As we hypothesized, the applicability of using SPM as source for fish eDNA metabarcoding has been confirmed and used for first time in this study. Fish species were found in all samples, irrespective of the location or characteristics of the SPM sampled.
Comparing the different extraction methods used, the eDNA extracted from SPM samples using a modified protocol of the DNeasy PowerSoil Kit, presented the highest purity (260/280 nm ratio) in combination with high DNA concentration, therefore it was the method selected for metabarcoding the eDNA extracted from the nine sampling sites. The isolation method was chosen due to its simplicity and scalability to perform a high number of extractions. However all tested methods resulted in high DNA concentration, making them suitable for metabarcoding, even if post extraction cleanup would have been needed (e.g. the Magnetic Forensic kit showed lower purity 1.44 (260/280 nm ratio)).
While eDNA-based fish monitoring from filtered water samples has been widely used and described and has cheap setup costs, it provides only a snapshot of the diversity at the sampling point, while continuous integration and eDNA settling in time-integrative sampled SPM would provide a better reflection of long-term site occupancy15,16,17,18. On the other hand, eDNA extraction from water samples using filters are laborious and extractions yields are low. The process of particles sinking or binding of eDNA (or residues of, e.g. fish tissue, feces or shales containing eDNA) to organic or mineral particles in SPM18 may result in a progressive accumulation of eDNA in the SPM. This statement was confirmed in our study. The results showed that using one SPM sample yielded higher DNA amounts per extraction (400–2,500 ng) than what is reported for eDNA extracted from an individual water sample using filters (30–560 ng)18,21,22,23,24,25. Here a small amount of SPM (~ 250 mg) is sufficient to extract high amounts of eDNA, which is of particular importance for the detection of rare fish species, where the concentration of their DNA is expected to be low. For example, Salmo salar which is classified as endangered in German rivers26, was detected in the Koblenz, Weil, and Blankenese SPM samples. Another main advantage of using SPM (in particular archived in the ESB), is that it is possible to retrieve and reanalyze the source material, allowing repeats and other complementary analyses e.g. chemical analysis to determine the presence of contaminants or stressors responsible for changes in fish populations. This kind of repeat analysis are not possible with filtered water samples, unless multiple samples are taken in parallel or the water itself is retained, both costly options.
Here, eDNA metabarcoding of the 9 riverine sites detected a total of 29 fish species. Most taxa found belong to commonly detected species in large rivers in Germany. For example, Abramis brama, Rutilus rutilus, Barbus barbus, Squalius cephalus, and Perca fluviatilis and are largely overlapping with the regulatory monitoring data from the Water Framework Directive (WFD)27. This coherence of fish species identified from eDNA extracted from SPM with the commonly detected fish species demonstrated the suitability of this approach. However, the number of fish species found in the ESB samples is similar or lower to what was found using traditional fish monitoring techniques, e.g. electro- and netfishing under the WFD27. For example ,with regard to monitoring sites in Germany between 27 and 57 fish species have been detected in 2012 and 2013 along the Rhine28, between 19 and 24 fish taxa were counted in 2007 at four sites of the river Elbe and between 27 and 29 fish species were detected at three sites of the Danube29. However, it needs to be considered that the number of WFD surveillance monitoring sites is much higher than the ESB sampling sites investigated in this study.
The fish community analysis also evidenced the presence of two contaminant species: Danio rerio and Oryzias latipes. For this reason, the extractions from the 9 sampling sites were repeated retrieving new subsamples from SPM, and before sequencing the absence of contaminant species (e.g. Danio rerio) was validated using specific qPCR primers (See Supplementary information). The specie-specific qPCR and the metabarcoding results showed successful removal of exogenous lab- contaminant fish species (See Supplementary information). The detection of those reads in the first samples strongly suggests cross-contamination in the laboratory since Danio rerio is a specie that we used commonly in our facilities for other purposes. It is well known that the most serious pitfall of metabarcoding eDNA is the risk of contamination with exogenous DNA30,31.
At the stage of PCR during library preparation, several samples exhibited unspecific amplification (double banding), Prossen, Weil, Bimmen and Dessau, which might be indicative of bacterial amplification. This additional bacterial amplification might have resulted in less efficient fish-specific sequencing and in consequence, a lower number of species found in those samples (5–9 species found compared to 8–17 species found in the non-contaminated samples). However, the richness is not only attributable to the presence or absence of contamination but might be also inherent to the sample. Contamination of reagents with bacterial DNA, or contamination with exogenous DNA in the laboratory (e.g. Danio rerio), in combination with the bacteria inherent to the sample itself, is a major problem exacerbated by the highly sensitive nature of the PCR, in particular when using universal primers. Therefore, even minor presence of these species in the lab equipment (like pipettes, surfaces, etc.) might result in large non-target amplification. To avoid such risk, we performed decontamination procedures for laboratory spaces and equipment (with UV radiation) and physically separated pre- and post-PCR workspaces.
The results of this proof-of-concept study will open the door for the retrospective evaluation of SPM samples to study, for example, seasonal and temporal trends of invasive species. The present study can be regarded as a first step towards more comprehensive investigations using eDNA extracted from archived SPM of freshwater fauna, flora and microorganisms. The fish taxa detected in this study complement well with species sampled in fish monitoring with traditional methods, e.g. nets, fykes and electrofishing. However, to study the fish community of a particular sampling site and draw conclusions on differences among sites, further investigations and more stringent analyses are required. The definition of a methodology should include an eDNA extraction strategy considering, for example, SPM extraction volume, the number of replicate extractions, the number of independent sequencing analyses required vs pooling the extracted DNA, etc. In order to validate this proof-of-concept study, future work will focus on method optimization and comparisons with established monitoring approaches.
Source: Ecology - nature.com
