Unveiling structure of tropical estuarine communities through eDNA and implications for biomonitoring

Tropical estuaries are hyper-diverse ecosystems, hosting essential habitats for freshwater, euryhaline and marine life. Understanding how biological communities are distributed in these systems has long been a challenge because of their inherent dynamic nature, and the diversity of interacting natural pressures and anthropogenic stressors they are subjected to. In this study, we used environmental DNA (eDNA) metabarcoding to examine the structure of multi-taxonomic communities in estuarine ecosystems (diatoms, crustacean, fish and eukaryote as a whole) and their relationships with environmental drivers in three differentially impacted locations facing the Great Barrier Reef in Central Queensland (Australia). We first demonstrated that eDNA signals from sediment and water matrices provide complementary information, and that both should be monitored for a more holistic understanding of community trajectories in anthropogenically-impacted aquatic environments. We also observed that, independently of the taxonomic group considered, communities were primarily structured by the ecological conditions of the estuary. A within-estuary differentiation along an upstream-downstream gradient was detected but only for small-bodied organisms, which further adds credence of eDNA approaches as an ecologically relevant tool for monitoring fine-scale biodiversity patterns even in profoundly dynamic environments. Finally, the different communities exhibited contrasting response patterns, in terms of diversity, composition and uniqueness, to the anthropogenic gradient. Hence, our findings emphasize the need for multi-taxonomic assessments, for which eDNA is well-suited, to better understand the impacts of multiple stressors on biodiversity, and thereby assist decision makers in the protection and management of tropical estuaries.

Methods
This dataset comprises environmental DNA datasets (both raw and filtered datasets along with associated scripts) and physico-chemical data obtained from water and sediment collected in three different estuaries from Queensland, Australia in 2018. 10 sites were selected at each estuary, at approximately 1 km intervals from the mouth of the river going upstream.

Environmental parameters of the water column (including salinity, pH, dissolved oxygen, turbidity, chlorophyll-a and temperature) were measured using a calibrated EXO2 YSI multiparameter sonde (YSI, Yellow Springs, OH, USA). Standard sampling protocols from the Queensland Government Department of Environment and Science were followed for determining contaminant (nutrient and pesticide) concentrations in the water column. Analysis of nutrients in water was carried out by the DES Chemistry Centre laboratory. Total and dissolved organic carbon were measured using a non-dispersive infrared sensor (NDIS). Total Kjeldahl values for nitrogen and phosphorus were measured through low-level atomic absorption (AA) analyses. For sediments, total organic carbon analyses were performed using a high temperature total organic carbon analyzer (Dohrmann DC-190, Teledyne Tekmar, Mason, OH, USA; Chariton et al. 2010) and particle size analysis was conducted by successive sieving through 500-μm, 180-μm, and 63-μm meshes and gravimetry. Metal concentrations in sediments were determined at CSIRO using a dilute-acid (1 M HCl) extractable metals method. Pesticides analyses in both water and sediment samples were performed by the Queensland Government, Forensic and Scientific services laboratory. Pesticides in water samples were analyzed by direct injection via liquid chromatography with tandem mass spectrometry (LC-MS-MS) while pesticides in sediment samples underwent a solvent extraction followed by QuEChERS solid phase clean up prior to the LC-MS-MS analysis.

For eDNA analyses, water and sediment samples were collected in triplicate. On the same day as sampling, each sediment sample to be used for eDNA analyses was homogenized and sub-sampled into 5-mL tubes. One liter of each water sample (i.e., a total of 3 L per site) was filtered on two 0.45 µm pore-size cellulose nitrate membranes using a peristaltic pump within 24-h of collection. All samples were kept frozen until DNA extraction. Environmental DNA analyses were performed in a dedicated facility at Macquarie University, Sydney. Water eDNA was extracted from filter papers using PowerWater kits while sediment DNA was extracted from ~0.5 g of sediment using DNeasy PowerSoil kits.

The following primer pairs have been used for targeting four different taxonomic groups. Eight base pairs (bp) tags, each differing by at least 5 nucleotides, were added to the 5’ end of each primer to enable the multiplexing of multiple PCR products into the same library before high-throughput sequencing. PCR amplifications were performed in triplicate. The list of primers used is described below:

-Euka01_F:5'-TGGTGCATGGCCGTTCTTAGT-3' & Euka01_R:5'-CATCTAAGGGCATCACAGACC-3' (18S rDNA V7 region; Hardy et al. 2010) for eukaryotes

-Baci01_F:5'-TCCAGCTCCAATAGCGTA-3' & Baci01_R:5'-AACACTCTAATTTTTTCACAGTA-3' (18S nuclear rDNA V4 region; this study) for diatoms

-Tele01_F:5'-ACACCGCCCGTCACTCT-3' & Tele01_R::5'-CTTCCGGTACACTTACCATG-3' (12S mitochondrial rDNA; Valentini et al. 2016) for fish

-Crust16S_F:5'-GGGACGATAAGACCCTATA-3' & Crust16S_R:5'-ATTACGCTGTTATCCCTAAAG-3' (16S rRNA; Berry et al. 2017) for crustacea

After PCR, amplicons were pooled and purified to generate one library per marker (i.e. per primers pair). Library preparation and paired-end sequencing were performed at the Ramaciotti Centre for Genomics (University of New South Wales, Sydney, Australia). The eukaryote library was sequenced separately on an Illumina MiSeq platform (2 × 250bp paired-end) while those for other groups were sequenced collectively on an Illumina NextSeq 100 (2 × 150bp paired-end).

Raw sequencing data were first curated using OBITools v.1.2 (Boyer et al. 2016) and Sumaclust software (Mercier et al. 2013). Additional filtering steps were performed in R version 4.0.3 using custom-made scripts. The scripts are provided in this repository and the procedure is described in detail in Pansu et al. 2025 (Environmental DNA). Filtered datasets along with scripts for statistical analyses are also provided.