Methodology:

Soil samples were collected in January-February 2018 from 12 sites between Torres del Paine (Chile) and Lagotellerie Island (Maritime Antarctica). Four to five samples of barren soils (c. 20 g fresh weight) were sampled 1-2 m from Colobanthus quitensis plants into sterile containers that were stored at c. 4 °C before being frozen at -20 °C, prior to transfer to the lab, also at -20 °C.

The samples were defrosted and mixed. Soil pH value was measured in a suspension of 5 g soil and 20 ml of distilled water. Soil C/N ratio was calculated following measurement of total C and N in freeze-dried soils using a TruSpec CN analyser. Water-soluble carbon (C) and nitrogen (N) were extracted from 5 g of fresh soil in 25 ml distilled water for 60 min. After filtration through filter paper, the concentrations of total dissolved organic C were measured using a Shimadzu TOC-L CSH/CSN total organic C analyzer. Total dissolved N was measured with a FIAstar 5000 flow injection analyser after digesting extracts in sulphuric acid with selenium as the catalyst. Dissolved concentrations of NO3--N and NH4+-N were measured using a FIAstar 5000 flow injection analyser using cadmium reduction and indophenol blue and molybdenum blue methods, respectively.

DNA was extracted from 0.25 g subsamples of freeze-dried soil using a FastDNA Spin Kit for Soil. The internal transcribed spacer (ITS) 2 region of fungal DNA was amplified using the primers ITS4 and ITS7. PCR amplifications consisted of two steps: (i) primers and template DNA (1 µl) were added to Illustra puReTaq Ready-To-Go PCR Beads and were thermocycled at 94 °C for 2 min, then 35 cycles at 94 °C for 30 s, 56 °C for 30 s and 72 °C for 30 s, followed by an extension step at 72 °C for 5 min and (ii) 2 µl of the diluted PCR product was added to a reaction mixture consisting of 0.15 µl DNA polymerase (AccuPrimeTM Taq DNA Polymerase High Fidelity, Invitrogen), 2 µl 10 x AccuPrimeTM PCR Buffer II, 1 µl (770 nM) of custom-tagged forward and reverse primers and 13.85 µl sterile water. Mixtures were thermocycled at 94 °C for 2 min, then 14 cycles at 94 °C for 30 s, 56 °C for 30 s and 72 °C for 30 s, followed by an extension step at 72 °C for 5 min. Amplicons were purified in 1% agarose gels using a Montage Gel Extraction Kit. The concentrations of amplicons were measured using Qubit dsDNA HS assays and fluorescence was measured using a Qubit fluorometer. Samples, pooled in equimolar amounts, were paired-end sequenced over two Illumina MiSeq flowcells.

Demultiplexed sequences from each sample were trimmed of primers using cutadapt 1.18. DADA2 v1.22.0 was subsequently used in R v 4.1.1 for quality filtering, amplicon sequence variant inference and taxon assignment. Quality filtering removed all sequences with ambiguous bases (trncQ=11), >2 expected errors and lengths <50 bp. Error rates were estimated for forward and reverse sequences and amplicon sequence variants inferred before merging with a 30 base pair minimum overlap. Chimeras were removed and taxonomies assigned to resulting amplicon sequence variants using the RDP classifier with the sh_general_release-dynamimc_all_25.07.2023 reference from UNITE. The decontam v1.14.0 package was then used to remove putative contaminants identified in negative controls (threshold value 1.0). Taxonomies and occurrences of the removed amplicon sequence variants were cross-checked through manual inspection. The taxonomy-assigned, decontaminated amplicon sequence variants were clustered into operational taxonomic units (OTUs) at 98% similarity over 90% of sequence length using the BLASTCLUST algorithm in the package blast-legacy v2.2.26. For each OTU, the taxonomies for its constituent amplicon sequence variants were compared to validate the OTU taxonomy used for all subsequent analyses.

Le...(20)

Data collection:

Instruments

Qubit fluorometer (Thermofisher, USA)

Illumina MiSeq System (Illumina, San Diego, California, USA)

TruSpec CN analyser Leco (St Joseph, MI, USA)

Shimadzu TOC-L CSH/CSN total organic C analyzer (Shimadzu, Kyoto, Japan)

FIAstar 5000 flow injection analyser (Foss Tecator, Hoganas, Sweden)

Software

R versions 4.1.1 and 4.1.3

R packages: terra v 1.6-47, glmnet v4.1-8, DADA2 v1.22.0, decontam v1.14.0 and indicspecies v1.7.15

Python: cutadapt 1.18

UNITE: RDP classifier with sh_general_release-dynamimc_all_25.07.2023

blast-legacy v2.2.26

Web of Science

CHELSA database v. 2.1

CHELSA-BIOCLIM+ database

iNEXT package

Data quality:

Raw data are shown. The data have not been cleaned.