Comparison of eDNA and eRNA methods
@Zerina Rahic
eDNA
Dead biomaterial or extracellular DNA can be transported into a sampling region from a significant distance, therefore the detection of eDNA does not necessarily confirm the presence of living organisms, nor automatically indicate that live organisms occur in close proximity (Macher and Leese, 2017; Cristescu and Hebert, 2018).
eRNA
In contrast, environmental RNA (eRNA) is believed to deteriorate more rapidly due to the chemical composition (hydroxyl groups) which makes this molecule more prone to hydrolysis or degradation (Dowle et al., 2015; Guardiola et al., 2016; Laroche et al., 2016, 2017). Environmental RNA may therefore provide a better proxy for inferring the presence of living organisms (Thomsen et al., 2012a; Sassoubre et al., 2016; Pochon et al., 2017; Cristescu, 2019). However, working with eRNA requires specialized storage of samples, and expensive and time-consuming workflow protocols; potentially limiting its applicability to routine monitoring programs (Wood et al., 2019a).
Methods
Metabarcoding
Metabarcoding utilizes universal primers that target taxonomically informative genes such as, the nuclear small subunit ribosomal RNA (18S rRNA) or the mitochondrial Cytochrome c Oxidase subunit I (COI) genes (Tanabe et al., 2016; Stat et al., 2017; Bista et al., 2018; Wangensteen et al., 2018). In the context of surveillance for marine NIS, this approach holds great potential but has limitations, including challenges in identifying NIS at species level due to the lack of sufficiently resolved phylogenetic markers, incomplete reference databases, primer biases and sequencing artifacts, which all may lead to false positive or negative results (Brown et al., 2016; Ammon et al., 2018; Cristescu and Hebert, 2018)
Target method
Targeted methods, e.g., species-specific qPCR, may offer a more sensitive approach for effective detection of specific marine NIS (Wood et al., 2017). However, species-specific assays need to be designed based on a priori knowledge of target organisms. Droplet digital PCR (ddPCR) is a real time PCR technology that divides eDNA/eRNA template into thousands of nanoliter droplets, each containing a single target molecule. Within each droplet, a PCR is conducted, and the outcome visualized via the presence or absence of a fluorescence signal. The number of target copies can be calculated on the positive-negative droplet relation, allowing direct quantification without the need for standard curves (Baker et al., 2018). When using ddPCR, the parallel processing of thousands of reactions enables the detection of very low target concentrations while minimizing PCR inhibition and removing the need for technical replicates, thereby reducing analysis costs and time (Nathan et al., 2014; Doi et al., 2015).
Limitations of both method
There is still limited knowledge on the factors affecting detection probabilities. (Wood et al., 2019b). For example, there is a need for more research to determine if the complexity of sampling matrices affects the detection efficiency and whether eDNA binds to certain environmental matrices for longer periods of time. Furthermore, additional information on the relationship between eDNA and eRNA signals will assist in determining whether the use of eDNA in isolation can accurately predict if living organisms are present near the collection source, thus making these tools more cost-effective for routine biomonitoring programs.
Material and Method
Briefly, water from the three depths was combined and pre-filtered through a 20 μm plankton mesh. To condense all biomaterial for eDNA/eRNA extraction, seawater (ca. 50 ml) from each site was used to re-suspend the material captured on the mesh and the re-suspended material was filtered onto nitrocellulose membrane filters (pore size 0.45 μm; Merck KGaA, Darmstadt, Germany). Filters were cut in half, each half stored in LifeGuardTM Soil Preservation Solution (QIAGEN, Hilden, Germany). Samples were frozen (−80°C) immediately until further processing.
Between each sampling location, all sampling equipment was soaked in 5% bleach (sodium hypochlorite) solution for at least 5 min and rinsed with water from the new sampling location to prevent cross-contamination.
- Laboratory analysis
- DNA/RNA Extraction
- Droplet Digital Polymerase Chain Reaction
- Metabarcoding of Cytochrome c Oxidase I (COI) and Bioinformatics
- Statistical analysis
Results
Sabella spallanzanii was detected in 35% of RNA samples (14% plate and 55% water) and 68% of all DNA samples (55% plate, 77% water). Sabella spallanzanii was detected in 44% of DNA samples with no corresponding RNA detections (20 plate samples and 23 water samples). In contrast it was detected in 4% of the RNA samples with no detections in the corresponding DNA samples (1 plate sample and 3 water samples) (Supplementary Table S1).
At phylum level, the taxonomic composition of the water samples was more diverse than the plate samples (see also Supplementary Table S2). The majority of eRNA ASVs were Arthropoda (34%), Mollusca (12%), Bacillariophyta (11%) and Porifera (9%). The eDNA dataset was similar with lower abundance of Arthropoda (25%) and Mollusca (10%), but higher abundance of Porifera (19%), and was characterized by a slightly higher proportion (21%) of unassigned taxa
Discussion
For the purposes of marine biosecurity surveillance discriminating between the detection of “legacy” signals and live organisms, for example when evaluating the success of an eradication program or compliance control, is preferable (Pochon et al., 2017; Zaiko et al., 2018). Since RNA is directly linked with active gene expression of metabolic pathways and deteriorates rapidly after cell death, it may be a better proxy for detecting “live” signals in environmental samples (Darling et al., 2017; Cristescu, 2019). However, cellular RNA production can vary enormously (over 3 orders of magnitude), largely due to varying transcription rates of ribosomal RNA (Fegatella et al., 1998). Additionally, working with RNA requires the conversion of RNA into cDNA which introduces additional costs and processing time. While ddPCR technology is particularly sensitive (Wood et al., 2019b), it is not immune to inhibition (Racki et al., 2014; Goldberg et al., 2016). The influence of community diversity and different sample matrices is still relatively unexplored (Zaiko et al., 2018).
eDNA vs eRNA
There is not much empirical information available on eRNA degradation in the marine environment, but it is assumed to degrade significantly faster than eDNA due to its more fragile chemical structure (Dowle et al., 2015; Guardiola et al., 2016; Laroche et al., 2016, 2017). A recent in situ study specifically on S. spallanzanii could trace DNA signals for up to 42 h, while RNA could not be recovered after 13 h of organism removal from the tank (Wood et al., in preparation). The decay rate is likely accelerated in natural settings depending on biotic and abiotic parameters (Dell’Anno and Corinaldesi, 2004). Unexpected observations showed a decline in specific eDNA degradation under increasing oxygen demand, chlorophyll and total eDNA, further highlighting the need to better understand these correlations (Barnes et al., 2014). Another complexity is a lower RNA recovery rate for certain preservation and extraction procedures, which can yield as little as just 5% of the original concentration and should be tested individually for each study (Lebuhn et al., 2016).
Effect of Sample Matrices on the ddPCR Detection Signals
Most marine or freshwater studies that use eDNA for monitoring purposes currently focus on water as the preferred sample medium, as it provides a homogenously distributed sample matrix and can be concentrated relatively easily through filtration (Thomsen et al., 2012b; Rees et al., 2014; Smith, 2017) but contradicting opinions exist, e.g., Goldberg et al. (2016). Sampling other matrices such as sediments may reduce detection probability due to patchy distributions of target eDNA (Andersen et al., 2012) and overall lower spatial coverage due to restricted starting material afforded by current eDNA isolation methods.
Sample type and volume have been shown to have a significant effect on recovered biodiversity but also on particular detections of certain organisms (Moyer et al., 2014; Nascimento et al., 2018). Sampling strategies need to be adapted toward the target species’ biological traits and life cycle (Rees et al., 2014; Furlan and Gleeson, 2017; Harper et al., 2018; Holman et al., 2019).
References