We demonstrated that RNAlater was highly effective for preserving fish eRNA on a filter at approximately − 20°C. The eRNA concentration was tens to hundreds of times higher in filter samples preserved in RNAlater than in those without any preservative (control samples). In particular, eRNA concentrations relative to the control samples were much higher for the September samples (eRNA was extracted a week after filtration) than the August samples (eRNA was extracted five hours after filtration) when using RNAlater. The longer storage time prior to RNA extraction for the September samples than the August samples during sample transportation in a freeze box and storage in a refrigerator may have resulted in higher relative eRNA concentrations from the samples in RNAlater. Our results support the technical validity in previous studies using RNAlater to preserve and reliably detect microbial, bulk, and fish eRNA in natural environments (e.g., Bachoon et al., 2001; Gorokhova, 2005; Miyata et al., 2021), and will boost the application of eRNA for the non-destructive and cost-effective monitoring of macrobial physiology at the population- and community-levels (Yates et al., 2021).
Although the LBP buffer had a certain capacity for eRNA preservation, its performance was much lower than of RNAlater. According to the user manual (https://www.mn-net.com/media/pdf/83/75/16/Instruction-NucleoSpin-RNA-Plus.pdf), LBP buffer contains guanidinium thiocyanate, a type of chaotropic salt frequently used for efficient RNA extraction (Chomczynski, & Sacchi, 2006). It is likely that the LBP buffer froze in a freeze box (− 16°C or below) and may not have been able to fully function to preserve RNA molecules and inactivate ribonucleases (RNases). Furthermore, it was surprising that ayu eRNA could be detected even from the filters that were stored at − 20°C without any preservative. Of course, this was mainly attributed to its much higher concentrations in an aquarium than in the natural environment. However, although RNA is thought to be a very fragile and instable molecule and rapidly degrade in vitro, the detection of eRNA from control samples implies that macrobial eRNA is likely to exist in the water in the form of a protected and intra-cellular state (i.e., protected by cells, extracellular vesicles, and protein capsids), which can increase its persistence and abundance, as hypothesized in some previous studies (Cristescu, 2019; Sigsgaard et al., 2020; Wood et al., 2020; Jo et al., 2022).
The methodological difference in RNA extraction between preservation treatments could also affect our results. Because RNAlater itself is unnecessary for RNA extraction, filter samples in RNAlater were centrifuged before the cell lysis step (i.e., the addition of LBP buffer to the filter) to remove excess RNAlater from the filter. This preliminary centrifugation could also remove residual water in the filter, likely increasing subsequent cell lysis and overall eRNA extraction efficiency. In the other treatments without the preliminary centrifugation (control and LBP buffer), residual water in the filters may have diluted the reaction concentrations of some reagents for RNA extraction, decreasing cell lysis and RNA binding efficiencies. In addition to residual water, multiple abiotic factors (e.g., minerals, nutrients, and chlorophyll) can negatively influence the efficiencies of DNA and RNA extraction and even PCR (Schrader et al., 2012; Thomsen et al., 2012; Eichmiller et al., 2014; Uchii et al., 2019). Regardless of the preservation method used, filter dehydration is preferable before eRNA (and likely eDNA) extraction from a filter (Fabre et al., 2014; Thomas et al., 2019).
It is clear that eRNA preservation via RNAlater can makes large-scale eRNA applications simpler and more efficient, although further optimization is still required for eRNA preservation. First, we did not compare the yields of ayu eRNA in each preservation treatment with the “fresh” eRNA sample that was extracted immediately after water filtration, which was due to the limitations on preparing space for eRNA extraction and cDNA synthesis at the sampling station. Second, Rissanen et al. (2010) reported that RNAlater preservation could drastically decrease the yields of bacterial RNA in environmental samples that contained inhibitor substances such as humic acids. Although filter centrifugation and dehydration can mitigate the negative effect to some extent, the utility of RNAlater for preserving macrobial eRNA independent of a deep freezer should be tested in multiple environmental conditions. Third, eRNA preservation reagents other than RNAlater should be considered. For example, isolated and purified RNA molecules have traditionally been preserved in an alcohol such as ethanol (Rio et al., 2010; Walker & Lorsch, 2013), and β-mercaptoethanol is occasionally used to inactivate ribonucleases in a sample (Chomczynski, & Sacchi, 2006; Mommaerts et al., 2015). Littlefair et al. (2022) preserved filter samples in a buffer for RNA isolation (RLT buffer; Qiagen) with β-mercaptoethanol at − 20°C until eRNA extraction, showing that the eRNA analysis detected more fish species known to inhabit the study lakes than the eDNA analysis. However, as β-mercaptoethanol is a pungent and toxic reducing agent and harmful to humans, the specific context may inform whether the reagent is suitable for eRNA preservation.
Moreover, although our study targeted mitochondrial genes as an eRNA biomarker, it would be more desirable to target nuclear genes, especially their messenger RNAs (mRNAs), to profile the physiological information of a community. However, compared with mitochondrial genomes for which tens to hundreds of copies exist per cell and nuclear ribosomal RNAs (rRNAs) that account for more than 90% of all RNAs in the cell, the proportion of mRNA abundance per cell is too low (less than 1%) and therefore their abundance and availability in the field is expected to be much lower. If RNAlater can effectively preserve such environmental transcriptomes (Yates et al., 2021), the reagent will contribute further to the practical application of eRNA-based meta-transcriptomics to achieve the non-invasive profiling of gene expression within a community and its response to specific environmental conditions. We believe that our findings have provided valuable information to refine the methodology for improving eRNA quality and quantity in environmental samples.