequences that represent different species of fungi present in the same environmental sample.

Sampling of environmental DNA has already had great impact on our understanding of fungal diversity, indicating higher than expected levels of species richness in various habitats, and even revealing entirely novel lineages. Once researchers have obtained DNA sequences from environmental samples, these are typically compared with sequence data existing in the large, open access database known as GenBank. Ideally, molecular sequence data from the environmental samples will be a close match to a sequence in GenBank, or an existing molecular data set that was obtained from a known and properly identified fungal specimen. Depending on the level of similarity, such matches allow researches to make inferences regarding the identity of environmental samples. As you can imagine, there has been a bit of disagreement among researchers regarding what level of similarity is sufficient to identify unknown sequences at various taxonomic levels. One aspect nearly all studies agree upon, however, is that all environments sampled thus far contain greater than expected levels of diversity for saprotrophic, mycorrhizal, and endophytic fungi (Lynch & Thorn, 2006; Higgins et. al.,2006; O’Brien et. al., 2005). A study of tundra soils in Colorado (Schadt et. al., 2003), and a subsequent analysis of soils from North America, Europe, and Australia (Porter et. al., 2008), revealed the presence of previously unknown lineage of Ascomycetous fungi. These fungi, which are currently being referred to as “Soil Clone Group I,” are known only from molecular sequence data. This presents a bit of a problem. If these organisms are to be recognized as formal taxa, how does one designate a type specimen for a taxon known only from sequence data? Designate the soil from which the DNA for such sequences was isolated? Such issues will undoubtedly become more common as the sampling of environmental DNA increases.

We must, of course, realize that such methods are not just applicable to samples of DNA isolated from the environment, but may also be applied to any fungi we encounter. I am sure that we have all collected mushrooms and other fungi whose identification eluded us. And for those familiar with attempting to key out such specimens, you know one of the most difficult hurdles to overcome can often be simply accessing the necessary literature. The situation is no different for researchers, and in many cases may even be worse if they have large numbers of specimens to identify, especially those from taxonomically difficult groups. A recent study by Geml et. al. (2008) assessed the molecular diversity of one such genus, Agaricus, in boreal and arctic habitats across Alaska. Molecular sequence data from these undtermined specimens was subjected to phylogenetic analyses within a broad sampling of sequences from identified Agaricus species. The results of their analyses indicate that two sections of the genus, Arvenses and Agaricus, are prevalent in the boreal and arctic habitats of Alaska, respectively. Based upon various levels of sequence identity, these authors also concluded that between 11 and 13 “operational taxonomic units” of Agaricus are present in the areas of Alaska sampled.

The concept of operational taxonomic units, or OTUs, is just one of the many pitfalls that one must consider when conducting or reading about these types of studies. As I mentioned earlier, researchers disagree as to the level of DNA sequence identity that can be used to reliably place undetermined sequences at various taxonomic levels. For these reasons, most researchers are hesitant to recognize undetermined or environmental sequences at the species level for anything less than a 100% match in sequence identity, preferring rather to consider then as OTUs. For example, in the study by Geml et. al. (2008) of Alaskan Agaricus, different numbers of OTUs could be recognized depending on whether the researchers used a 95% or 98% sequence identity criterion, and also depending on the method used for sequence comparison. We know that various regions of genomes evolve at different rates, and therefore that gene selection will also have a large influence on the amount of divergence we would expect to see between closely related species, genera, etc. To confound things even further, we also know that different lineages of fungi evolve at different rates, so that the amount of change we observe in a genus like Agaricus may be quite different than that we observe in Morchella for a given gene. For these reasons, it will likely not be possible to find discrete levels of sequence identity we can utilize to recognize undetermined sequences of fungi at the levels of species, genera, etc.

One final point I would like to make about environmental DNA sampling methods, and bar coding in general, is that they are not and likely never will be a replacement for traditional morphological systematic studies, nor are they meant to be. Remember that all of these molecular methods rely upon a certain level of baseline data. If there were not researchers out collecting mushrooms and other fungi—which are in turn identified, preserved in herbaria, sequenced, and placed in GenBank—none of the environmental or otherwise undetermined sequences would have a framework upon which to be evaluated. We must realize, however, that given the current estimates of fungal diversity, it is not realistic to assume that we will ever be able to document all fungi using traditional methods. Environmental DNA sampling methods appear to promise a quick, efficient, and in some cases the only means of assessing fungal diversity in a number of habitats. Such methods could be especially useful to assess the fungal diversity of endangered or otherwise threatened environments. It is for these reasons that I believe both traditional methods of taxonomy and systematics should work in concert with studies utilizing environmental DNA sampling and molecular identification of undetermined fungal specimens.

Further Reading: