Have the cake and eat it: Optimizing nondestructive DNA metabarcoding of macroinvertebrate samples for freshwater biomonitoring

DNA metabarcoding can contribute to improving cost-effectiveness and accuracy of biological assessments of aquatic ecosystems, but significant optimization and standardization efforts are still required to mainstream its application into biomonitoring programmes. In assessments based on freshwater macroinvertebrates, a key challenge is that DNA is often extracted from cleaned, sorted and homogenized bulk samples, which is time-consuming and may be incompatible with sample preservation requirements of regulatory agencies. Here, we optimize and evaluate metabarcoding procedures based on DNA recovered from 96% ethanol used to preserve field samples and thus including potential PCR inhibitors and nontarget organisms. We sampled macroinvertebrates at five sites and subsampled the preservative ethanol at 1 to 14 days thereafter. DNA was extracted using column-based enzymatic (TISSUE) or mechanic (SOIL) protocols, or with a new magnetic-based enzymatic protocol (BEAD), and a 313-bp COI fragment was amplified. Metabarcoding detected at least 200 macroinvertebrate taxa, including most taxa detected through morphology and for which there was a reference barcode. Better results were obtained with BEAD than SOIL or TISSUE, and with subsamples taken 7-14 than 1-7 days after sampling, in terms of DNA concentration and integrity, taxa diversity and matching between metabarcoding and morphology. Most variation in community composition was explained by differences among sites, with small but significant contributions of subsampling day and extraction method, and negligible contributions of extraction and PCR replication. Our methods enhance reliability of preservative ethanol as a potential source of DNA for macroinvertebrate metabarcoding, with a strong potential application in freshwater biomonitoring.