SHW_montane_willow_fungi_eDNA_2026

Raw sequence data and OTUs of fungi collected from 90 soil samples in 2022. Study sites Our research was conducted at three mountainous areas in Scotland – Ben Lawers National Nature Reserve (NNR), Corrour Estate, and Mar Lodge Estate. These sites are situated across the centre, east, and west of the Scottish Highlands respectively, providing a representative selection of geographic and geological regions within the core range of S. lapponum in this country (Stroh et al. 2023). Each site has a total of at least 150 S. lapponum relict plants distributed across numerous smaller populations. A total of 50,000 montane willows have been planted at Ben Lawers in extensive large herbivore exclosures since 1991 (Watts 2024). The main quantity of planting took place in 2004-2006, including 18,285 S. lapponum from which the planted individuals in this study have been sampled. Planting occurred without fertilizer or ground preparation in clusters to create an open habitat mosaic. The stock was typically 2 years old at the time of planting, and of local origin produced by tree nursery propagation from seed and cuttings collected from the Ben Lawers range and populations on three nearby mountains within 20 km. Sampling design A total of 90 soil samples were collected for eDNA metabarcoding during September and October 2022. Ten samples were gathered from individual relict S. lapponum plants at each of the three study sites (30 samples in total). These ten samples were selected from at least four separate S. lapponum populations distributed across a representative range of relict montane scrub habitat present at each study site. Ten samples were also collected from locations without willows present at each of the three study sites (30 samples in total). These “non-Salix” control samples were characteristic of potential S. lapponum restoration sites separate from relict cliff ledge populations or other ECM tree species, and selected based on relevant altitude, geology, vegetation type, and the associated vascular plant indicator species currently guiding local-scale planting (Table 1). In addition, 10 samples from established S. lapponum individuals planted during 2004-2006 were also obtained from Ben Lawers. We could not replicate this collection elsewhere in Scotland, since no other population is comparable in age and extent of planted montane willows. Finally, a further 10 samples were gathered from established populations of each of the more common generalist species S. aurita and S. repens from across Corrour Estate. Rather than being restricted to cliff ledges, these 10 sample locations at this study site were scattered throughout open hill locations at lower altitudes than S. lapponum Sample collection A sterile apple corer was used to collect five small soil cores from the top 5 cm of soil from below the canopy of every shrub chosen for sampling (Fig. 2). The five soil samples were combined to form a single composite sample. Associated vegetation, moss, and leaf litter was carefully parted to allow access to the soil below, while minimising ground disturbance as much as possible to the sensitive montane locations. Only robust and healthy shrubs at least 50 cm in height and 1 m in diameter were used, to avoid harming young, very old, or already damaged individuals. An exception of a minimum height of 30 cm was made for S. repens which generally has a more creeping and low-statured growth form than S. lapponum and S. aurita. A different apple corer was used for each composite sample to avoid the need for sterilisation during fieldwork in the montane environment. Corers were previously sterilised by immersion in boiling water for 30 minutes. Composite samples from each of the 90 sampling points (individual shrubs or non-Salix locations) were stored in sealed sterile bags, placed in a fridge at the end of each day of fieldwork, and delivered to the laboratory within 48 hours for initial processing involving removal of large stones and freezing prior to eDNA extraction. Sampling was conducted with landowner permission and NatureScot consent for designated areas. DNA extraction Frozen soil was homogenized by cryo-milling with liquid nitrogen, and total nucleic acids were extracted following the method of Griffiths et al. (2000) as modified by Freitag & Prosser (2009). Briefly, liquid N2-chilled 250 mg subsamples were placed in lysing matrix ‘B’ tubes (MP Biomedicals), preloaded with 0.5 ml 5% CTAB in 0.7M NaCl/120mM K2HPO4 extraction buffer, brought to ice temperatures, before the addition of 250 µL of ice-cold Tris-buffered phenol at pH 6.7 ± 0.2 and 250 µL chloroform/isoamyl alcohol (24:1). Samples were homogenised for 2 x 15 s at 5000 rpm on a Precellys 24 tissue homogenizer with a five-minute incubation on ice after each homogenisation cycle. After centrifugation for 10 minutes at 16,000 x g, 4 °C, the upper, aqueous layer was transferred to ice-cold 2 mL MaXtract High Density gel barrier tubes (Qiagen), pre-loaded with 100 µL of each of phenol and chloroform:isoamyl alcohol (24:1). Tubes were gently inverted for ca. 30 s and centrifuged at 20817 x g, 10 min, 4 °C, to allow gel barrier separation of aqueous and organic layers. Traces of phenol were removed with the addition of 500 µL of ice-cold chloroform:isoamyl alcohol. Tubes were gently mixed by inversion and centrifuged at 20817 x g for 10 min. The aqueous phase was transferred to 1.5 ml tubes and 1 mL of 30 % (w/v) PEG 6000 in 1.6 M NaCl was added and mixed gently by inversion before resting on ice for 2 hours. Nucleic acids were pelleted by centrifugation for 20 min at 20817 x g and the supernatant was removed. Pellets were washed twice with 1 mL of 70 % (v/v) ethanol and centrifuged for 10 min at 20817 x g. Ethanol was removed by pipette aspiration and pellets dried at 55 °C, before resuspension in 30 µL of nuclease-free water. Tubes were incubated for 5-15 min at 55 °C. Resuspended extracts were flash frozen in liquid N2 and stored at -80 °C until further analysis. Contamination during the extraction procedure was examined using control extraction blanks that were mock samples of 250 µL molecular grade sterile nuclease free water (Sigma-Aldrich), extracted and cleaned up alongside field samples. Extraction controls were included at a ratio of 1:24 samples. The control extracts were diluted in a similar ratio as the experimental samples and submitted for sequencing along with the sample set. Sequencing 200 ng of DNA from each sample and control extract were shipped to LGC Biosearch Technologies, Berlin, Germany. Amplicons were generated using the fITS7-ITS4 primers for fungi (White et al. 1990; Ihrmark et al. 2012). As expected, controls failed to amplify and were excluded from sequencing. Library preparation and sequencing was performed according to the Ovation Rapid DR Multiplex System (Tecan Trading AG, Switzerland) using the manual 15027617 MiSeq System Guide (15027617), on an Illumina MiSeq V3 (2x300bp) sequencer. Bioinformatics All data were processed using the Qiime2 environment version 2023.5 (Bolyen et al. 2019). Raw data were assessed for read count and quality. ITS amplicon data were trimmed using the ITSxpress plugin to remove the conserved regions surrounding the ITS region and improve taxonomic classification (Nilsson et al. 2008; Rivers et al. 2018). ITS data were clustered into OTUs (Operational Taxonomic Units) at a threshold of 97.5% similarity using the VSEARCH plugin (Rognes et al. 2016). Taxonomic assignments were performed using Qiime2’s Naive Bayes classifier trained using the UNITE database (Quast et al. 2012; Nilsson et al. 2019), without read extraction as this is not seen to improve classification. OTUs were assigned to trophic groups by cross-referencing with the primary lifestyles listed in the FungalTraits database (Põlme et al. 2020).