We exposed natural spring communities from coastal Svalbard (Norway) for 2-3 weeks to stable temperature treatments (2°C, 6°C, 9°C), as well as to repeated 5-day heatwaves of differing intensity (6°C and 9°C, Figure 1). By excluding grazers and ensuring nutrient replete and stable light conditions, we focused on temperature effects only. Species composition was assessed by 18S rRNA metabarcoding. In the same filtration setup as described above, 400-500 ml samples were filtered on 0.8 µm PC filters (Nucleopore, Whatman, UK), and immersed in 650 µL of preheated extraction buffer (SL1 of the NucleoSpin Soil extraction kit (see below) at 50°C) to be stored at -20°C until further analysis. After thawing and cell disruption with a MagNa Lyser (Roche Diagnostics, Switzerland), DNA extraction was performed according to the manufacturer’s protocol using the NucleoSpin Soil extraction kit (Macherey-Nagel GmbH, Germany). Amplicon libraries of the V4 region (18S rRNA gene) were generated using the standard 16S Metagenomic Sequencing Library Preparation protocol (16S Metagenomic Sequencing Library Preparation, Part #15044223 Rev. B. Illumina, USA) using the forward primer CCAGCASCYGCGGTAATTCC and reverse primer ACTTTCGTTCTTGAT (Badley et al. 2016). Single samples were indexed using the Nextera XT Index Kit v2 primers (Illumina, USA) and pooled for sequencing on a MiSeq sequencer (Illumina, USA). Results were demultiplexed and FASTQ sequence ?les generated using the Generate FASTQ work?ow of the MiSeq sequencer software, yielding a total of ~10x106 raw amplicons. Primers were removed with cutadapt v2.8 (Martin 2011), and further processing of the sequence data was performed using the DADA2 R package v1.18 (Callahan et al. 2016). Reads were trimmed (forward reads after 240- 260 bp, reverse reads after 200 -210 bp) and denoised, before paired-end reads were merged (min. overlap 50 bp, no mismatches), and predicted chimeras were removed, yielding a total of ~6.8x106 filtered amplicons. Taxonomic assignment of the resulting amplicon sequence variants (ASVs) was performed using the reference databases PR2 (v4.12.0). For downstream analyses in the software R (vers. 4.3), non-phototrophic taxa were removed, as well as ASVs with a count of less than 10 reads in replicate sample means.
