The environment is a powerful selective pressure for sessile organisms, such as plants, and adaptation to the environment is particularly important for long-lived species, like trees. Despite the importance of adaptive trait variation to the survival and success of trees, the molecular basis of adaptation is still poorly understood. Gene expression patterns in three closely related, but phenotypically and ecologically divergent, pine species were analyzed to detect differentiation that may be associated with their adaptation to distinct environments. Total RNA of Pinus mugo, P. uncinata and P. sylvestris samples grown under common garden conditions was used for de novo transcriptome assembly, providing a new reference dataset that includes species from the taxonomically challenging Pinus mugo complex. Gene expression profiles were found to be very similar with only 121 genes significantly diverged in any of the pairwise species comparisons. Functional annotation of these genes revealed major categories of distinctly expressed transcripts including: wood trait properties, oxidative stress response, and response to abiotic factors such as salinity, drought and temperature. We discuss putative associations between gene expression profiles and adaptation to different environments, for example: upregulation of genes involved in lignin biosynthesis in the species which have adapted to mountainous regions characterized by strong winds and thick snow cover. Our study provides valid candidates for verification of the importance of the gene expression role, in addition to evidence for selection within genomic regions, in the process of ecological divergence and adaptation to higher altitudes in pine taxa.

Genetic material was obtained from needles of 11 two year-old seedlings of three pine species (P. mugo – 4 samples, M; P. uncinata – 3 samples, UN; P. sylvestris L. – 4 samples, PS) grown in a glasshouse facility of the UK Centre of Ecology and Hydrology in Edinburgh. The seedlings were derived from open-pollinated seeds collected in natural populations in Europe (for details, see Wachowiak et al., 2018). Raw paired-end Illumina reads from teams’ previous RNA-sequencing attempts were used, data is deposited in the European Nucleotide Archive under accession number PRJEB6877 (for precise sample identifiers, please see Table 1 in the article). Details regarding RNA isolation, library preparation and sequencing are specified in Wachowiak et al. (2015).

The raw reads of all 11 samples were quality checked in FASTQC v0.11.9 (www.github.com/s-andrews/FastQC) and used for de novo transcriptome construction by TRINITY v2.11.0 (Haas et al., 2013). The paired-end assembly procedure adapted default parameter values, except it was run with simultaneous cleaning and trimming of reads provided by TRIMMOMATIC v0.39 (Bolger et al., 2014) and contigs below 200 bp were rejected.

TRIMMOMATIC parameters used: "ILLUMINACLIP:anaconda3/envs/Trinity/share/adapters/TruSeq2-PE.fa:2:30:10:2:keepBothReads SLIDINGWINDOW:5:20 LEADING:3 TRAILING:3 MINLEN:30"

The only change in the file made was deletion of the 'TRINITY_' part from the generated sequence names. No more alternations have been introduced.