Study material and experimental design
The northern populations of V. alpina came from Scandinavia (Oppland, Norway and Västerbotten, Sweden; hereafter referred to as NOR and SWE, respectively ; Fig. 1), while the southern populations came from two Italian mountain regions (Monte Sobretta, Central Alps and Monte Prado, Northern Apennines; hereafter referred to as SOB and PRA, respectively; Fig. 1). The four provenance sites differed in terms of climatic conditions besides latitude and corresponding location within the species’ range: NOR and SOB were colder sites, while SWE and PRA were warmer sites (Abeli et al. 2015; Fig. 1). In summer 2012, seeds of V. alpina plants were collected at each site. The plants were cultivated at the Botanical Garden of the University of Pavia, at approximately 85 m above sea level, and produced seeds for the first time in 2015. Seeds of all populations were harvested and kept under standard seedbank conditions (-20°C after drying at 15%RH, 15°C, FAO 2013) until used for cultivation in controlled conditions. Leaves of six to eight mature plants per population were harvested and dried in silica gel for subsequent molecular analyses.
In 2016, the seeds were sown in Petri Dishes filled with 1% Agar in distilled water + 125 mg/L of Gibberellic Acid (GA3) and incubated at 20°C in light- and temperature-controlled incubators. Because the northern and the southern populations of V. alpina differed in germination timing and dormancy levels (Mondoni et al. 2018), we used GA3 to obtain synchronous germination and avoid different developmental stages at the beginning of the growth chamber experiment. The selected amount of GA3 proved to synchronise germination without inducing deformities in the seedlings. Soon after radicle emission, 120 seedlings per population were transferred each into a 3-cm plastic pot with standard soil (Hochmoor substrate, Terflor®, Italy) and assigned to two treatments (60 plants per treatment in 20 replicates of three plants each), using a light- and temperature-controlled growth chamber (ClimaCell®, MMM Group, Germany) as detailed below. The base treatment simulated the average summer temperature at the warmest southernmost site (PRA) based on temperature measured with a datalogger (MLog5W, Geoprecision, Germany) placed in that site from 30 August 2013 to 24 July 2015. The extreme treatment simulated the temperature values recorded by the same datalogger at PRA during a summer heat wave in July 2016. In particular, the extreme treatment differed from the base treatment for two day-temperature peaks of 22°C at dawn and 30°C during the day, respectively. Both temperature peaks lasted one hour (Fig. 2). In both treatments the plants were exposed to