Extremotolerant fungi in resinous soils: a unique diversity of generalists and specialized hydrocarbon degraders

Abstract

Filamentous fungi are ubiquitous and constitute more than 75% of the soil biomass. Fungal diversity increasingly emerges as a key factor in soil ecosystem resilience against climate change and pollution, yet much of this diversity remains hidden and potentially in decline. While investigations have primarily focused on fungal tolerance to extreme physical conditions, such as temperature and salinity, chemically stressed environments remain underexplored reservoirs of novel fungal diversity. These habitats may harbor strains with significant biotechnological potential. This study tested the hypothesis that long-term contamination of soil with Pinus resin alters fungal diversity and promotes the growth of specialized fungal lineages enriched in hydrocarbon-degrading capabilities. We analyzed a resinous soil sample collected from an inactive resin processing site undisturbed for nearly 50 years. Initial physicochemical and microscopy analyses confirmed the presence of viable fungi despite extreme environmental constraints. High-throughput sequencing of fungal ITS2 regions revealed a fungal community composition highly distinct from adjacent forest soil, characterized by unusual taxonomic profiles and a high proportion of poorly classified or novel lineages. Functional inference and taxonomic analyses identified hydrocarbon-associated taxa including Sorocybe resinae (one of the most abundant OTUs) and Amorphotheca resinae (detected at low abundance). These fungi are known resinicolous and extremophilic species, illustrating the unique ecological adaptation of fungi within resin-rich, chemically stressful soils.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-026-05128-y.

Background

Fungi are ubiquitous in topsoil, displaying cosmopolitan distributions but also patterns of habitat filtering that reflect their exceptional adaptability [1, 2]. Global fungal species richness is estimated between 2.2 and 3.8 million [3], yet to date only ca. 160 000 fungal species have been formally described [4], including hundreds that can influence human health, both as pathogens or drivers of environmental change [5]. High-throughput DNA-based surveys have revolutionized our understanding of fungal biogeography and ecological roles, establishing soils as among the most diverse biological habitats on Earth, though their spatial and temporal distributions remain insufficiently understood [1, 6, 7].

Fungi adaptability allows colonization across extremes, from the stratosphere, Antarctic glaciers to the human gut [8]. This ecological success is driven by unique physiological strategies, and both metabolic and morphologic plasticity [9, 10]. Within soil ecosystems, fungi span from specialists constrained to narrow, extreme niches (e.g. Amorphotheca spp. found in polluted habitats with petroleum derivatives) [11, 12] to broadly tolerant generalists (e.g. Aspergillus fumigatus that thrive across a wide array of habitats, from compost to croplands) [9, 13, 14]. Adaptation strategies are most apparent in soils with extreme conditions such as temperature, water availability, or nutrient stress: thermophiles like Thermomyces lanuginosus inhabit compost [13, 15], psychrophiles such as Thelebolus microsporus persist in frozen soils [15–17] and xerophiles, including Wallemia sebi occur in arid habitats [18, 19]. In these settings, selective pressures exclude all but a restricted set of stress-adapted colonizers [10, 18]. Anthropogenic environments often impose additional selective pressures, featuring extremes of pH, temperature, radiation, and chemical pollutants; conditions unfriendly to most microorganisms [18].

Resin-contaminated soils are particularly challenging for microbial colonization due to low nutrient availability, acidic pH, and high resin acids concentration [20–22]. Resin acids, including abietic (AA) acid and dehydroabietic (DHA) acid, are potent allelochemicals for many soil microbes and aquatic organisms (with LC50 values in the sub-mg/L range) [20, 23]. As a result, resin acts as a chemical filter in conifer bark and in contaminated soils, shaping microbial succession and community structure during resin acid breakdown events [20, 24]. Resin degradation mechanisms are better characterized in bacteria, usually involving dit gene clusters (e.g. Pseudomonas and Paraburkholderia), whereas in fungi the process is less well defined and is thought to begin with cytochrome P450 monooxygenases that oxidize hydrophobic resin compounds, thereby facilitating their subsequent breakdown [24–26]. Examples of fungal species capable of degrading resin acids include Phlebiopsis gigantea and Trametes versicolor [27]. Fungi inhabiting resin-rich environments plausibly possess traits conferring tolerance and/or degradative capabilities toward these compounds, positioning them as models for studying extremophilic adaptation. However, little is known about how long-term, high-dose resin exposure in industrial environments alters soil mycobiome composition and function, or how these chemically extreme soils compare to other soils. Filling this gap is essential for advancing soil eco-health, biogeochemical management, and strategies for ecosystem recovery or bioremediation.

In this study, we use high-throughput, culture-independent (metabarcoding) community profiling to characterize the mycobiome of resin-impacted soil from an inactive industrial site. We compare its taxonomic and predicted functional composition to that of nearby soils (also under anthropogenic pressure) and integrate these findings with global databases of fungal diversity. Our aim is to better classify the ecological distinctiveness of resinous soils within a broader fungal biodiversity framework.

Materials and methods

Study design

We investigated whether long-term exposure to Pinus resin fosters specialization within the colonizing fungal communities, leading to enrichment of hydrocarbon-degrading fungi. Resinous soil (RS) was collected inside a resin refinery in Portugal, inactive since the 1970s. A nearby forest soil (FS) sample was collected for comparison. Both RS and FS were analyzed as representative composite samples to reduce small-scale spatial heterogeneity. Scanning electronic microscopy (SEM) confirmed the presence of numerous fungal structures within RS, which when placed onto culture media (with chloramphenicol) showed growing hyphae. The soil mycobiota was analyzed via high-throughput sequencing of environmental DNA. For mycobiome profiling, Operational Taxonomic Units (OTUs) were computed and identified with public sequence databases. Bioinformatic and statistical analyses enable resolution of fungal sub-communities, including resinicolous, extremophiles, and generalists, supporting inferences about ecological specialization. Complementary physiochemical characterizations of the RS included analysis of resin acid content by nuclear magnetic resonance (NMR), alongside assessment of their abiotic and biotic degradation.

Chemicals

Malt Extract Agar (MEA), mycological peptone and chloramphenicol (Oxoid, SR0078E) were obtained from Oxoid (Hants, UK). Other culture media components were acquired from Fischer Chemical (Waltham, MA, USA) or Sigma-Aldrich (Darmstadt, Germany). All solvents used were analytical grade or higher: methanol (p.a., > 99.9%, Fisher Chemical), acetone (p.a., > 99.8%, Fisher Chemical), phosphoric acid (p.a., 85%, José M. Gomes Santos), benzene (p.a., > 99.7%, Sigma-Aldrich), and deuterated dimethyl sulfoxide (DMSO-d₆; p.a., > 99.99%, Merck). Resin acid standards: abietic acid (AA, p.a., ≥ 85%), dehydroabietic acid (DHA, p.a., > 95%), and isopimaric acid (IPA, p.a., ≥ 98%) purchased from Acros Organics (Verona, Italy), Fluorochem (Hadfield, UK), and Sigma-Aldrich, respectively. Water was obtained from a Milli-Q system (Millipore).

Soil sampling and processing

Soil samples were collected in August 2018 at the inactive resin refinery in Oleiros, Portugal (39° 56’34.6 “N, 7° 55’44.3” W). The study area comprised two distinct sampling locations within the same industrial facility, embedded in a forest landscape (Fig. 1A; Fig. S1). A resinous soil (RS) sample was collected from inside the former processing pavilion (enclosed structure). A forest soil (FS) was taken at the edge of the industrial yard (resin receiving area), within a stand dominated by Pinus species. This served as a local reference sample collected under comparable environmental conditions but without prolonged resin exposure. RS consisted of a compact block of resin infused soil (~ 2m x1m x 0.005 m; 4.90 ± 0.42 kg), without any distinguishable litter layer or vegetation cover. Due to its compact nature, the entire block was manually fragmented on site using a steel crowbar and divided into four segments (Fig. S1D). FS samples were collected using a transect sampling method, consisting of four subsamples (Fig S1C) spaced ~ 15 m apart. After litter removal, soil was collected from a depth of ≤ 10 cm, totaling 0.72 ± 0.03 kg. All samples were transported in sterile plastic bags under cooled conditions and stored at -20 °C in dark until processing. The four frozen RS segments were cut into ~ 2 cm² pieces using an electric wood saw (Fig. S1D). From each segment, 25–35 pieces were homogenized by cryomilling at − 196 °C (RESTCH cryomill with a 5 mL jar, two 5-mm aluminum balls) employing a 3-min precool, 1-min milling at 30 Hz, and 0.5-min of cooling at 5 Hz. FS samples (four in total) were brought to room temperature, then sieved to particles ≤ 2 mm (ISO 11464:2006). Composite RS and FS samples were prepared by pooling 25 g of homogenized material from each segment and sample, respectively. The pooled composites were sealed and stored at 4 °C, dark until analysis. These composite samples were used in all subsequent experiments.

Fig. 1.

Origin, composition, and chemical characterization of resinous soil samples. Resinous soil samples were collected in Oleiros, Portugal (A), specifically within an inactive resin refinery at the resin processing pavilion (B). The sample, obtained in the form of a compact block (C), preserve a diverse array of typical resin acids, as demonstrated by a representative HSQC NMR spectrum of the RS (red) organic extract; a representative spectrum of FS organic extract (blue) is shown for comparison (D). Additionally, the chemical structures of resin acids commonly found in Pinus pinaster resin are illustrated for reference (E). The diagnostic H resonances of the resin acids were only detected in RS; their average concentrations (mg·g^− 1) in the RS organic extracts (n = 3) are shown in the table above. A schematic of the resin receiving area and the resin processing pavilion is shown in the Supplementary information Fig. S1. Maps and satellite images (A and B) were prepared using QGIS v3.34.7 (QGIS Development Team, 2024). Scale bars were added to panels (B) and (C). Panel (B) was calibrated in ImageJ (v.2.35) using a reference distance from Google Maps measurements, whereas panel (C) used direct measurements of the sample, with the lateral view scaled using an estimated thickness of approximately 5–6 cm

Characterization of soil samples

Physicochemical analyses of either soil were performed: pH (ISO 10390:2005, in distilled H₂O), gravimetric water content (θg; ISO 11465:1993), total nitrogen content (TKN; ISO 11261:1995, modified Kjeldhal Method), water activity (a_w) (HygroPalm HP23-AW-A probe (Rotronic AG, Bassersdorf, Switzerland)) [28], and organic carbon (OC; Walkley-Black wet oxidation method); organic matter (OM) was estimated as OC x 1.724 [29, 30], except that the OM for RS was determined by the European Norm 12879:2000, loss-on-ignition (dry) method.

Resin acids were extracted from the composite soil samples using a fast solvent-sonication [31]: soils were mixed with methanol (1:2 w/v), sonicated thrice (45 Hz, 30 °C, 30 min), centrifuged (4500 g, 15 min), and the pooled methanolic fractions (two extractions) dried under nitrogen flow, and stored at -20 °C until NMR analysis (n = 3). NMR analysis on a Bruker Avance III 800 CRYO (Germany) at 25 °C, employed 1D and 2D experiments (¹H, ¹H-¹H COSY, ¹-¹³ C HSQC and ¹-¹³ C HMBC). The methanolic soil extracts (5-mm-diameter NMR tubes) were dissolved in DMSO-d₆ (500 µL) with benzene as internal standard (10 µL of benzene solution, 125 g·L^{− 1}) (n = 3). Signals for DHA (H-14 at 6.84 ppm), IPA (H-11 at 5.79 ppm) and Pimaric acid (PA, H-13 at 5.12 ppm) were integrated for quantification as previously defined [32]. MestReNova, Version 11.04–18,998 (Mestrelab Research) was used to process the raw data acquired.

RS pieces (2 cm²) and RS cryomilled were imaged via SEM at Instituto Técnico de Lisboa Electron Microscopy Lab (MicroLab). Samples were mounted on aluminium stubs with carbon tape, sputter coated with Au/Pd using a Quorum Technologies coater (model Q150T ES), and observed with a ThermoScientific Phenom ProXG6 SEM (model with a CsB6 filament) equipped with an energy dispersive spectroscopy (EDS) light element detector.

Decay of abietic acid (AA) in synthetic soil

The decay of AA (0.3 mgꞏml^{− 1}, pH = 7) was accessed in synthetic soil (sand, kaolinite, bentonite, CaCO₃, and humic acid) inoculated with a 6 h-old peptone water extract of the RS. Briefly, RS was dispersed in peptone water (1:10, w/v) supplemented with chloramphenicol (50 µgꞏmL^{− 1}) for 6 h (30 °C, 100 rpm, dark), then decanted to obtain the microbial extract [31, 33]. For each condition and time point, three independent microcosms were established. Abiotic controls were also performed using 50 µgꞏmL^{− 1} chloramphenicol in sterile MQ water. Samples were incubated for 7, 14 and 29 days (20 °C, dark), with time-zero controls included. At each endpoint, AA was recovered (fast solvent-sonication, n = 3), and the extracts stored at − 20 °C until chromatographic analysis. AA content in the synthetic soil extracts was determined by HPLC using an Alliance 2695 system (Waters Corporation, Milford, MA, USA) with a PDA detector (Bromma, Sweden) and Symmetry C18 column (4.6 mm × 250 mm, 5 μm), isocratic elution with 2 mM MeOH: H₃PO₄ (86:14), flow 1 mLꞏmin^{− 1}, at 30 °C [34]. AA (240 nm) was quantified using an external standard method with a commercial pure compound; detection and quantification limits were 0.0001 mgꞏmL^{− 1} and 0.4 mgꞏmL^{− 1}, respectively. Data acquisition was accomplished with Empower 2 software (Waters).

Soil mycobiome profiling

DNA was extracted from 0.5 g composite soil samples (three independent DNA extraction replicates; n = 3) using an optimized phenol/chloroform/CTAB methodology [35]. The ITS2 fungal rDNA region was amplified in two steps: initial PCR with standard primers, followed by a second PCR using the same primers carrying a barcode: gITS7 (5’-GTGARTCATCGARTCTTTG-3’) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3’). In the second amplification conditions were modified: denaturation at 95 °C, 25 cycles, annealing at 55 °C. PCR products were purified (Zymo DNA Clean & Concentrator-5) and quality-checked by gel electrophoresis and NanoDrop™ spectrometry (One Microvolume UV-Vis Spectrophotometer, Thermo Fisher Scientific Inc., USA). RS replicates were sequenced in two independent Illumina MiSeq runs, whereas FS samples (n = 2) were sequence once.

Sequence data processing and taxonomic assignment

Raw data was pre-processed using fastp v0.23.2, including adapter trimming and read-pair merging [36]. Sequence clustering and OTU inference followed SEED v2.1 pipeline, as applied before [35, 37] (UPARSE v8.1.1861 [38], ITSx v1.0.11 [39]). The most abundant sequences were retrieved as representative for each OTU and used for downstream analyses. Taxonomic assignments targeted the “Full UNITE+INSD dataset for Fungi” (v04.04.2024) [40, 41], and alternatively the NCBI fungal subset [42], with a default e-value threshold of 1e^{− 5} using either the megablast algorithm or blastn.

Bioinformatics and statistical analyses

Statistical analyses were conducted in R v4.4.1. Alpha diversity metrics: observed richness, Shannon, Simpson, inverse Simpson, evenness and Chao1, were computed using Vegan v2.6.10 [43] and otuSummary v0.1.2 [44]; rare OTUs were defined as ≤ 1% relative abundance. Statistical comparisons included Wilcoxon rank-sum test or t-tests depending on assumptions verified by Shapiro–Wilk and Levene’s tests. To assess effect sizes, Cliff’s delta was calculated when necessary. Beta diversity was assessed using Bray–Curtis dissimilarity, visualized via Principal Coordinates Analysis (PCoA), and tested using PERMANOVA; homogeneity of dispersion was verified by PERMDISP (ANOVA) [43]. Taxa visualization employed Metacoder v0.3.8 (heat trees) [45], ggVennDiagram [46], and ComplexHeatmap v2.25.0 [47]. RS and FS mycobiota (depleted from singletons) were compared with the Global Soil Mycobiome Consortium (GSMc) dataset [48] using phyloseq v1.48.0 [49] and the retrieved taxonomic composition was visualized using fantaxtic v0.2.1 [50]. Boxplots and additional statistical graphics were generated with ggplot2 v3.5.1 and associated tidyverse (v2.0.0) tools [51–53]. Final figures were prepared for publication in Inkscape v1.2.1 [54].

Functional characterization of the mycobiomes

The ecological traits of the combined RS and FS genus-level dataset (composed of 1079 OTUs, including singletons) was assigned using the FungalTraits v1.2 database [55]. Trait annotation focused the following trait categories: lifestyle (guild), growth form, substrate preference, decay type, and plant association. Genera lacking trait information (e.g. incertae sedis or unclassified taxa) were excluded from functional summaries and trait coverage estimates. Relative abundance (%) per trait level was calculated independently for each soil, based on aggregated genus-level abundances. Dominant genera were reported only when the RS–FS difference > 1.0% (Δ_pp) and their relative abundance reached ≥ 0.5% in at least one soil.

Predictive functional analysis was performed using FunFun v0.1.15 [56, 57] to infer metabolic potential from sequence data. Default parameters were applied, with the exception that the epsilon distance for the nearest neighbor was fixed at 0.6. The resulting functional predictions were weighted by OTU abundance, and subsequently normalized. Enrichment analysis for functional categories was then performed using R v4.4.1, DESeq2 v1.48.2, and dplyr v1.1.4 packages [58–60]. Enriched functional categories display log₂fold change > 1, and adjusted p-value < 0.05.

Results and discussion

Resinous soils reveal viable fungi despite selective physicochemical conditions

Fungi adapted to resinous environments are often proficient in degrading complex hydrocarbons [20, 61], but their diversity and functional potential remains largely underexplored. Our study system, a resin industry facility inactive for over fifty years (Fig. 1A-B), offers distinct niches: a resin receiving area (open field) and a resin processing pavilion (mostly enclosed, reduced erosion and disturbance) (Fig. S1A-B). Resinous soil (RS), sampled within the pavilion, was obtained as a compact block with the bottom surface in contact with cement flooring and high humidity (Fig. 1C). Forest soil (FS) was collected in the resin receiving area within a Pinus stand (Fig. S1C). For both sites, composite samples were produced (Fig. S1D). Physicochemical analysis (Fig. S2) revealed that RS (cryomilled) is moderately acidic (pH 5.9), extremely rich in OM (84%), and N-poor (0.05%), reflecting resin chemical imprint. FS values (7.41% OM, 0.15% N) were within the expected range reported for Portuguese P. pinaster forest soils (7–12%, > 5 cm deep) [62, 63], yet RS far exceeded typical OM values. This marked difference is likely due to the analytical approach used for OM estimation: in FS, it was determined via oxidizable organic carbon, whereas in RS, due to its compact nature, OM was analyzed by loss-of-ignition method (EN 12879), which quantifies all combustible organic material, including resin-derived compounds and any embedded organic matter. Gravimetric water content was higher in RS (95.38 gꞏKg^{− 1}) compared to.

FS (13.75 gꞏKg^{− 1}), due to moisture preserved in the resin-block’s cement interface. However, a_w, a critical determinant for microbial life, was 0.78 in RS, favoring xerotolerant taxa, and 0.9 in FS, generally permissive for most microorganisms. NMR analyses of the methanolic extracts of RS confirmed the dominance of DHA, PA, and IP (Fig. 1D-E). AA was not detected, despite its prevalence in local Pinus resins [32]. Extraction controls of a synthetic matrix (containing 0.6 mg AA per g of soil) indicated ≥ 80% recovery efficiency. No resin acids were detected in FS using NMR (Fig. 1D, blue).

SEM of untreated resinous blocks revealed abundant mycelial structures: hyphae, spores and/or fruiting bodies, distributed throughout the matrix (Fig. 2) and sometimes embedded within resin (Fig. 2F-G) or on Pinus sp. pollen grains (Fig. 2C). Calcofluor white-stained fungal mycelial expanded outward from RS particles (Fig. S3A-B), further confirmed by light microscopy (Fig. S3C-D). Moreover, synthetic soil spiked with AA (0.12 mg/g) and inoculated with the RS microbial extract (peptone water supplemented with chloramphenicol) exhibited clear biotic degradation of AA at 14 and 30 days (> 50% AA loss by day 30; Fig. S4). In the sterilized controls (abiotic controls), AA decay reached a plateau after 14 days, reflecting abiotic transformations (e.g. oxidation, isomerization) and/or interactions with the soil matrix (sorption to soil’s constituents). In contrast, inoculated microcosms showed continued depletion over time. These results indicate that viable microorganisms are not only present but are metabolically responsive in RS. Given that bacterial growth is largely suppressed by chloramphenicol supplementation, the observed AA degradation can be primarily attributable to fungal activity. These results are consistent with previous reports of resin acids degradation by e.g. Trametes versicolor, Phlebiopsis gigantea, and Aspergillus niger [20, 27, 64]. Collectively, these results confirm that RS retains colonizing fungi, regardless of high resin acid diversity and concentration (≥ 0.15 g per g of RS), and C-rich, N-poor and low a_w conditions. These conditions resemble those of hydrocarbon-polluted soils, which are often inhabited by hydrocarbon-degrading fungi and bacteria [65, 66].

Fig. 2.

Morphological characterization of resinous soil. Scanning electron microscopy (SEM) micrographs of the resinous soil matrix and associated fungal structures. A The surface morphology of the resinous soil sample under analysis, providing a general overview of the sample. At this scale, no clearly identifiable fungal structures are observed. Scale bar: 200 μm. B Filamentous structures resembling fungal hyphae. Scale bar: 100 μm. C A Pinus pollen grain (~ 10–15 μm in diameter) undergoing fungal parasitism. The pollen exine remains largely intact, but fungal hyphae are observed penetrating and surrounding the grain. Scale bar: 30 μm. D A dense, filamentous fungal network interspersed with small spherical structures (~ 1–3 μm) developing at the tips of the hyphae, suggestive of spores. Scale bar: 20 μm. E A smooth, porous matrix containing embedded unidentifiable material along with a free-floating fungal spore (~ 5–10 μm in diameter). The surrounding matrix may represent an area of fungal colonization or preservation. Scale bar: 20 μm. F A fragmented matrix entrapping elongated, oval structures (~ 3–5 μm long), which are morphologically consistent with spores from filamentous fungi. Scale bar: 20 μm. G A cavity-like structure (~ 7 μm in diameter) with multiple rounded elements (~ 3 μm in diameter) enclosed within the resin matrix, either a sporangium or a degraded fungal fruiting body containing spores. The spores appear loosely aggregated, suggesting a prior release event. Scale bar: 15 μm. H A filamentous fungal structure resembling a conidiophore. Scale bar: 15 μm. Yellow arrows indicate fungal hyphae and cyan arrows indicate different fungal spores. See also Fig. S3 for confocal and light microscopy on the viability of fungal material from resinous soil particles

Resinous soil mycobiome: richness and classification uncertainty

Aiming to validate unusual diversity and dominance by specialists in the RS mycobiome, two independent high-throughput ITS2 sequencing runs (RS1 and RS2) were conducted on the same DNA-extraction triplicates. Raw reads were 51 million and 69 million (Table S1), yielding post-filtering 53,779 and 65,188 high-quality reads, which were subsequently clustered (97% similarity) into 341 (Table S2) and 476 OTUs (Table S3), for RS1 and RS2, respectively. Alpha diversity metrics: observed richness (number of detected OTUs), Chao1 (total richness including undetected rare taxa), Shannon diversity (accounting for both richness and evenness with sensitivity to rare and abundant taxa), Simpson diversity indices (1–D and 1/D, emphasizing evenness and dominance structure), and evenness (the balance between relative abundances across taxa), confirmed normality and homogeneity (Shapiro–Wilk and Levene’s tests, p > 0.05) (Table S4). Most metrics showed no differences between RS1 and RS2 (Table S5, Fig. S5A, Table S6), except higher Shannon diversity (p ≈ 0.02) and Inverse Simpson (p ≈ 0.05) for RS2 (t-test). Analyses of abundant and rare fungal taxa (Table S7) likewise indicated stable dominant species across runs. RS2 appears marginally more even than RS1, but differences were minor (likely technical). β-diversity (Bray–Curtis, PERMANOVA) showed no significant compositional differences between runs (F = 0.83, R² = 0.17, p = 0.7, Fig. S5B) and similar within-group variance (F = 0.25, p = 0.65; Fig. S5C), with only 17% of explained variance attributable to dataset (Fig. S6A, B). Moreover, 63% of the 156 identified genera were shared between runs (Fig. S5D), with nearly all abundant taxa (> 0.5%) present in both, and few unique genera exceptionally rare: 16 in RS1 and 38 in RS2 (Table S8).

Consequently, the RS mycobiome (integrating both runs) comprises 620 fungal OTUs (Table S9). Alpha-diversity metrics showed high fungal richness and relatively stable community structure, but also remarkable taxonomic uncertainty (Table S6 and S7). The mean observed richness was 322 OTUs (IQR: 36), with the Chao1 estimator indicating a potentially higher richness (mean: 476, IQR: 93), suggestive of presence of undetected taxa. Shannon diversity (mean: 3.1, IQR: 0.07), Simpson index (mean: 0.9, IQR: 0.007), and evenness (mean: 104, IQR: 10) all displayed narrow interquartile ranges. The consistently high and stable Shannon, Simpson, and evenness values indicate a diverse community with relatively uniform taxon contributions, reflecting low dominance and a balanced composition structure; in these metrics, higher values indicate greater diversity and evenness (Table S6).

The RS mycobiome spans 6 phyla, 26 classes, 64 orders, 107 families, and 154 unique genera (Table S8), dominated by Ascomycota (65.5%) and Basidiomycota (26.2%). Minor phyla (Mortierellomycota, Chytridiomycota, and Mucoromycota) were present at negligible abundances (< 0.01%; Table S9, Fig. 3A). About 9.4% of OTUs could not be classified beyond Kingdom (“no hit” reads). Taxonomic uncertainty was also substantial at lower ranks: 3.9% of OTUs at phylum level, and up to 30.8% and 56.6% were assigned to incertae sedis at genus and species, respectively. Although ITS2-based assignments do not provide sufficient resolution for robust species-level identification, these results suggest that the resin’s selective environment may be shaping an uncharted “microbial dark matter” (Fig. 3B), consistent with UNITE classification of ‘dark taxa’ in ITS-based species hypotheses [40, 67]. In ITS-based classifications, poorly resolved OTUs, including sequences with no database match (“no hit”) or assigned as incertae sedis, are often considered part of the ‘dark taxa’, i.e. taxa lacking formal taxonomic resolution in UNITE species hypotheses [40, 67, 68]. This pattern is particularly evident when compared with FS (see below). Moreover, the structure of the community (Fig. S7) was highly uneven with a small group of dominant taxa: 19 classified genera plus unclassified reads, Eurotiomycetes, Fungi, Basidiomycota, and Dothideomycetes incertae sedis, accounted for 95.1% of abundance, with 131 genera contributing to the remaining 4.9%, each ≤ 1%. Key dominants included Eurotiomycetes gen. incertae sedis (26.2%), Sporobolomyces (12.3%), Sorocybe (7.1%), Penicillium (5.6%), along several other genera at intermediate abundances (0.5–5%; Table S9).

Fig. 3.

Taxonomic composition and structure of the resinous soil (RS) mycobiome. A Relative abundance of fungal phyla detected in the RS mycobiome, shown as mean relative abundance (%) across RS replicates. Ascomycota and Basidiomycota dominate the community, whereas other phyla (Mortierellomycota, Chytridiomycota, and Mucoromycota) are present only at trace levels. A noticeable fraction of reads could not be confidently assigned beyond higher taxonomic ranks, including taxa classified as incertae sedis and unclassified (“no hit”) sequences. B Heat tree visualization of the RS fungal community generated using the metacoder framework, illustrating taxonomic relationships from phylum to genus. Node color reflects mean relative abundance (%) across RS replicates, with darker colors indicating higher abundance, while node size corresponds to the number of OTUs assigned to each taxon. The presence of numerous nodes assigned only at high-level taxa (e.g. not resolved to genus) and incertae sedis groups reflects limited taxonomic resolution and highlights the prevalence of taxonomically unresolved fungal lineages in resinous soil

Many of the genera detected in RS mycobiome represent ecological specialists (Table S10), including resinicolous taxa, both as dominants (Sorocybe (best-hit S. resinae) Sarea, Lophium and Mytilinidion [69]) and at low-abundance (Chaenothecopsis [69]), and taxa associated with conifer hosts (Colacogloea, Symbiotaphrina, Infundichalara, Phialocephala, Pragmopora [70–73]). Penicillium and Aspergillus spp. build the dominant group of generalists, consistent with their frequent isolation from hydrocarbon-rich environments (e.g. A. tubingensis and A. niger, Penicillium chrysogenum [14, 74]) and ability to transform plant terpenes (e.g. P. digitatum [75] and A. niger [76]). In addition, RS is rich in multiple specialist taxa repeatedly reported from extreme and/or petroleum-rich environments:

Candida keroseneae (originally described from aviation fuel) [77], Debaryomyces hansenii, Exophiala spp. (black yeasts), Rhodotorula (including R. mucilaginosa) [18, 78] and Scedosporium [14, 79]. Within low-abundant taxa it is noted the presence of Amorphotheca resinae (= Hormoconis resinae), known to utilize fuel components [11, 80], Rhinocladiella, Cladosporium, Aureobasidium and Alternaria, frequently documented as hydrocarbon-tolerant/degraders [14], and Neodevriesia s.l., Devriesia, and related Neodevriesiaceae lineages, usually found to colonize nutrient-poor, xeric and chemically stressful surfaces [81]. Collectively, the RS mycobiome displays both high taxonomic diversity and uncertainty, ecological uniqueness, and potential novelty, highlighting strong selection for specialized, stress-adapted, and potentially metabolically versatile fungal lineages.

High taxonomic turnover and distinctiveness between resinous and nearby forest soils

We compared the RS mycobiome to an adjacent FS mycobiome, testing if resin impacts extend to local patterns (Table 1; Table S11). FS showed slightly higher richness (352 OTUs vs. 281 OTUs in RS) but nearly identical Shannon diversity values (3.03 in RS, 3.02 in FS), and marginally lower Simpson’s indices (0.90 in RS; 0.87 in FS), indicating more pronounced dominance structure in FS (Table S12). Despite these observable differences, Wilcoxon rank-sum tests yield no statistical p-values (> 0.05), likely constrained by small sample sizes. Nonetheless, non-parametric effect size (estimated via Cliff’s delta) indicated large biological relevant effect for richness (δ = 1.00) and Simpson’s evenness (1–D; δ = − 0.67), whereas Shannon diversity effect size was negligible (δ = 0.00).

Table 1.

Summary of key features of resinous soil (RS, integrated dataset) and forest soil (FS) mycobiomes. Alpha diversity metrics include richness, Shannon and Simpson diversity. Compositional differences are highlight by Bray Curtis dissimilarity coefficient, genus-level overlap and level of unclassified/incertae sedis genera. Functional differences highlight each microbiome functional profile. Additional information is displayed in the Supplementary Information (Table S11-S17)

Parameter	RS	FS	RS vs. FS
Observed richness (OTUs)	281	352	Slightly higher in FS
Shannon diversity	3.03	3.02	Similar
Simpsons diversity	0.90	0.83	Slightly higher in RS
Dominant taxa	Eurotiomycetes incertae sedis, Sporobolomyces, Sorocybe, Penicillium	Penicillium, Sordariales incertae sedis, Phialocephala	Different dominant assemblages
Bray Curtis dissimilarity	—	—	0.968
Genus-level overlap	—	—	~ 17% shared genera
Unclassified / incertae sedis OTUs	30.8%	9.3%	Increased taxonomic uncertainty in RS
Functional profile	Saprotrophic; Antagonistic; Resin-associated traits	Plant-associated, including foliar endophytes; Soil-associated traits	Functionally distinct

βeta-diversity analyses (Whittaker’s β, β_W; Bray-Curtis dissimilarity, D_BC) quantified this contrast: richness-based turnover (i.e., compositional dissimilarity among samples) was higher among RS (1.95) than FS (1.41) replicates, reflecting greater internal heterogeneity in RS, while abundance-weighted dissimilarity between environments was extreme (D_BC = 0.968; Table S13). Consistently, genus-level composition mirrored this separation. FS was dominated by Penicillium (54.2%), Sordariales gen. incertae sedis (6.4%), and Phialocephala (5.6%) (all common soil genera), while RS had Eurotiomycetes gen. incertae sedis (26.4%), Sporobolomyces (12.1%), and Sorocybe (7.7%) as dominant (the last two exclusives to RS), sustaining habitat-specific selection (Figs. 4 and 5A). Together, dominant genera comprised 94.4% and 91.9% of the RS and FS communities, respectively. Taxonomic overlap of classified genera was limited: of 359 genera, only 60 (17%) were shared between soils, whereas 204 (57%) were unique to FS, and 95 (26%) to RS, with several clades strongly enriched or depleted in either soil (Fig. 5B). Each community is composed of a few dominant taxa plus a large pool of low-abundance taxa, supporting both richness/high turnover and distinct ecological structuring.

Fig. 4.

Genus-level composition of the fungal communities in resinous soil (RS) and adjacent forest soil (FS). Genus-level composition of the RS and FS mycobiomes, shown as mean relative abundance (%) across replicates for each condition. Genera with relative abundance exceeding 1% in at least one condition (RS or FS) are shown. In both soils, community structure is highly uneven and dominated by a small number of taxa, and the identity of dominant genera differs strongly between habitats

Fig. 5.

Local comparison of fungal communities in resinous soil (RS) and adjacent forest soil (FS). A Heat tree representation highlighting taxonomic differences between RS and FS at the genus level. Node color indicates the log₂ fold change in relative abundance between soils, with enrichment toward RS or FS shown by negative (blue) or positive (green) values, respectively, while node size reflects the number of OTUs associated with each taxon. The distribution of colored branches illustrates clades that are consistently favored or suppressed in each soil type. B Overlap of classified fungal genera between RS and FS. Only a small subset of genera is shared between the two soils, whereas most taxa are unique to one environment, underscoring strong habitat-specific filtering and limited taxonomic overlap despite close spatial proximity

To assess taxonomic resolution independently of the combined clustering approach, classification uncertainty was evaluated using the individually processed RS and FS datasets. FS showed lower uncertainty at higher taxonomic ranks, with fewer OTUs unclassified beyond kingdom level (3.17% vs. 9.4%) and fewer incertae sedis assigned at phylum (1.12% vs. 3.9%) and genus levels (9.29% vs. 30.8%) than RS (Table 1). Because ITS2 does not reliably resolve species in either dataset, species-level assignments remained poorly resolved, with similarly high proportions of unresolved taxa (62.26% in FS vs. 56.6% in RS). Together, these results indicate that RS contains a larger fraction of poorly classified taxa at lower ranks down to genus.

The RS mycobiome compositions retrieved from either the clustering of RS reads alone, or RS and FS reads combined, are virtually identical, except that in the last, five new OTUs could be assigned in RS and the number of unclassified OTUs slightly decreased (Table S11). This reflects how UPARSE defines centroids when clustering pooled data and do not represent biological differences [38]. When running the taxonomic assignment against NCBI nt database (BLASTn; e-value 0.05) instead of UNITE, the number of unclassified OTUs also reduced slightly (Table S14) [40–42]. Systematically unclassified OTUs (e.g. OTU20) likely index true novelty.

Overall, RS and FS, though similar in gross α-diversity, diverge strongly in community structure and in the prevalence of poorly classified or novel fungal lineages. Similar patterns have been reported in other chronically polluted substrates, indicating that richness can coexist with relatively even abundance structures due to the replacement of sensitive dominants by stress-tolerant specialists [79, 82]. Minimal genus-level overlap, divergent taxonomic profiles and higher proportions of unclassified and incertae sedis OTUs in RS, support habitat-driven differentiation shaped by selective pressures in resin-rich soils.

Functional characterization of the resinous soil mycobiome

Trait profiling at the genus level using FungalTraits (Table S15-S17) and FunFun (Table S18) databases [55–57] revealed major functional differences between RS and FS mycobiomes. Substrate preference (Table S16 and S17) showed the highest Bray–Curtis dissimilarity between mycobiomes (≈ 55), followed by growth form (≈ 53), lifestyle composition (≈ 51), and plant association (≈ 31), while decay type showed lower divergence (≈ 26). Annotation completeness varied across traits, with RS harboring a substantially higher proportion of poorly characterized genera compared to FS. The RS mycobiome exhibited minimal plant-association coverage (17.4%) with sharp reduction of host-linked interactions, mostly putative foliar endophytes (8.3%; Penicillium, Rhodotorula, Aspergillus) and others lacking endophytic capacity (5.4%, dominated by Colacogloea). RS lifestyle composition was enriched in saprotrophic guilds (mostly Sporobolomyces, 12.3%, Colacogloea, 4.6%, and Sorocybe, 7.1%) and antagonistic guilds (mostly mycoparasites, 14.7%), along with substrate preferences for fungal-derived substrates (17.2%), sugar-rich resources (8.6%) and resin-associated substrates (7.1%, exclusively detected in RS). In contrast, FS was strongly dominated by soil-associated substrates (59.0% vs. 6.4% in RS) and plant-associated guilds, with foliar endophytes comprising nearly 28% of total abundance (mostly Penicillium), besides ectomycorrhizal taxa (4.3%), soil saprotrophs (9.0%), and root endophytes (3.8%). These patterns indicate that RS mycobiome is dominated by saprotrophic taxa and exhibits specialization on resin-derived substrates and fungal resources over plant-linked symbiotic roles.

Predictive functional pathway analyses revealed substantial divergence between the two mycobiomes (Table S18). The RS mycobiome showed significant enrichment in surface adherence and biofilm formation pathways, including extracellular matrix (ECM) interaction and synthesis: cell adhesion molecules (ko04514), ECM-receptor interaction (ko04512), glycosphingolipids biosynthesis (ko00601), and glycosaminoglycans metabolism (ko00532, ko00534, ko00535) (Fig. 6A). These traits are consistent with RS conditions, promoting biofilm-like growth and spatial stability on hydrophobic, low-moisture substrates [11]. By contrast, FS communities exhibited enrichment in functional categories associated with: secondary metabolites biosynthesis (ko00902, ko00997, ko01059), peptidoglycan degradation proteins (ko01011), and metabolism of amino acid precursors (ko00471, ko00473), involving L/D-alanine and D-glutamic acid (Fig. 6A). This includes functions associated with bioactive compound production (e.g. polyketides and non-ribosomal peptides) and microbial cell wall degradation, which are consistent with niche competition in nutrient-variable forest soils [83]. FS also retained brown-rot, chitinolytic, and keratinolytic decay profiles (all absent in RS; Table S17), reflecting broader degradative versatility in forest litter. Overall, these patterns reflect substrate chemistry rather than OM quantity: RS contains abundant but chemically constrained, resin-derived carbon, whereas FS OM is mostly litter-derived and therefore more accessible.

Fig. 6.

Functional differentiation of resinous soil (RS) and forest soil (FS) mycobiomes and their placement within European fungal diversity. A Predictive functional pathway analysis inferred using FunFun, showing differentially enriched KEGG pathways between RS and FS mycobiomes (log₂ fold change > 1 or < − 1; adjusted P < 0.05). Positive log₂ fold change indicates enrichment in RS, whereas negative values indicate enrichment in FS. Pathways related to surface adherence, extracellular matrix interaction, and biofilm-associated processes are enriched in RS, whereas FS is enriched in pathways associated with secondary metabolite biosynthesis, microbial competition, and amino acid metabolism. Point color indicates the direction and magnitude of enrichment (log₂ fold change), and point size reflects statistical significance (− log₁₀ Pₐdⱼ). B Shannon diversity indices of fungal communities across biome categories defined in the Global Soil Mycobiome consortium (GSMc), including anthropogenic, forest, grassland, shrubland, tundra, woodland, and unknown biomes. Boxplots show the distribution of diversity values within each biome. RS (green dot) and FS (yellow dot) fall at the lower boundary of their respective biome categories, indicating reduced diversity relative to regional reference soils. C Relative abundances of the 25 most abundant fungal orders across selected European sampling sites, including RS and FS both placed in Portugal. Bars are grouped by country and colored by phylum and order. RS is dominated by Ascomycota and show reduced representation or absence of fungal orders commonly abundant in regional soils, including Mortierellomycota, Zoopagomycota, and Mucoromycota, highlighting RS compositional divergence from geographically and climatically similar European soils. Only 30% of RS diversity can be explained by the 25 most abundant fungal orders observed in these soils, whereas ~ 80% of FS diversity is explained by these orders

Global context: biogeography of the resinous soil mycobiome

To contextualize the ecological uniqueness of RS, its mycobiome (alongside that of the neighboring FS) was compared to reference datasets from the Global Soil Mycobiome consortium (GSMc) [48]. RS and FS were placed within the spectrum of terrestrial biomes across European and Mediterranean regions, classified as anthropogenic and forest biomes, respectively. Comparison of Shannon diversity values across GSMc biome categories revealed similar medians, ranging from 5.48 in woodland to 5.26 in forest/shrubland (Fig. 6B; Table S19). Considering only RS and FS classified OTUs that greatly reduces their dissimilarity, both exhibited much lower Shannon values (3.06 and 3.11, respectively), placing them at the lower boundary of their biome categories and among the least diverse soils in the analyzed regions with values ranging from 5.26 (subtropical broadleaf forest in Madeira, Portugal) to 6.19 (forest in Sicily, Italy). This result underscores the local divergence of RS and FS from regional diversity patterns, highlighting how chronic chemical stress and anthropogenic history can produce island-like fungal assemblages even within similar climatic zones [84, 85]. Principal Coordinates Analysis (PCoA, based on Bray–Curtis dissimilarity) clustered RS and FS within the Southern European cloud, suggesting partial compositional similarity to these soils, although the global dataset showed no clear geographic clustering and no strong affiliation with any single regional or biome cluster (Fig. S8A-B). RS and FS clustering based on classified OTUs suggests that shared site-related factors (e.g. geography and climate) contribute to similar diversity patterns, outweighing the influence of their dissimilar components (Fig. 5). This effect is amplified when considering a large dataset at the continental scale. Analysis of the 25 most abundant fungal orders across European samples further illustrated RS distinctiveness (Fig. 6C). RS is compositionally divergent from the other Portuguese and Spanish soils, including FS: only 30% of RS diversity can be explained by the 25 most abundant fungal orders observed in these soils. In contrast, nearly 80% of FS diversity is explained by these orders, indicating a much closer compositional overlap with the dominant GSMc taxa. Most reference Portuguese soils were dominated by Basidiomycota, especially Agaricales [12, 86]. Contrarily, both RS and FS are dominated by Ascomycota (RS: Chaetothyriales, Saccharomycetales, Eurotiales; FS: Eurotiales, Helotiales, Sordariales), besides that fungal orders usually common in regional soils (Mortierellomycota, Zoopagomycota, and Mucoromycota) are either at low abundance or absent. Collectively, despite that the analysis is narrowed to classified OTUs, data suggest that the RS mycobiome (and to a much lesser extent also that of FS) is compositionally distinct compared to geographically or climatically similar soils.

Ecological context of RS mycobiome composition

Although bacteria are well-established agents of diterpenoid metabolism [24], our findings underscore the significant, yet underappreciated, role of fungi in resin degradation under extreme conditions. Whereas fungi typically do not dominate in transient, resin-amended environments [24], the RS mycobiome supports their persistence and potential functional relevance. Microcosm experiments suggested that the RS mycobiota is actively involved in resin-acid turnover: under antibiotic suppression of bacteria, > 50% of AA was lost in 30 days, confirming viable fungal terpenoid degradation capacity (Fig. S4). This result aligns with established fungal metabolic pathways, particularly those involving cytochrome P450 monooxygenases (CYP) and related oxidative enzymes known to transform hydrocarbons and terpenoids [87]. The taxonomic composition of RS supports this interpretation, being dominated by Eurotiomycetes, mainly Penicillium with Aspergillus and Exophiala at lower abundance. Members of this class are well recognized for oxidative xenobiotic metabolism via CYP systems. Additionally, the presence of Sorocybe, a genus associated with resin-rich environments, suggest that the observed community reflects adaptation to and selection for fungi capable of degrading resin-associated compounds [25, 69]. Comparative analysis with mycobiomes from resin-rich conifer tissues (pollen, needles, bark, galls) [24, 88–90] reveals that typical core classes in those environments: Dothideomycetes, Tremellomycetes, Mortierellomycota, and Mucoromycota, are only detected in RS at vestigial levels (e.g. Dothideomycetes 3.9%, Tremellomycetes 0.3%). In sharp contrast, Eurotiomycetes are strongly enriched (Fig. 5, Table S9), consistent with previous reports on spruce-derived resin enrichments, and resinicolous surveys [24, 90]. Experimental enrichment studies confirmed that resin compounds act as long-term selective agents, rapidly filtering fungal and bacterial populations [24]. Early-stage Ascomycota (e.g. A. terreus, A. flavus, P. decumbens) are only transient community members and are found in RS solely as rare traces (≤ 0.03%). With the persistence of resin, only extremotolerant Eurotiomycetes and Microbotryomycetes (red yeasts) [12], remain abundant, forming stable, stress-adapted consortia distinct from opportunistic colonizers.

Previous studies have reported as dominant genera from pollen and needles: Epicoccum, Vishniacozyma, Alternaria, Aureobasidium, and Cladosporium [88, 89]. These genera occur in RS at very low abundance, suggesting limited recruitment from above-ground conifer sources, and further supporting the uniqueness of chemical filtering in RS. Satellite taxa reported in association with P. pinaster needles, such as Mortierella clonocystis, Cryptococcus podzolicus, Aspergillus piperis, and Metschnikowia sp [88]., are largely absent in RS (< 0.01%, except Aspergillus at 0.65–1.06%). This pattern suggest that the RS assemblage arises through soil-chemical selection, consistent with soils under pine plantations favoring Ascomycota (especially Trichocomaceae: Aspergillus, Talaromyces, Penicillium) over Basidiomycota prevalent in undisturbed forest soils [86].

Conclusions

The described resinous soil mycobiome is radically distinct from a nearby forest soil and microbial communities associated with natural conifer tissues, indicating that chronic resin saturation from long-term anthropogenic activity acts as an intense ecological filter. RS is dominated by resin acids (≤ 15% of soil mass), recalcitrant markers of conifer resin [32], representing an anthropogenically amplified natural chemical stress. Combined with low water activity and extreme C: N imbalance (C/N > 1600), RS creates an extreme microhabitat filtering for functionally specialized fungal taxa. It parallels chronic heavily oil-polluted soils where hydrocarbon-degrading and/or stress-tolerant fungi dominate [14, 74]. These convergent selective pressures underscore the broader relevance of RS as a model for fungal adaptation to persistent chemical stress.

Our findings elevate the significant yet frequently overlooked functional contribution of fungi under persistent resin pressure. Microcosm experiments confirmed fungal viability and active resin acid turnover, though the specific taxa driving these modifications remain unidentified. The RS mycobiome is defined by extremotolerant Eurotiomycetes (42.7%) and resilient yeasts from Microbotryomycetes, distinguishing it from conifer needle, pollen, and bark mycobiomes [88, 89]. Moreover, early-stage colonizers of resin-rich substrates [24, 90], appear only at trace levels, highlighting a succession toward stress-adapted, substrate-specialized lineages. Classical resinicolous genera (e.g. Sorocybe, Sarea, and Chaenothecopsis (Mycocaliciales) [69]) and hydrocarbon-tolerant genera (e.g. Exophiala, and Debaryomyces [18, 78]), comprise the most dominant, functionally important groups. Conversely, canonically dominant aerial or foliar associates are rare, reflecting limited persistence of above-ground propagules [88, 89, 91]. The ecological uniqueness of RS relies on many low abundant taxa, yet some culturable strains (e.g. Amorphotheca spp., 0.33%; T. versicolor, 0.001%) remain viable under standard conditions (unpublished data).

Beyond known resinicolous and hydrocarbon-associated fungi, RS harbors a considerable fraction of unclassified or incertae sedis OTUs (> 30% at genus level), far exceeding ambiguity in FS (> 9% at genus level). This suggests that resin-saturated habitats may function as potential refugia for fungal “dark matter”: poorly characterized fungal diversity encompassing lineages underrepresented in current databases and possibly including novel taxa [40, 67, 68]. However, this high proportion of unclassified OTUs also highlights methodological limitations inherent to ITS2-based taxonomic resolution and incomplete database coverage. Thus, while some sequences may correspond to undescribed fungi, others likely reflect the restricted discriminative power of the ITS marker, which requires complementary loci (e.g. LSU, EF1α, β-tubulin) for robust identification [40, 67]. This interpretation is consistent with reports of specialized and previously undescribed resin-associated fungi [92]. Overall, the elevated share of unclassified sequences in RS underscores a distinct and insufficiently characterized fungal assemblage, warranting targeted multi-locus and culture-based follow-up studies. The persistence of ancient resinicolous fungi like Sarea and Chaenothecopsis in Paleogene fossil records underscores the deep evolutionary continuity of resin-associated niches [69, 93, 94], while modern resinicolous fungi (e.g. Ganoderma adspersum) colonizing diverse polymeric matrices [95, 96] extend the ecological significance of resin adaptation to anthropogenic substrates. This evolutionary breadth reflects the fundamental selective power of resin chemistry across natural and industrial timescales.

RS resin acid persistence mirrors chemical legacies encountered in pulp-mill effluents and industrial sites [20]. Thus, selecting for fungal consortia capable of modifying such stressors offers significant bioremediation potential. Notably, the dominant RS taxa (e.g. Exophiala, Aspergillus) include genera with opportunistic pathogenic members, suggesting a link between environmental extremotolerance and traits associated with opportunism [97, 98]. However, given the taxonomic resolution of this study, no conclusions can be drawn about the pathogenic potential of the specific taxa present. This observation nonetheless highlights a One-Health/Eco-Health-relevant research direction that warrants further investigation [99, 100].

Far from being marginal, RS mycobiome highlights the potential of fungi to colonize, persist, and drive transformations in extreme ecosystems. Our conceptual strategy of analyzing a composite sample aimed to reduce within site-variability, whereas its comparison with that of a nearby forest soil, allowed to analyze major differences caused by resin-exposure and not by altered geography and climate. These findings expand our view of natural product-shaped environments, proposing resin-rich habitats as “laboratories” for adaptive innovation with applications from pollutant attenuation to biocatalysts discovery. Future work integrating targeted cultivation with deep molecular analyses will further unravel, both the known and unknown, fungal diversity revealed here.

Abbreviations

AA

Abietic acid

CYP

Cytochrome P450 monooxygenases

DBC

Bray–Curtis dissimilarity

DHA

dehydroabietic

ECM

Extracellular matrix

EDS

Energy-dispersive spectroscopy

FS

Forest soil

GSMc

Global Soil Mycobiome consortium

HPLC

High-performance liquid chromatography

HPLC

High-performance liquid chromatography

IPA

Isopimaric acid

IQR

Interquartile range

ITS

Internal transcribed spacer

LC50

Median lethal concentration

LSU

Large subunit ribosomal RNA

MEA

Malt Extract Agar

NCBI

National Center for Biotechnology Information

NMR

Nuclear magnetic resonance

PA

Pimaric acid

PCoA

Principal coordinates analysis

PCR

Polymerase chain reaction

PERMANOVA

Permutational multivariate analysis of variance

PERMDISP

Permutational analysis of multivariate dispersions

PDA

Photodiode array

OC

Organic carbon

OM

Organic matter

OTUs

Operational Taxonomic Units

RS

Resinous soil

SEM

Scanning electronic microscopy

SRA

Sequence Read Archive

TKN

Total Kjeldahl nitrogen

UNITE

Unified system for DNA-based fungal species linked to the classification

UPARSE

Ultra-fast sequence analysis pipeline

βW

Whittaker’s beta diversity

Authors’ contributions

All authors have made important contributions to the acquisition, analysis and interpretation of data and contributed to the drafting of the manuscript: CAMA (sampling); AP, AV (fungal experiments, analytics); PD, IM (technical support); AP, TM (bioinformatics); RE, AB (NMR analyses); AP (manuscript draft); CSP (conceptualization, resources, supervision and final manuscript). All authors read and approved the final version of the manuscript.

Funding

We acknowledge funding from the Fundação para a Ciência e Tecnologia (FCT) through the projects ‘FATE’ (PTDC/CTA-AMB/6587/2020), FATE4FUTURE (nº 2023.16924.ICDT), MOSTMICRO-ITQB Unit (UIDB/04612/2020 and UIDP/04612/2020) and LS4FUTURE Associated Laboratory (LA/P/0087/2020), and by Fundo Europeu de Desenvolvimento Regional (FEDER) under the project “BIOPINUS” (CENTRO-01-0247-FEDER-072630). ÂP and RE are grateful to FCT for the fellowships references SFRH/BD/144593/2019 (10.54499/SFRH/BD/144593/2019), and 2021.06435.B (10.54499/2021.06435.BD), respectively. TM is grateful for the working contract (2023.11076.TENURE.076) financed by national funds under the FCT-TENURE Programme.

Data availability

Raw sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession PRJNA1392999. All processed data supporting the findings of this study, including taxonomic tables, diversity metrics, and functional analyses, are provided in the Supplementary Information.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.