A Quaternary Sedimentary Ancient DNA (sed aDNA) Record of Fungal-Terrestrial Ecosystem Dynamics in a Tropical Biodiversity Hotspot (Lake Towuti, Sulawesi, Indonesia)

Abstract

Short-term observations suggest that environmental changes affect the diversity and composition of soil fungi, significantly influencing forest resilience, plant diversity, and soil processes. However, time-series experiments should be supplemented with geobiological archives to capture the long-term effects of environmental changes on fungi-soil-plant interactions, particularly in undersampled, floristically diverse tropical forests. We recently conducted trnL-P6 amplicon sequencing to generate a sedimentary ancient DNA (sedaDNA) record of the regional catchment vegetation of the tropical waterbody Lake Towuti (Sulawesi, Indonesia), spanning over one million years (Myr) of the lake's developmental history. In this study, we performed 18SV9 amplicon sequencing to create a parallel paleofungal record to (a) infer the composition, origins, and functional guilds of paleofungal community members and (b) determine the extent to which downcore changes in fungal community composition reflect the late Pleistocene evolution of the Lake Towuti catchment. We identified at least 52 members of Ascomycota (predominantly Dothiodeomycetes, Eurotiomycetes, and Leotiomycetes) and 12 members of Basidiomycota (primarily Agaricales and Polyporales). Spearman correlation analysis of the relative changes in fungal community composition, geochemical parameters, and paleovegetation assemblages revealed that the overwhelming majority consisted of soil organic matter and wood-decaying saprobes, except for a necrotrophic phytopathogenic association between Mycosphaerellaceae (Cadophora) and wetland herbs (Alocasia) in more-than-1-Myr-old silts and peats deposited in a pre-lake landscape, dominated by small rivers, wetlands, and peat swamps. During the lacustrine stage, vegetation that used to grow on ultramafic catchment soils during extended periods of inferred drying showed associations with dark septate endophytes (Ploettnerulaceae and Didymellaceae) that can produce large quantities of siderophores to solubilize mineral-bound ferrous iron, releasing bioavailable ferrous iron needed for several processes in plants, including photosynthesis. Our study showed that sedaDNA metabarcoding paired with the analysis of geochemical parameters yielded plausible insights into fungal-plant-soil interactions, and inferred changes in the paleohydrology and catchment evolution of tropical Lake Towuti, spanning more than one Myr of deposition.

Keywords: Lake Towuti; felsic; fungi; quaternary; sedaDNA; tropics; ultramafic.

Figure 1

General overview of the sampling location modified from [43]. (A) Lake Towuti is situated on the Indonesian Island of Sulawesi and forms part of the Malili Lake system. (B) Bathymetry of Lake Mahalona and Lake Towuti, displaying the coring location of Site 1. The coring location (Site 1) is indicated by a black circle with white filling in panel (B).

Figure 2

General overview of core 1F and the recovered sequence data. (A) Total extracted sedimentary DNA content (nanograms per gram of sediment). (B) Total number of reads, comprising (C) identified lab contaminants found in amplified and sequenced extractions and background blanks, either exclusively or in conjunction; (D) Non-fungal eukaryotes; (E) Unclassified reads; (F) Paleofungal reads utilized for downstream analysis in this study. All panels present quantities on a Log10 scale. The lithology graph on the left displays the meter composite depth (mcd), coring method (hydraulic piston vs Alien), the positions of tephra deposits T1-T23, and sediment ages. Radiocarbon dating of the bulk OM at 9.79 mcd revealed a sediment age of ~44.7 Ka, whereas ⁴⁰Ar/³⁹Ar dating of the tephra T18 layer at 72.95 mcd indicated a sediment age of 797.3 ± 1.6 Ka. The U2/U1c transition at ~98.8 mcd is estimated to have occurred 1 Ma ago through extrapolation (see the work of [40], for details). Also presented is a stratigraphic column illustrating previous endmember (EM) modelling of X-Ray Fluorescence (XRF) core-scanning data, modified based on the work of [44]). The column was generated by coloring each XRF data point according to its highest-scoring EM. The resulting six EMs (see the legend above the figure) represent changes in ecological (EM1), climatic (EM2-4), tectonic (EM5), and geomorphic processes that determine changes in sediment composition. Refer to [44], and the main text for further details.

Figure 3

A bubble plot illustrating the relative read abundance of Ascomycota (n = 52 ASVs) and Basidiomycota (n = 12 ASVs) across the 117 m long sediment record of Lake Towuti (n = 80 intervals). Lower taxonomic ranks include class (c_), order (o_), family f_), and genus g_). The legend details the size and color coding of the bubbles to represent the variation in relative abundance (1, 10, 20, 50, and 100%) against depositional units (stages) and their transitions. The meter composite depths (mcd) of each sample are color-coded according to sediment lithology: lacustrine red sideritic clays (RCs), green clays (GCs), diatom ooze (DO), and pre-lake U2 silts and peats. The red sideritic clays between 4.12 and 7.95 mcd span the Last Glacial Maximum; LGM [36]. Maximum likelihood transfer bootstrap trees for the taxonomic affiliation of each ASV, along with their closest described relatives or environmental isolates and clones from NCBI’s nr/nt database, are included in the Supplementary Information (Figures S3–S5).

Figure 4

Similarity percentage (SIMPER) results illustrate the top 11 fungal taxa and their percentage contributions to the Bray–Curtis similarities (cut-off = 90%) observed between samples representing the main sediment lithologies: sideritic red clays (RCs), green clays (GCs), and diatom ooze (DO) of Unit 1, alongside felsic silts and peats of Unit 2. Taxonomic ranks are abbreviated as follows: phylum Ascomycota (A) and phylum Basidiomycota (B), followed by c_ (class), sc_ (subclass), o_ (order), f_ (family), and g_ (genus). Refer to Figure 3 for an overview of the downcore distribution of each ASV and Figures S3–S5 for their taxonomic affiliations with the nearest-described relatives or environmental isolates and clones from NCBI’s nr/nt database.

Figure 5

A heatmap illustrating Spearman Rank Correlations (rho values are indicated in the color key) between fungal ASVs (normalized and square-root-transformed read abundances) and relative changes in available geochemical proxy data [40,56] within the upper 94 mcd: Horizontal Cluster Y1 consists of fungal ASVs that demonstrated significant positive Spearman correlations, primarily with %Si and the TLE/TOC ratio, suggesting enhanced nutrient availability/diatom primary productivity (PP) and increased preservation of sedimentary OM, respectively. Collectively, these parameters align with the paleobiology end member 1 (EM-1) as per the work of [44]. Many taxa in this cluster exhibited significantly negative Spearman correlations with ultramafic parameters and non-significant but positive Spearman correlations with the AL/Mg ratio, indicating an increased contribution of the Loeha River draining organic-rich felsic substrates into Lake Towuti during periods of heightened precipitation. Conversely, Cluster Y3 includes taxa that primarily correlated positively with ultramafic parameters, indicative of increased sediment discharge from the Mahalona River during inferred drying periods, most notably with %Ni and %Cr (EM2 markers for increased bedrock erosion) and %Fe and %Mn (EM4 redox markers implying water column oxygenation). Cluster Y2 comprises fungal taxa exhibiting significant positive Spearman correlations with ultramafic substrates and, to a lesser extent, inorganic felsic substrates (EM3; %Ti, %K), which drained into Lake Towuti from the Loeha River during periods of intensified precipitation. Cluster Y4 displays the opposite trend and consists of taxa that show significant positive Spearman correlations with inorganic felsic parameters and, to a lesser extent, ultramafic substrates. Both Clusters Y2 and Y4 include fungal ASVs predominantly linked with the drainage of U1c sediments during ongoing tectonic activities that resulted in high transport energy (EM5, %Ca) and frequent oscillations between the drainage of ultramafic versus felsic sediments. Taxa in Cluster Y5 primarily drained into Lake Towuti following the connectivity of the Lampenisu and Mahalona Rivers between Lakes Mahalona and Towuti (i.e., top 30 mcd), as evidenced by the (significant) positive and negative Spearman correlations with %Mg (i.e., EM6) and negative correlations with the Al/Mg ratio. Significance levels (p): 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05. This dataset was also interpreted using canonical analysis of principal coordinates (CAP), illustrating the spatial separation of samples in the three lacustrine subunits based on dissimilarities in the fungal ASV compositions (Figure S6). Taxonomic ranks are abbreviated as follows: phylum Ascomycota (A) and phylum Basidiomycota (B), followed by c_ (class), o_ (order), f_ (family), and g_ (genus).

Figure 6

A heatmap illustrating Spearman Rank Correlations (with Rho-values indicated in the color key) between downcore variations in normalized and square-root-transformed 18S read abundances of fungal ASVs and catchment paleovegetation inferred from paired sedimentary trnL-P6 profiling, as shown by [36]. Arbitrary boxes highlight the transitions between the main horizontal clusters (fungal communities) and vertical clusters (plant communities). Taxonomic ranks are abbreviated as follows: phylum Ascomycota (A) and phylum Basidiomycota (B), followed by c_ (class), o_ (order), f_ (family), and g_ (genus). For simplicity and to save space “mycetes” has been removed from the class names. Plant taxonomic ranks consist of order (o_), family (f_), clade (cl_), subfamily (sf_), tribe (tr_), and genus (g_). Letters in brackets signify the vegetation categories to which the identified plant members belong: [a] herbs; [B] TRSH; [c] warm-climate steppe C₄ grasses; and [d] wetland/humid-climate C₃ grasses. Significant Spearman Rank Correlations are indicated with asterisks: significance levels (p): 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05.