2 MATERIALS AND METHODS
2.1 Taxon sampling, plant material, and growth conditions
Because Cochlearia and Ionopsidium species are rare and often endangered, our study is based upon our germplasm collections. These collections were compiled over the last three decades (BrassiBase [https://brassibase.cos.uni-heidelberg.de/]: Koch et al., 2012; Kiefer et al., 2014; Koch et al., 2018) because our taxon sampling required sufficient and high-quality seed material from wild populations to propagate the respective material used in the freezing experiments. Cochlearia species under study were selected from a range of different habitats, including arctic regions (C. groenlandica ), alpine habitats (endemic C. excelsa from Austrian South-Eastern European Alps/endemic C. tatrae , High Tatra Mountains), cold calcareous springs and creeks (Central EuropeanC. pyrenaica /polish C. polonica ), coastal dune areas, and salt marshes (C. danica /C. aestuaria /C. anglica ).Cochlearia danica , which originally adapted to coastal sand dune areas, migrated along road routes into central Europe, where it remains today (Koch, 1996; Koch, 1997). Accessions of Ionopsidium came from Mediterranean habitats. A total of 34 different accessions, comprising 13 species, were used in the experiments (Fig. 1, Suppl. Material Table 1). Plants were grown and cultivated in a growing chamber for 34-38 weeks with a day/night cycle of 14/10 h and a 30 min transition period, during which the light intensity was gradually lowered. The growth conditions included a permanent temperature of 20°C (± 2°C) and a relative humidity of 50%. Plants were watered regularly and no light or drought stress was omitted. For each accession, 20 individuals were propagated to finally select three randomly selected individuals each (non-flowering and healthy rosette-forming plants) for the two acclimation treatments: (i) three individuals were acclimated at 4°C in a climate chamber for five days at a 65% relative humidity (a slight increase in humidity, compared to 20°C growth conditions, was implemented because of the lower temperature). (ii) Three randomly selected individuals were not acclimated and remained in the original growth chamber at 20°C as a control. The same day/night regime was applied for both treatments, as was chosen for the initial cultivation. Six week old individuals of the Arabidopsis thaliana Col-0 ecotype were used as an internal control.
The aim of this study was to performed electrolytic conductivity measurements on leaves subjected to a freezing regime from 0°C to -10°C at different time points and to calculate lethal experimental freezing temperatures (LT ) for representative species from both genera.LT50 and LT100 (50% and 100% cell membrane damage) values were calculated then using logistic functions fitted to the measured electrolytic leakage values (PEL).
2.2 Electrolytic assays
Electrolytic leakage of detached leaves is commonly used to quantitatively assess the freezing tolerance and cold acclimation potential of plants (Armstrong et al., 2020; Hincha & Zuther, 2014; Wos & Willi, 2015, 2018). The cell membranes are the primary sites of freezing damage. When cell membranes are damaged, the cell’s contents leak out; this leakage can be detected by measuring electrical conductivity because the ionic composition of the water in which the leaves were immersed changes. Therefore, electrolytic leakage is expressed in terms of relative conductivity (Lee & Zhu, 2010; Hatsugai & Katagiri, 2018; Armstrong et al., 2020). Electrolytic leakage was measured following cold exposure and again after boiling at 100°C to ensure 100% leaf damage. The relative leakage, or percentage of leakage, was calculated from the ratio of the two measurements. The lethal temperature (LT50 ), which refers to the temperature that causes 50% electrolytic leakage from cells (or otherwise causes 50% leaf damage), was derived based on the relative leakage. LT100 is the temperature at which 100% of the electrolytes have leaked from the cells and is calculated from the percentage of electrolytic leakage. The electrolytic leakage was measured according to the method described by Thalhammer et al.. al (2014), with some modifications outlined below in detail.
For the electrolyte leakage assay, six different freezing temperatures (0°C, -2°C, -4°C, -6°C, -8°C, -10°C) were selected to follow a respective temperature cline for further regression analyses. One leaf per temperature treatment (0, -2, -4, -6, -8, and-10°C) and three leaves for negative controls were harvested from acclimated (4°C) and non-acclimated individuals (20°C). In total, nine leaves of similar size (appr. 1 cm2) were harvested from each of the three individuals from a single accession, which resulted in 54 leaves and measurements for each accession. Leaves of approximately the same size and thickness were chosen and placed into ddH2O-filled (3 ml) 10 ml DURAN glass tubes that were closed tightly afterwards with metal lids. The negative controls were incubated at 4°C on a shaker at 100 rpm. To control the down-cooling freezing of other samples, tubes were placed in a cooling bath (LAUDA RP2045; LAUDA Scientific, Lauda-Königshofen, Germany). An automatic temperature ramping program was used to ensure a steady, standardized temperature change from 0°C to -10°C for all samples. The temperature was lowered by 2°C during a 3 min period between 45 min stable cooling intervals. Sample collection commenced at the end of each cooling interval, just before the 3 min cooling period started.
Crystallization of the remaining samples was induced using liquid nitrogen 20 min after cooling at-2 °C. Inoculation loops (steel, 0.5 mm) were used to initiate the crystallization. Loops were immersed in ddH2O and subsequently in liquid nitrogen until ice crystals formed; the loops were then placed carefully into the tubes, carefully damaging the samples to induce immediate crystallization of the entire sample. The samples were then removed at various freezing temperatures (0°C to -10°C) and incubated for 48 h at 4°C and 100 rpm on a shaker. Wires were carefully removed, and an additional 2 ml cold ddH2O was added to the tubes. The measurements were performed 24 h later. The electrical conductivity of the solution was measured using a conductivity meter (METTLER TOLEDO LE703; Mettler-Toledo, Albstadt, Germany) after the samples were brought to room temperature. Measurements were performed before (EL0) and after (EL1) boiling the solution for two h at 100°C. Boiling ensured complete destruction of the leaves, leading to 100% electrolytic leakage out of the cells. Finally, the analysis of 34 accessions resulted in 1836 measurements.
2.3 Analysis of electrolytic leakage data
The following equations were used to calculate the percentage of electrolytic leakage (PEL):
EL = EL0/EL1 (2.1)
ELcontrol = EL0,control/EL1,control (2.2)
PEL = (EL–mean(ELcontrol)) x 100 (2.3)
Equations 2.1 and 2.2 calculate the ratio of electrical conductivity of un-boiled (EL0) and boiled (EL1) samples for frozen and control samples. Equation 2.3 calculates the percentage of electrolytic leakage, which is corrected by subtracting the mean ELcontrol value for each pretreatment, as leaf damage could occur from harvesting, incubating in water, and cooling at 4°C. Negative PEL values were corrected to a 0.00% leakage. This correction occurred occasionally for samples cooled at 0°C or -2°C, as leaf damage remained relatively low at these temperatures compared to the control samples. This made it possible for control values to increase.
All calculations were performed in the R statistical environment using R Studio (Version R.4.1.2; R Core Team, 2021). A self-starting model was used to apply a logistic function to the measured PEL values as follows:
PEL = A/[1 + e(xmid− T/scal)] (2.4)
The inflection point (xmid) gives the LT50, and scal is a scale factor. The asymptote (A) was set to 100%, and the input(T) was the temperature. For this purpose, the stats package in R (Version 4.1.2; R Core Team, 2021) implementing functions nls() and SSlogis() was used to write an r-script (Supplementary Material File 1). Values were derived using predict(); this process was performed for measured PEL values of each accession for acclimated and non-acclimated samples, separately. Using this method, the lower asymptote of the curve approached a 0% electrolytic leakage (or zero leaf damage) and the upper asymptote approached a 100% electrolytic leakage (or maximum leaf damage).LT50 and LT100 values for the acclimated and non-acclimated samples were calculated using the model data. Because some leaf damage was caused by the harvesting, incubation in water, and cooling at 4°C, the mean ELcontrol of each pretreatment was subtracted from 100%, and the resulting percentage was used to calculateLT100 values. To assess the variation in the measurements, the mean PEL value and standard error were calculated from replicates of each temperature. The difference between the acclimated LT and non-acclimated LT values was computed to calculate the ΔLT50 and ΔLT100 values. Measured values and calculated values were exported, and the script allowed the computation of a graph showing all PEL values and mean values of the measured data, LT50 ,LT100 , and sigmoidal curves for acclimated and non-acclimated sample accessions.
Further statistical analyses to test differences inLT50 and LT100 values, both acclimated and non-acclimated, were performed using t-tests. To confirm whether variation in lethal values existed among species, an analysis of variance (ANOVA) was applied. A correlation analysis of theLT50 and LT100 values with geographic coordinates (latitude and longitude) as well as with ploidy level of the different species (diploid versus polyploid) was performed using R Studio (Version R.4.1.2; R Core Team, 2021); the data were checked for normality using the Shapiro-Wilk normality test. This was performed using the shapiro. test() function of the stats package. Information on chromosome number and ploidy level is provided in Suppl. Material Table 1 (Koch et al., 1996; Koch et al., 1998; Koch et al., 1999; Koch, 2002; Koch et al., 2003; Koch & Bernhardt, 2004; Cieslak et al., 2007; Koch, 2012; Wolf et al., 2021).
Pearson’s correlation measures a linear dependence between two variables that have a normal distribution: lethal values and latitude/longitude. For visualization, scatterplots were produced using ggscatter() from the ggpubr package (Version 0.4.0), showing lethal values of acclimated and non-acclimated samples at different latitudes and longitudes. A simple regression line was calculated (through add = “reg.line”).
2.4 BioClim data analyses of investigated accessions
Principal coordinate analysis was performed to identify different species groups according to the bioclimatic characteristics of the accession habitats. For this purpose, nineteen bioclimatic variables were downloaded from the WorldClim climate data grid (https://www.worldclim.org; Hijmans et al., 2005) for allCochlearia and Ionopsidium accessions. These included temperature-related (BIO1-BIO11) as well as precipitation-related variables (BIO12-BIO19), namely annual mean temperature (BIO1), mean diurnal range (BIO2), isothermality (BIO3), temperature seasonality (BIO4), maximum temperature of warmest month (BIO5), minimum temperature of coldest month (BIO6), temperature annual range (BIO7), mean temperature of wettest quarter (BIO8), mean temperature of driest quarter (BIO9), mean temperature of warmest quarter (BIO10), mean temperature of coldest quarter (BIO11), annual precipitation (BIO12), precipitation of wettest month (BIO13), precipitation of driest month (BIO14), precipitation seasonality (BIO15), precipitation of wettest quarter (BIO16), precipitation of driest quarter (BIO17), precipitation of warmest quarter (BIO18), and precipitation of coldest quarter (BIO19). Principal coordinate analysis (PCoA) was performed using the multivariate statistical package (MVSP 3.22; Kovach, 2007). Bioclimatic variables and accession numbers were imported in the csv format, data were centered and standardized, and Kaiser´s rule was used to extract the axes that retained factors with eigenvalues greater than one.