A 9,000-year marine diatom assemblage reconstruction off Eastern Tasmania, Australia, using microfossil and molecular methods

Brief description

Over the past fifty years, Eastern Tasmanian waters have experienced rapid warming, primarily due to the extension of the East Australian Current. This has driven expansion of warm-water biota and decline of those adapted to cooler conditions, including phytoplankton. Presently, plankton monitoring, including diatoms along Eastern Tasmania, spans <100 years. This study reconstructed diatom communities throughout a sediment core spanning 9,000 years before present (9 kyrs BP), using microfossil analysis and molecular techniques, including sedimentary ancient DNA (sedaDNA) and NRS 18S rRNA from a 10-year water column archive at the Maria Island IMOS NRS mooring. Microfossil analysis revealed a dominance of strongly silicified benthic taxa (Campylodiscus, Diploneis, Paralia, Pyxidicula, Triceratium). Notably, Paralia sulcata showed a shift ~6 kyrs BP from small to larger cells, possibly reflecting a transition from a coastal to shelf ecosystem. However, microfossils underrepresented lightly silicified planktonic diatoms. Molecular methods detected higher diatom diversity, though up to 50% of sedaDNA reads remained unclassified due to reference library limitations. Lightly silicified planktonic genera (Chaetoceros, Corethron, Lithodesmium, Rhizosolenia) were identified only via molecular approaches and comprised 73% of sedaDNA and 88% of 18S rRNA records. Of 10 shared diatom families, 5, 15, and 4 were unique to microscopy, sedaDNA, and 18S rRNA, respectively. SedaDNA also captured greater benthic diversity. Our findings revealed limitations in reconstructing historic diatom assemblages from sediment cores. Microfossils faced constraints due to difficulties in morphological identification and preservation biases. In contrast, sedaDNA analysis yielded finer taxonomic resolution, provided access to high-quality reference sequence libraries were available.

Lineage

Maintenance and Update Frequency: asNeeded

Statement: 2.2 Sediment core collection and preparation In May 2018, a marine sediment gravity core (GC02-S1) measuring 268 cm in length was collected during the RV Investigator voyage INV2018_T02 at a water depth of 104 m near the continental shelf edge to the east of Maria Island (42.845°S; 148.240°E) (Figure 1b, red cross). To minimise disturbance caused by large gravity corers during sediment contact, a shorter, complementary core (12 cm) was obtained using a KC Denmark Multi-Corer at the same location, denoted as MCS1-T6 (where T6 refers to tube number 6 of the multi-corer). Additionally, a 35 cm-long multi-core (MCS3-T2) was collected from an inshore site within Mercury Passage on the western side of Maria Island (Figure 1b, orange triangle). All cores were hermetically sealed, labelled, and transported to the Australian Nuclear Science Technology Organisation (ANSTO) in Lucas Heights, NSW, Australia, where they were stored at 4°C. Further information on core collection and preparation can be found in Paine et al [11]. For this study, the inshore core MCS3-T2 was analysed at depth intervals of 5 cm from the core base (35 cmbsf) and then at intervals of 2 cm from 26 cmbsf to the surface. All MCS1-T6 samples were analysed at 2 cm depth intervals, while all GC02-S1 samples were analysed at 5 cm depth intervals. 2.3 Sediment age model Dating of MCS3-T2 was derived from eight Lead-210 (210Pb) measurements. Dating of MCS1-T6 was based on 210Pb at six depth intervals, and the dating of offshore deep core GC02-S1 used both 210Pb (7 depths) and radiocarbon (14C) from bryozoan fragments at three depths. A Bayesian age-depth model for each of these measurements was determined using the rbacon [30] software package within the R platform [31] with the SHCal20 curve for radiocarbon age calibration [32]. The deepest segment of GC02-S1 positioned 268 cmbsf was dated at ~8.9 kyrs BP. From comparisons of the 210Pb profiles of GC02-S1 and MCS1-T6, we estimate that 3.5 cm representing approximately 30 yrs were missing from the top of GC02-S1. This missing period was represented by the top of MCS1-T6. The deepest sample from MCS3-T2 inshore (35 cmbsf) was dated at ~144 yrs BP. 2.4 Diatom microfossil sample preparation and enumeration 2.4.1 Sample collection and subsampling A total of 74 sediment samples were collected from the three cores, 12 from MCS3-T2, 6 from MCS1-T6, and 56 from GC02-S1. Each sample was obtained by subsampling bulk raw sediment. A volume of approximately 2 cm3 was placed into individual 15 mL polycarbonate centrifuge tubes (Falcon) and dried in a Thermoline Scientific laboratory oven at 50°C and their weights recorded. The samples were then rinsed with distilled water to remove salts. 2.4.2 Hydrogen peroxide Treatment To eliminate organic matter from the sediment samples, 2 mL of 30% hydrogen peroxide was added to each sample centrifuge tube. The tubes were loosely covered with a lid to allow for evaporation and then placed in a stainless-steel test-tube rack. The test-tube rack was partially submerged in a heated water bath at 60°C. The reaction time varied from 30 minutes to 2 hours depending on the organic content of the sample. 2.4.3 Sediment washing and sieving After the hydrogen peroxide reaction was complete, the screw top was removed, and the centrifuge tube was topped up with distilled water. The screw top was replaced, and the tube was shaken and centrifuged at 700 x g for 10 minutes at room temperature. The supernatant was discarded, and the washing step was repeated a further two times (three times total). The sediment samples were then wet sieved through a 125 µm mesh, and the product that passed through the mesh was retained. The retained product was sieved again through a 15 µm mesh, and the captured product was retained in a clean polycarbonate centrifuge tube. The samples were then dried in a sample oven at 40°C for 48 hours. 2.4.4 Abundance reference marker Latex fluorescent beads (CC Size Standard L10, 10 μm average diameter, Lot 9747486, Beckman Coulter, Brea, CA) were rinsed twice with distilled water by centrifugation and resuspended in a solution at a concentration of 1650 beads/μl, as determined by a haemacytometer to create an "abundance reference marker," [33]. After drying, each sediment sample was weighed at room temperature, and 995 µL of distilled water and 5 µL of the microbead solution combined to create a slurry for smear slide analysis. 2.4.5 Smear slide and SEM preparation Standard protocols for preparing smear slides for light microscopy examination, as described by Tada et al. [34], were followed. Briefly, 150 µL of sediment slurry was pipetted onto a glass microscope coverslip (Menzel Gläser 22 x 22 mm) and smeared thinly with a stainless-steel laboratory spatula. The coverslip was allowed to evaporate over a hotplate, and one drop of Norland optical adhesive #61 was used to mount the smeared coverslip to a Knittel G300 26 x 76 x 1.0 mm microscope slide. The slide was then cured in sunlight. To prepare the sample for SEM analysis, the processed samples were filtered onto Isopore Millipore 13mm disc filters with a pore size of 1.2µm. Subsequently, the filters were allowed to air dry at room temperature and then affixed to Ted Pella standard SEM pin stubs with double-sided conductive carbon tabs measuring 12 mm in diameter. A layer of approximately 3 nm of platinum was deposited onto the prepared SEM sample stubs using a BalTec SCD 050 sputter coater. 2.4.6 Microscopy Diatom microfossils and fluorescent beads were observed and quantified using a Nikon Eclipse Ci light microscope. Systematic transects covering the entire slide were employed at a magnification of 400x using phase contrast and UV light. We counted damaged valves or frustules in cases where at least half of the structure was intact, or if we observed at least one pole along with the central area. Additionally, when deemed suitable, we considered fragments as equivalent to one-third of a complete valve. Our emphasis was on maintaining a consistent method for counting fragments, rather than strictly adhering to a rigid, predefined approach. The number of microfossils and fluorescent beads in each sample was tallied, and a quantification per g-1 of sediment calculated by relating the two counts. The formula used to calculate microfossils per unit measure was Ct = ((Tc/Lc) Ls / Wts), where Ct represents the concentration of microfossils, Tc/Lc represents the microfossil count divided by the fluorescent bead count, Ls represents the concentration of the fluorescent bead solution added, and Wts represents the weight of the sample. Taxonomic details and associated digital imagery were recorded throughout the enumeration process. Further taxonomic analysis utilised a Hitachi SU-70 scanning electron microscope (SEM). Diatom identification was aided by material from Jameson and Hallegraeff [35]; Saunders et al. [36] and the references within. Taxonomic assessment of diatom valves was extended at a minimum to genus level. However, for comparison with sedaDNA, categorisation to taxonomic ranks such as subclass and family proved more appropriate. We included size measurements for Paralia sulcata due to established correlations between environmental parameters and both its abundance and size, rendering it a potentially valuable paleo-indicator species [37], [38]. Preservation of diatom valves was assessed by implementing the following criteria documented by Tada et al. [34]: good (valves are intact, including the most fragile species; some breaking occurs), moderate (valves and most broken material show areolae and/or outside dissolution), poor (extreme dissolution and/or fragmentation prevents species identification). 2.4.7 Sedimentary ancient DNA (sedaDNA) extractions, library preparations and sequencing Work involving sedaDNA was carried out at the Australian Centre for Ancient DNA ultraclean ancient (GC02-S1) and forensic (MCS1-T6, MCS3-T2) facilities following decontamination standards [39]. Extraction of sedaDNA from sediment samples followed a ‘combined’ protocol developed explicitly for marine eukaryotes [40]. All extractions included seven extraction blank controls (EBC’s) whereby empty tubes were treated using the same protocols to determine potential systematic (cross) contamination. Metagenomic shotgun libraries were prepared from the samples and EBC’s according to Armbrecht et al. [24], [40], [41]. Library sequencing was conducted using an Illumina NextSeq platform (2 x 75 bp) at the Australian Cancer Research Foundation Cancer Genomics Facility & Centre for Cancer Biology, Adelaide, Australia, and the Garvan Institute of Medical Research, KCCG Sequencing Laboratory Kinghorn Centre for Clinical Genomics, Darlinghurst, Australia. Bioinformatic processing of shotgun sequencing data followed protocols detailed in Armbrecht et al. [40], with software and analytical parameters described in Armbrecht et al. [24], [42]. The National Center for Biotechnology Information (NCBI) nucleotide database (ftp://ftp.ncbi.nlm.nih.gov/blast/db/FASTA/nt.gz, downloaded November 2019) was used as the reference database to build a MEGAN alignment tool (MALT) index and sequences aligned using MALT (version 0.4.0; semi-global alignment) [43]. Subtractive filtering (i.e., subtracting reads for species identified in EBCs from samples) was conducted for the final data set. Hereafter, the term ‘samples’ refers to sediment-derived data after subtraction of EBC reads. We used MEGAN CE 6.21.10 to export the read counts for eukaryotes, including diatoms, at their lowest identified taxonomic level.