Discussion

eDNA detection

Prevention of false negatives is an issue that receives much attention in monitoring freshwater biodiversity using environmental DNA (e.g. Darling & Mahon 2011; Buxton et al., 2017). Therefore, a better understanding of the limits of eDNA detection is essential. This study shows that eDNA of live shrimps degrades faster in the presence of OM, resulting in reduced amounts of detectable eDNA, especially when pH is low, as might be found in peat bogs. We found the level of PCR inhibition to be unaffected by pH or the presence of OM. Therefore, detection of reduced amounts of eDNA when OM was present must be explained by a decline in rate of decay and by a failure to sample eDNA instead of by PCR inhibition. As spiked DNA degraded significantly faster than eDNA we believe most eDNA detected in natural systems must be contained inside cells or mitochondria. This is in line with findings of Turner et al., (2015) who found that only a minor fraction of carp eDNA to be extra-cellular. Dupray et al., (1997) reports that heat-killed cells of Salmonella typhimurium persist in seawater longer than purified DNA. Nielsen et al., (2007) show that the residence time of bacterial DNA in soil is generally longer when dead cells are used as DNA source compared to purified DNA.
In aquatic environments, DNA is known to degrade faster in the presence of DNA-consuming microorganisms (Alvarez et al., 1996; Duprayet al., 1997). The longer persistence of cellular DNA can be explained by the presence of cellular compounds such as cell membranes that form a barrier against DNA consuming microorganisms and nucleases in the environment (Dupray et al., 1997).
Humic acids can strongly adsorb DNA, probably by ligand binding, hydrophobic interaction, aggregation or precipitation (e.g. Saekiet al., 2011), and eDNA therefore might have been adsorbed to organic particles that were deposited at the bottom of the aquariums. Stotzky (2000) found that DNA bound to humic acids and clay-humic acid complexes becomes more resistant to degradation by DNases. However, as we infer that most eDNA is cellular, these processes might have a minor effect on eDNA contained in cells or mitochondria. Sampling of organic material or sediments might increase the yield of target eDNA, though PCR might be inhibited by organic acids in such cases. However, sampling of organic material or sediments might result in detection of historical eDNA, not representing the actual presence of target species (Olajoset al., 2018).
In a comparable experimental set-up to ours, using tanks, Buxtonet al., (2017) found the effect of pH on eDNA survival to be insignificant (which is in line with our findings), but that sediment has a strong effect. The authors conclude that especially “ponds with organic sediment types—or sediments that become suspended easily—can be a source of false negative results” (Buxton et al., 2017). Remarkably, in our aquarium treatment B (high pH and no added OM), eDNA could be detected more than six weeks later, whereas other studies found that eDNA degrades beyond detection ability within two weeks (Dejeanet al., 2011; Thomsen et al., 2012a; Thomsen et al., 2012b; Strickler et al., 2015; Eichmiller et al., (2016). However, the eDNA concentrations in these aquariums were unnaturally high, thus not reflecting a natural situation. This might have resulted in relatively high amounts of detectable eDNA and probably lengthened eDNA survival.
Several studies show a correlation between eDNA concentration and population density (Maruyama et al., 2015; Wilcox et al.,2016; Baldigo et al., 2017). This study, as well as previous studies (Strickler et al., 2015; Echmiller et al., 2016) show that environmental conditions strongly affect eDNA concentration. We therefore believe caution is warranted when using eDNA concentrations as proxy for population density. Environmental conditions might specifically affect eDNA concentrations on the sampling site. Therefore, it is necessary to correct measured eDNA concentrations for local environmental conditions such as pH and amount of OM.
Our study, as well as previous studies, focused on selected environmental factors only and was conducted in an artificial ecosystem (i.e. an aquarium) (Nielsen et al., 2007). Complex interactions between eDNA degradation and additional factors such as the presence of DNA-consuming microorganisms remain largely unknown, and future studies should therefore include microbial activity as well. In addition, species that occur in a wide range of habitats should be used to investigate the relation between amount of detectable eDNA and other environmental conditions in the field such as seasonality (de Souzaet al., 2016) or soil type (Buxton et al., 2017).