Planta https://doi.org/10.1007/s00425-018-2939-1

ORIGINAL ARTICLE

Mechanisms of frost resistance in Arabidopsis thaliana Imke I. Hoermiller1 · Moritz Ruschhaupt2 · Arnd G. Heyer1  Received: 29 March 2018 / Accepted: 8 June 2018 © Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract Main conclusion  Freezing resistance strategies vary in Arabidopsis depending on origin. Southern accessions may avoid or tolerate freezing, while northern ones are always tolerant and reduce the proportion of freezable tissue water during acclimation. Survival of sub-zero temperatures can be achieved by either avoiding or tolerating extracellular ice formation. Conflicting evidence has been presented showing that detached leaves of Arabidopsis thaliana are either freeze avoiding or tolerant. Here, we used three different natural Arabidopsis accessions from different habitats to investigate the frost resistance strategy of whole plants in soil. Plants were cooled to fixed temperatures or just held at their individual ice nucleation temperature for different time intervals. Tissue damage of whole plants was compared to the standard lethal temperature determined for detached leaves with external ice nucleation. While all detached leaves survived freezing when ice nucleation was externally initiated at mild sub-zero temperatures, whole plants of the southern accession behaved as freeze avoiding in the non-acclimated state. The northern accessions and all cold acclimated plants were freezing tolerant, but the duration of the freezing event affected tissue damage. Because this pointed to cell dehydration as mechanism of damage, the proportion of freezable water in leaves and osmolality of cell sap was determined. Indeed, the freezing tolerant accession Rsch had a lower proportion of freezable water and higher cell sap osmolality compared to the sensitive accession C24 in the cold acclimated state. Keywords  Cold acclimation · Differential scanning calorimetry · Differential thermal analysis · Freeze avoidance · Freezing tolerance · Ice nucleation Abbreviations AC Cold acclimated INT Ice nucleation temperature LT50 Lethal temperature for 50% NA Non-acclimated

Electronic supplementary material  The online version of this article (https​://doi.org/10.1007/s0042​5-018-2939-1) contains supplementary material, which is available to authorized users. * Arnd G. Heyer [email protected]‑stuttgart.de 1

Department of Plant Biotechnology, Institute of Biomaterials and Biomolecular Systems, University of Stuttgart, Pfaffenwaldring 57, 70569 Stuttgart, Germany



Department of Botany, Science Center Weihenstephan, Technical University Munich, Emil‑Ramann‑Straße 4, 85354 Freising, Germany

2



Introduction Climate change will increase the incidence of weather extremes, and thus investigation of abiotic stress resistance mechanisms will become increasingly important. The ability to withstand freezing temperatures is critical to the survival of many plant species (Wisniewski et al. 2014), and it is also an important factor in agricultural production. The US Food and Agriculture Organization declares that in the USA higher economic losses have been caused by freezing damage of crops than by any other weather hazard (White and Haas 1975; Snyder and de Melo-Abreu 2005), and corresponding reports exist for European countries (Nannos et al. 2013; Papagiannakis et al. 2014). As an example, a severe cold wave had come over almost all of the USA in January 2018, and even down to Florida, with temperatures down to 17 °C below normal. The extent of crop damage by low temperature depends on several factors, such as the minimum temperature, the duration of the frost event, and the state of development of the plant (Papagiannakis et al. 2014). It

13

Vol.:(0123456789)

Planta

is, therefore, important to understand the mechanisms of damage and plant winter hardiness to support breeding for frost resistance. Plants have evolved several strategies to increase their freezing resistance when exposed to low, non-freezing temperatures in a process known as cold acclimation. Cold acclimation is the result of highly complex physiological adjustments that include induction of genes, enhancement of anti-oxidative defense, changes in lipid and protein composition, and increases in metabolite concentrations (Gusta and Wisniewski 2013; Arias et al. 2015). Among the latter, accumulation and sub-cellular re-distribution of soluble sugars is considered as an essential component (Nägele and Heyer 2013; Hoermiller et al. 2016). Historically, frost resistance strategies are differentiated between freeze avoidance and freezing tolerance. Both are under genetic control and have evolved in response to selection pressure (Wisniewski et al. 2014). Avoidance of freezing may occur by thermal insulation of freezing-sensitive organs, such as, e.g., flowering stalks, or by super-cooling, i.e., cooling of a liquid below its freezing point without ice crystal formation (Arias et al. 2015). While the ability to avoid freezing is often accompanied by certain structural features, e.g., rigid cell walls, reduced intercellular spaces, and low apoplastic water content (Goldstein et al. 1985; Arias et al. 2015), freezing tolerance is associated with biochemical mechanisms that allow plants to control ice formation in extracellular spaces and tolerate cellular dehydration without severe tissue damage (Sakai and Larcher 1987; Lipp et al. 1994; Jacobsen et al. 2007; Wisniewski et al. 2014). Freezing tolerant plants profit from high concentrations of low-molecular weight solutes that ameliorate cellular dehydration but also provide non-colligative protection of cell membranes (Hincha et al. 2002) and proteins (Xie and Timasheff 1997). Interestingly, also the super-cooling capacity of freeze avoiding plants increases with higher concentrations of low-molecular weight solutes (Schulze et al. 2005; Kasuga et al. 2007), probably due to a reduced proportion of apoplastic water (Arias et al. 2015). Given that cell wall rigidity is also advantageous for freezing tolerant plants, as it confers greater mechanical resistance to the physical pressure exerted by extracellular ice growth, it is difficult to discriminate freeze avoiders and tolerators based solely on morphological or metabolic parameters. Thus, similarity of the ice nucleation temperature (INT), which is the temperature of tissue freezing, and the leaf lethal temperature ­(LT50), which is defined as 50% leakage of the plasma membrane, is often used to classify freeze avoiders. In contrast, freezing tolerant plants are characterized by a L ­ T50 that is significantly lower than their INT (Lipp et al. 1994). Cold tolerance of Arabidopsis thaliana accessions varies considerably depending on its origin (Lipp et al. 1994; Zhen and Ungerer 2008; Zuther et al. 2012). This raises the

13

possibility that also the cold resistance strategy may vary among populations. The objective of this study was the analysis of cold resistance strategies of three A. thaliana accessions, originating from the Iberian Peninsula (C24), Central Europe (Col-0), and Russia (Rsch). Plants were analyzed in the acclimated as well as non-acclimated state to figure out whether the significant differences in frost resistance found among populations coincides with different resistance mechanisms that could be traced back to specific traits.

Materials and methods Plant material and growth conditions Arabidopsis thaliana (L.) Heynh (Brassicaceae), accessions Col-0, C24, and Rsch (www.arabi​dopsi​s.org), were grown in GS90 soil and vermiculite (1:1) in a growth chamber at 8 h light (110 µmol m−2 s−1; 22 °C)/16 h dark (16 °C) to avoid floral induction and foster biomass formation. After 5 weeks, plants were transferred to long day, 16 h light (110 µmol m−2 s−1; 22 °C)/8 h dark (16 °C). The relative humidity was 70%. Plants were watered every second day and fertilized with standard nitrogen–phosphate–potassium fertilizer 3 and 5 weeks after sowing. One week after transfer to long-day conditions, a set of plants was used for measurements and another set was shifted to a growth chamber with a 16 h/8 h light–dark regime at 4 °C and a light intensity of 90 µmol m−2 s−1 for one week. All plants were still in the vegetative state, when used for experiments.

Thermal imaging (TI) and differential thermal analysis (DTA) Soil-grown plants in 5  cm pots were put in styrofoam containers, and the soil was separated from the rosette by two pieces of styrofoam sheet, letting a small hole for the hypocotyl. This allowed isolation of the rosette against freezing of the soil, which would have interfered with TI and DTA recordings. For measurements, plants were cooled in a jacketed vessel coupled to a thermostat at a rate of − 2 °C ­h−1 and monitored with an infrared camera (A35, Flir Systems, Frankfurt a.M., Germany) at a frequency of one picture per 20 s using the Flir Research IR software. A PT100 resistor was mounted on one of the leaves to measure the absolute temperature by a linearization circuit based on a Max4236_37A operational amplifier (Maxim Integrated, San Jose, CA, USA). Temperature recording, using an AD converter coupled to a computer, was used for controlling the thermostat CC 305 (Huber AG, Offenburg, Germany). Control of the thermostat was provided by custom software allowing temperature recording, setting of the cooling rate, the hold-time at the ice nucleation temperature (INT), and

Planta

the gradient for thawing the plant to a temperature of 5 °C. Detection of INT was achieved by setting a threshold for a temperature increase of more than 0.7 °C within 40 s.

Electrolyte leakage measurements The release of electrolytes from injured cells was used to quantify tissue damage. In contrast to re-growth this method yields quantitative results for individual samples and was therefore preferred, because data recording for each individual sample lasted between 8 and 18 h, thus precluding the analysis of several hundreds of samples necessary to evaluate re-growth. The electrolyte leakage method is widely used and was shown to yield reliable values for tissue damage (Zuther et al. 2012). In brief, following freeze–thaw cycles, whole rosettes were detached from roots at the hypocotyl and transferred to 50 ml polyethylene vessels filled with 35 ml of deionized water. After incubation over night at 4 °C on a shaker, 500 µl of the bathing solution were mixed with 2.5 ml deionized water before the measurement of electrical conductivity by an LR 325 electrode coupled to an Inolab 740 system (WTW, Weilheim, Germany). The vessels were then incubated at 95 °C for 30 min, conductivity was measured again, and relative damage was expressed as the conductivity ratio before and after boiling of the sample.

LT50 determination on detached leaves The 50% lethality temperature was determined as described by Knaupp et al. (2011). In brief, three rosette leaves, taken from three individual plants, were placed in glass tubes containing 200 µl of deionized water. Eleven tubes per sample set were transferred to a programmable cooling bath set to − 1 °C, while one control tube was left on ice during the entire experiment. After temperature equilibration at − 1 °C, ice crystals were added to the tubes to initiate freezing. Samples were then cooled at a rate of − 2 °C h­ −1 over a temperature range of − 1 to − 15 °C for non-acclimated (NA) plants, and − 1 to − 18 °C for acclimated (AC) plants, and samples were taken from the bath at various time points. After slowly thawing the samples at 4 °C, relative electrolyte leakage was determined as described in the previous paragraph. Kill curves were fitted to the dataset using the drm function of package drc (Ritz et al. 2015) provided by the R statistical software package (R Core Team 2016). The ­LT50 was calculated using the maED function with Buckland model averaging.

Differential scanning calorimetry Thermal analysis was performed with a differential scanning calorimeter DSC1 (Mettler-Toledo, Gießen, Germany) equipped with an Intercooler II unit and a MultiSTAR HSS7

ceramic sensor. Indium (melting point 156.6 °C, enthalpy 28.7 J g−1) was used as enthalpy standard for calibration. Leaf disks of at least 3 mg were sealed in 40 μl aluminum crucibles (Mettler-Toledo). An empty crucible served as reference. The temperature program, starting at 25 °C, consisted of the following segments: (1) 25–5 °C at − 10 °C/s, (2) 5–0 °C at − 1 °C/s, (3) 0 to − 18 °C at − 0.067 °C/s, (4) − 18 °C for 2 min, and (5) − 18 to 25 °C at 10 °C/s. Dry nitrogen was employed as purge gas for the calorimeter head. Data analysis was performed with the STARe SW 9.20 software (Mettler-Toledo). After adjusting a baseline to the heat flow curve, the heat flow of the freezing exotherm was integrated and compared to the theoretical energy of fusion expected for complete freezing of tissue water. The amount of tissue water per sample fresh weight was determined by weighing leaf blades of the same plant without midrib before and after complete drying in a vacuum oven at 105 °C.

Cryoscopy of tissue sap Entire plant rosettes were squeezed rapidly with a custom build juice extractor, and the collected tissue sap was boiled for 5 min at 95 °C before transfer to ice until measuring. A cryoscope (Knauer, Berlin, Germany) consisting of a universal temperature recorder and a cooling unit was used for measuring osmolality of the samples based on freezing point depression. The number of replicates for each accession × treatment was 7–12.

Statistical analysis The R software package, version 3.3.1, was used for statistical analysis (R Core Team 2016). If not otherwise specified, two-way ANOVA was performed using genotype (C24, Col0, Rsch) and growth condition (NA, AC) as factors. Tukey’s HSD, as implemented in the “agricolae” package (de Mendiburu 2016), was used as post hoc test at a significance level of P < 0.05. Data are visualized as bar charts of the mean with standard error of the mean.

Results Determination of freezing damage and INT in different Arabidopsis accessions To assess tissue damage caused by freeze–thaw events in intact plants, whole plants in soil were cooled down to a fixed temperature of −  10  °C at a rate of −  2  °C ­h−1 before thawing at a rate of 3 °C h­ −1 until a temperature of 5 °C was reached. Tissue damage was measured as relative electrolyte leakage from the rosette (Fig. 1). All NA plants suffered electrolyte loss of approximately

13

Planta

Fig. 1  Relative electrolyte leakage from whole plant rosettes after a freeze–thaw cycle with minimal temperature −  10  °C. Shown are means for the three Arabidopsis accessions Rsch, Col-0, and C24 in non-acclimated (NA) and acclimated (AC) condition ± standard error of the mean (SE; n = 5). Letters above the bars indicate the result of an ANOVA for genotype and treatment effects. Groups sharing the same letter are not significantly different (P < 0.05)

Fig. 2  Ice nucleation temperature (INT) of whole plant rosettes, accessions Rsch, Col-0, and C24, in non-acclimated (NA) and acclimated (AC) conditions. Bars represent means ± SE for at least ten replicates. Letters below the bars indicate the result of an ANOVA for genotype and treatment effects. Groups sharing the same letter are not significantly different (P < 0.05)

Tissue damage at INT and impact of freezing time 60%. However, following cold acclimation, Col-0 and Rsch plants showed conductivity values of less than 15%, which matched background levels of control plants not exposed to freezing temperatures. C24 plants showed severe damage no matter of the acclimation status. This shows that either they do not tolerate ice formation, or their ­LT50 was higher than − 10 °C. Because ice formation is an exothermal process that converts kinetic energy of water molecules into heat, when molecules become locked in the crystal lattice, an exotherm in the temperature recording during the cooldown is an immediate indicator of a freezing event. Recordings of whole plant rosettes with an IR camera mounted above the rosette visualized the freezing process by an intense illumination of the rosette (Suppl. video V1). With very few exceptions, ice formation spread rapidly throughout the entire rosette, indicating distribution of the freezing event over the vascular system as reported in other species (Hacker and Neuner 2008). Only plants with homogeneous ice formation within the whole rosette were included in the subsequent analysis of tissue damage to make sure that the tissue passed a freeze–thaw cycle. Large differences between genotype and treatment groups were observed for INT with the means ranging from − 4.3 to − 6.0 °C. Based on a large dataset of at least ten replicates per group, a two-way ANOVA for genotype and condition (non-acclimated, NA, vs. acclimated, AC) showed that INT was not different in NA plants and shifted to lower values during acclimation in Rsch and Col-0, but not in C24 (Fig. 2). Thus, a significant genotype x condition interaction was revealed (F2,148 = 8.127, P = 0.000448), proving that the three genotypes responded differently to the cold acclimation treatment.

13

The cooldown to −  10  °C revealed that all genotypes underwent tissue freezing, and only acclimated Col-0 and Rsch survived this treatment. However, it remained unclear whether damage occurred at INT or at lower temperatures. Thus, an experiment was designed where plants were monitored by an IR camera during the cooldown to capture their individual INT, which then triggered the holdtemperature that was kept for 5 h before thawing again at a rate of 3 °C h −1. The program for the cooling bath was set to ensure an interval of 300 min between reaching INT and then 0 °C during the warm-up. This implied that, because of variation in INT, the time span with constant temperature varied. Schematics of the various temperature schedules used in this study are presented as Suppl. Fig. S1. Only C24 NA plants became severely damaged under these conditions, while all other plants displayed electrolyte leakage between 10 and 20% (Fig. 3). Although leakage values for Col-0 NA and Rsch NA appeared slightly higher than in AC plants, no significant difference could be detected in an ANOVA for genotype and treatment effects. The results clearly showed that a short-term freezing event could be tolerated by all AC plants, while accessions differed in the NA state. It has been demonstrated that the duration of a freezing event has an impact on the mode and severity of damage to isolated mesophyll protoplasts (Nagao et al. 2008). Thus, in a second experiment (Fig. 4), the time interval from ice formation until thawing was extended to 10 h and results were compared to those shown in Fig. 3 by genotype-specific t tests. This revealed a significant increase in tissue damage for Col-0 NA and C24 AC (P < 0.05), demonstrating that the effect of duration of the freezing event was influenced by genotype as well as condition.

Planta

Fig. 3  Relative electrolyte leakage from whole plant rosettes after a freeze–thaw cycle with INT as minimal temperature and a freezing hold-time of 300  min. Shown are means for the three Arabidopsis accessions Rsch, Col-0, and C24 in non-acclimated (NA) and acclimated (AC) condition ± standard error of the mean (SE; n = 5). Letters above the bars indicate the result of an ANOVA for genotype and treatment effects. Groups sharing the same letter are not significantly different (P < 0.05)

Fig. 4  Relative electrolyte leakage from whole plant rosettes after a freeze–thaw cycle with INT as minimal temperature and a freezing hold-time of 600  min. Shown are means for the three Arabidopsis accessions Rsch, Col-0, and C24 in non-acclimated (NA) and acclimated (AC) condition ± standard error of the mean (SE; n = 5). Letters above the bars indicate the result of an ANOVA for genotype and treatment effects. Groups sharing the same letter are not significantly different (P < 0.05)

Nucleated freezing in detached leaves In the experiments reported so far, it turned out that C24 NA did not tolerate ice formation, either because INT was below or at ­LT50, or the duration of the freezing experiment was too long already at 5 h. Thus, an experiment with detached leaves was performed, because ice nucleation can easily be initiated at temperatures near the freezing point of water. Again differential thermal analysis, i.e., the temperature difference between the leaves and the environment, was used to monitor ice formation. A broad exotherm was registered starting immediately with the

Fig. 5  Ice nucleation temperature (INT) of detached leaves with external ice nucleation; accessions Rsch, Col-0, and C24, in nonacclimated (NA) and acclimated (AC) conditions. Bars represent means ± SE (n = 5). Letters above the bars indicate the result of an ANOVA for genotype and treatment effects. Groups sharing the same letter are not significantly different (P < 0.05)

Fig. 6  LT50 values calculated from relative electrolyte leakage of detached leaves exposed to minimal temperatures in the range of 0 to − 12 °C. Shown are means for the three Arabidopsis accessions Rsch, Col-0, and C24 in non-acclimated (NA) and acclimated (AC) conditions. Shown are means ± SE (n = 5). Letters above the bars indicate the result of an ANOVA for genotype and treatment effects. Groups sharing the same letter are not significantly different (P < 0.05)

addition of crystals at around − 1 °C (Fig. 5). Although slightly lower values of INT were recorded for C24 under both conditions, no significant effect of genotype or acclimation state could be detected. Parallel to the thermal analysis, leaves were sampled at different temperatures to determine the L ­ T50 using the classical electrolyte leakage test (Fig. 6). This revealed significant effects of genotype (F2,15 = 23.701; P < 0.001) and condition (F1,15 = 75.440; P < 0.001) as well as a genotype × condition interaction (F2,15 = 4.168; P < 0.05). Most important, the L ­ T50 of C24 NA was about 2 °C lower than INT, demonstrating that also in these plants ice formation on its own did not cause severe injury, when it occurred at a sufficiently high temperature. The kill curves, from which ­LT50 values were taken, are shown in Suppl. Fig. S2.

13

Planta

Differential scanning calorimetry and cell sap osmolality For various biological systems, it has been demonstrated that reducing the amount of freezable water (­ fH2O) is part of the strategy to avoid freeze injury (Vertucci and Stushnoff 1992; Block 2002). Therefore, we investigated whether the amount of freezable water in leaves of the differential freezing tolerant accessions was correlated with tissue damage during freeze–thaw cycles. To quantify ­fH2O, Differential Scanning Calorimetry (DSC) was employed, and the integral over time of the heat flow during ice formation in leaf disks was used to calculate the amount of frozen water. The proportion of freezable water in NA leaves was not different between genotypes (Suppl. Fig. S3), while a significant genotype effect was obtained for AC samples (Fig. 7a). Rsch AC had a significantly lower proportion of freezable water compared to C24 AC (65 vs. 69%), while Col-0 AC behaved intermediate (67%). To figure out whether differences in the amount of freezable water depended on cell sap osmolality, cryoscopy of cell sap squeezed from entire plant rosettes was

Fig. 7  Proportion of freezable water in leaf disks (a) and osmolality of tissue sap from whole rosettes (b) of the three Arabidopsis accessions Rsch, Col-0, and C24 in acclimated condition. Shown are means ± SE (n = 5). Letters above the bars indicate the result of an ANOVA for genotype effects. Groups sharing the same letter are not significantly different (P < 0.05)

13

employed (Fig. 7b). This revealed that among acclimated plants, Rsch had the highest osmolality (408 mosmol/kg), followed by Col-0 (375 mosmol/kg) and C24 (340 mosmol/ kg). Again no difference between genotypes was found for the NA condition (Suppl. Fig. S2).

Discussion Frost resistance strategies Plants are considered freezing tolerant, when they survive extracellular ice formation and the associated dehydration of the protoplast, the latter being strongly dependent on osmolyte concentration of the cell sap, the degree of hardening, and the state of maturity of the whole plant (Levitt 1980; Sakai and Larcher 1987). In the present study, three accessions of Arabidopsis thaliana that span the latitudinal range of the species were used to study mechanisms of cold resistance in intact plants instead of detached leaves to keep plant integrity and to resemble the natural situation as close as possible. Since most of the free water in a plant tissue forms ice at temperatures between 0 and − 10 °C (Burke et al. 1976; Levitt 1980), plants were exposed to a 5 h temperature ramp from 0 to − 10 °C to ensure complete freezing, and tissue damage was determined after slow thawing. During the cooldown, INT of the different accessions varied in the range of − 4 to − 6 °C. It is well documented that ice formation is a stochastic process, with variation of about 2 °C especially for samples in the µl range (Wilson et al. 2003). However, in our study with 10–30 replicates per genotype and condition, variation primarily depended on the acclimation status: for Col-0 and Rsch, INT was shifted by about − 1.5 °C during acclimation, while no significant change occurred for the cold sensitive accession C24. A shift of INT by about − 1 °C in Col-0 as an effect of cold acclimation has already been described (Reyes-Diaz et al. 2006), and a lowered INT in cold acclimated, freezing tolerant plants has also been observed in other species, such as, e.g., rape seed (Gusta et al. 2004). When exposed to a minimal temperature of − 10 °C, Col-0 and Rsch clearly showed tolerance to ice formation in the AC state with only background levels of electrolyte leakage, while all NA plants and also C24 AC were severely damaged. This contrasted with earlier work by Reyes-Diaz et al. (2006), stating that detached Col-0 leaves would avoid but not tolerate freezing, regardless of their cold acclimation status. However, in a close relative, Arabidopsis kamchatica, freezing tolerance in the acclimated state was clearly demonstrated (Armstrong et al. 2015). Considering conflicting evidence in the literature, Armstrong et al. (2015) suggested that for A. thaliana survival of a freezing event might depend

Planta

on the cooling rate and the presence or absence of external ice nucleation. To better discriminate between damage caused by the incidence of freezing and the absolute lowest temperature experienced, we designed an experiment, where plants were exposed to exactly their INT and held in a frozen state for a defined time interval of 5 h. This takes into account the stochastic nature of ice formation: all samples were cooled to their individual INT and kept at that temperature, no matter at which temperature ice formation occurred. This was achieved by controlling the cooling process based on the detection of the freezing exotherm of the tissue. All plants survived ice formation except C24 NA. The fact that C24 AC tolerated ice formation, while C24 NA did not, clearly shows that C24, like the other accessions, is able to acclimate to low temperature as has been shown earlier (Zuther et al. 2012). Comparing the data shown in Figs. 1 and 3, this reveals that the L ­ T50 of C24 AC must lie in the temperature range between − 5 and − 10 °C, which is in agreement with published data (Zuther et al. 2012). Considering geographical origin, it appears that the middle European (Col-0) and northern (Rsch) accession tolerated freezing independent of the acclimation status, while the spring ephemeral C24 may show seasonal variability in its frost resistance strategy. This may be advantageous in warmer climates, because freeze avoidance by super-cooling is more useful for surviving short cold snaps during periods of metabolic activity and development, while freezing tolerance is superior in overwintering (Sakai and Larcher 1987). Thus, a strict distinction between freezing tolerance and freeze avoidance as mutually exclusive strategies may in fact not be appropriate as already indicated (Gusta and Wisniewski 2013; Armstrong et al. 2015).

Mechanisms of damage Different processes such as intracellular ice formation or lipid bilayer events, e.g., membrane phase transition or membrane fusions have been described to cause cell injury at low temperatures (Nagao et al. 2008). The latter as well as protein denaturation at low temperature result from cellular dehydration. While the degree of dehydration resulting from extracellular freezing depends upon the lowest temperature of exposure, the damage to cellular components such as membranes and proteins caused by dehydration also depends on the duration of the water deficit (Hacker et al. 2011). For isolated protoplasts of Col-0 an effect of duration of sub-zero exposure not only on survival but also on the mode of damage was observed: While short-term freezing caused interbilayer events such as membrane fusion at the plasma-lemma, long-term freezing caused vesiculation of the membrane similar to effects of high salt concentrations (Nagao et al. 2008).

A significant effect of extending the freezing time from 5 h to 10 h was observed for Col-0 NA, while C24 NA and Rsch NA did not respond, although due to different reasons: Rsch NA survived both treatments, while C24 NA was killed already after 5 h of freezing. Again this indicates that in a state of metabolic activity, long-term dehydration of the protoplast may be deleterious in Arabidopsis accessions not adapted to harsh climate conditions. It has been reported that “Spring accessions” of Arabidopsis respond to moderate water deficit by increasing photosynthetic activity to enhance root growth, while “Winter accessions” improve water use efficiency (Des Marais et al. 2012). Although both responses are components of a drought avoidance strategy, the former appears contra-productive under severe drought exposure and may thus indicate inability of the spring accessions to endure long-term dehydration. It is interesting to note that the Col-2 accession studied by Des Marais et al. (2012) was classified as Spring accession with a somewhat intermediate behavior. In the AC state, duration of freezing increased damage only in C24, indicating that both cold tolerant accessions, Col-0 and Rsch, developed dehydration tolerance or reduced the probability of intracellular ice formation during cold acclimation. While the above mentioned experiments clearly pointed to cell dehydration as the mode of damage, they still did not answer the question whether extracellular ice formation could cause injury at least in the cold sensitive accession C24. Armstrong et al. (2015) referred to conflicting evidence in the literature and concluded that experimental conditions strongly influenced the results. While in the whole plant experiments reported here, no external ice nucleation occurred, most experiments using detached leaves apply external nucleation to assure synchronous freezing. In an experiment with external ice nucleation applied to detached leaves, tissue freezing commenced at − 0.8 to − 1.1 °C (Fig. 5), while ­LT50 values ranged between − 3.5 and − 7 °C (Fig. 6), thus clearly demonstrating that extracellular ice formation by itself is not the reason for tissue damage even in the cold sensitive C24. However, Reyes-Diaz et al. (2006) as well as Armstrong et al. (2015) reported conformity of INT, measured without external ice nucleation, and ­LT50 values determined in the presence of external ice nucleation. The same was observed here for C24 NA. As put forward by Armstrong et al. (2015), at least the southern European accessions likely behave as freeze avoiders when not externally nucleated and may under these conditions be unable to control extracellular freezing, when temperature falls below INT.

Control of the freezing process Pearce (2001) denoted freezing-induced cell dehydration as “the single most important cause of freezing damage”

13

Planta

exerted by phase changes in cellular membranes with the plasma membrane, the tonoplast, and also the thylakoid being important targets. Thus, freezing tolerance must involve avoidance of intracellular ice formation as well as the attenuation of cell dehydration by reducing the proportion of freezable tissue water (Vertucci and Stushnoff 1992). According to Sakai and Larcher (1987), a decrease in water content to 60–70% of saturation depresses the freezing temperature by 1–2 Kelvin and would thus be a potent means to prevent intracellular ice formation. However, to avoid dehydration, the proportion of freezable water must be reduced. In winter wheat, adjustment of water relations during cold acclimation was observed to occur in two steps with a reduction of total tissue water content triggered by temperatures around 0 °C and a change in its accessibility to freezing at later stages of acclimation (Yoshida et al. 1997). Quantifying the energy of exotherms in leaf samples of the different Arabidopsis accessions revealed a lower proportion of freezable water in Rsch AC compared to C24 AC, while Col-0 AC behaved intermediate. Pairwise comparisons of NA and AC samples revealed a significant reduction of freezable water in Rsch and Col-0 (Student’s t test, P < 0.05), but not in C24. Cryoscopic determination of osmolality of tissue sap showed that during cold acclimation Col-0 and to an even greater extent, Rsch increased solute content and thus the proportion of bound water that is not accessible to freezing. This clearly reveals control of the freezing process in the acclimated cold tolerant accessions. Similar mechanisms have been observed in freezing tolerant animals like the Canadian wood frog Rana sylvatica (Costanzo et al. 2013). Here, glucose and urea are used to control extracellular freezing and prevent osmotic stress. It is widely known that accumulation of soluble sugars correlates with freezing tolerance in Arabidopsis (Zuther et al. 2012), although especially for hexoses no mode of protection of cellular components is known (Crowe et al. 1997). We propose that modification of the physical state of water to reduce the proportion of freezable tissue water could be the physiological role of hexose accumulation during cold acclimation.

Conclusion It is clearly shown that leaf tissue of all Arabidopsis thaliana accessions tested can survive a freeze–thaw cycle, when freezing is externally initiated at mild freezing temperatures. However, investigating whole plants revealed different strategies of low temperature survival. While C24 from the Iberian Peninsula behaved primarily as a super-cooler in non-acclimated conditions and shifted to freezing tolerance following cold acclimation, the Central European and Russian accessions Col-0 and Rsch were freezing tolerant in either condition. In the case of Col-0 NA, the duration of

13

the freezing event strongly influenced the extend of tissue damage, indicating that cell dehydration is likely to be the mechanism of tissue damage. Accordingly, a reduction in the proportion of freezable water, probably caused by accumulation of osmolytes, could be observed in freezing tolerant accessions. Author contribution statement  IIH conducted whole plant experiments, DTA, cryoscopy, analyzed data, and wrote the manuscript. MR conducted detached leaf experiments, analyzed data, and contributed to method development. AGH designed the study, developed freezing programs, and contributed to manuscript writing. All the authors read and approved the manuscript. Acknowledgements  This work was supported by a Grant from the Deutsche Forschungsgemeinschaft (DFG), Grant Nr. HE-3087/10-2 to AGH. We would like to thank Diether Gotthardt and Marvin Müller for plant cultivation and the members of the Department of Plant Biotechnology for fruitful discussions.

References Arias NS, Bucci SJ, Scholz FG, Goldstein G (2015) Freezing avoidance by super-cooling in Olea europaea cultivars: the role of apoplastic water, solute content and cell wall rigidity. Plant, Cell Environ 38:2061–2070 Armstrong JJ, Takebayashi N, Sformo T, Wolf DE (2015) Cold tolerance in Arabidopsis kamchatica. Am J Bot 102:439–448 Block W (2002) Interactions of water, ice nucleators and desiccation in invertebrate cold survival. Eur J Entomol 99:259–266 Burke MJ, Gusta LV, Quamme HA, Weiser CJ, Li PH (1976) Freezing and injury in plants. Annu Rev Plant Physiol 27:507–528 Costanzo JP, do Amaral MCF, Rosendale AJ, Lee RE (2013) Hibernation physiology, freezing adaptation and extreme freeze tolerance in a northern population of the wood frog. J Exp Biol 216:3461–3473 Crowe J, Oliver A, Hoekstra F, Crowe L (1997) Stabilization of dry membranes by mixtures of hydroxyethyl starch and glucose—the role of vitrification. Cryobiology 35:20–30 de Mendiburu F (2016) agricolae: Statistical procedures for agricultural research. R package version 1.2-4. https​://CRAN.R-proje​ct.org/ packa​ge=agric​olae Des Marais DL, McKay JK, Richards JH, Sen S, Wayne T, Juenger TE (2012) Physiological genomics of response to soil drying in diverse Arabidopsis accessions. Plant Cell 24:893–914 Goldstein G, Rada F, Azocar A (1985) Cold hardiness and supercooling along an altitudinal gradient in andean giant rosette species. Oecologia 68:147–152 Gusta LV, Wisniewski M (2013) Understanding plant cold hardiness: an opinion. Physiol Plant 147:4–14 Gusta L, Wisniewski M, Nesbitt N, Gusta M (2004) The effect of water, sugars, and proteins on the pattern of ice nucleation and propagation in acclimated and nonacclimated canola leaves. Plant Physiol 135:1642–1653 Hacker J, Neuner G (2008) Ice propagation in dehardened alpine plant species studied by infrared differential thermal analysis (IDTA). Arc Antarct Alp Res 40:660–670

Planta Hacker J, Ladinig U, Wagner J, Neuner G (2011) Inflorescences of alpine cushion plants freeze autonomously and may survive subzero temperatures by supercooling. Plant Sci 180:149–156 Hincha D, Zuther E, Hellwege E, Heyer A (2002) Specific effects of fructo- and gluco-oligosaccharides in the preservation of liposomes during drying. Glycobiology 12:103–110 Hoermiller II, Naegele T, Augustin H, Stutz S, Weckwerth W, Heyer AG (2016) Subcellular reprogramming of metabolism during cold acclimation in Arabidopsis thaliana. Plant Cell Environ 40:602–610 Jacobsen AL, Pratt RB, Ewers FW, Davis SD (2007) Cavitation resistance among 26 chaparral species of Southern California. Ecol Monogr 77:99–115 Kasuga J, Arakawa K, Fujikawa S (2007) High accumulation of soluble sugars in deep supercooling Japanese white birch xylem parenchyma cells. New Phytol 174:569–579 Knaupp M, Mishra KB, Nedbal L, Heyer AG (2011) Evidence for a role of raffinose in stabilizing photosystem II during freeze-thaw cycles. Planta 234:477–486 Levitt J (1980) Responses of plants to environmental stresses. Academic Press, New York, London Lipp CC, Goldstein G, Meinzer FC, Niemczura W (1994) Freezing tolerance and avoidance in high-elevation Hawaiian plants. Plant Cell Environ 17:1035–1044 Nagao M, Arakawa K, Takezawa D, Fujikawa S (2008) Long- and short-term freezing induce different types of injury in Arabidopsis thaliana leaf cells. Planta 227:477–489 Nägele T, Heyer AG (2013) Approximating subcellular organisation of carbohydrate metabolism during cold acclimation in different natural accessions of Arabidopsis thaliana. New Phytol 198:777–787 Nannos N, Bersimis S, Georgakellos D (2013) Evaluating climate change in Greece through the insurance compensations of the rural production damages. Global Planet Change 102:51–66 Papagiannakis K, Lagouvardos K, Kotroni V, Papagiannakis G (2014) Agricultural losses related to frost events: use of the 850 hPa level temperature as an explanatory variable of the damage cost. Nat Hazards Earth Syst Sci 14:2375–2386 Pearce RS (2001) Plant freezing and damage. Ann Bot 87:417–424

R Core Team (2016) R: A language and environment for statistical computing. https​://www.R-proje​ct.org/ Reyes-Diaz M, Ulloa N, Zuniga-Feest A, Gutierrez A, Gidekel M, Alberdi M, Corcuera LJ, Bravo LA (2006) Arabidopsis thaliana avoids freezing by supercooling. J Exp Bot 57:3687–3696 Ritz C, Baty F, Streibig JC, Gerhard D (2015) Dose-response analysis using R. PLoS One 10:e0146021 Sakai A, Larcher W (1987) Frost survival of plants: responses and adaptation to freezing stress. Springer, Berlin, Heidelberg Schulze E-D, Beck E, Müller-Hohenstein K (2005) Plant ecology. Springer, Heidelberg Snyder RL, de Melo-Abreu JP (2005) Frost protection: fundamentals, practice, and economics. In: Environment and natural resources series, Food and Agriculture Organization of the United Nations, Rome, no. 10, vol 1 Vertucci CW, Stushnoff C (1992) The state of water in acclimating vegetative buds from Malus and Amelanchier and its relationship to winter hardiness. Physiol Plant 86:503–511 White GF, Haas JE (1975) Assessment of research on natural hazards. MIT Press, Cambridge, Mass Wilson P, Heneghan A, Haymet A (2003) Ice nucleation in nature: supercooling point (SCP) measurements and the role of heterogeneous nucleation. Cryobiology 46:88–98 Wisniewski M, Gusta L, Neuner G (2014) Adaptive mechanisms of freeze avoidance in plants: a brief update. Environ Exp Bot 99:133–140 Xie G, Timasheff SN (1997) The thermodynamic mechanism of protein stabilization by trehalose. Biophys Chem 64:25–43 Yoshida M, Abe J, Moriyama M, Shimokawa S, Nakamura Y (1997) Seasonal changes in the physical state of crown water associated with freezing tolerance in winter wheat. Physiol Plant 99:363–370 Zhen Y, Ungerer M (2008) Clinal variation in freezing tolerance among natural accessions of Arabidopsis thaliana. New Phytol 177:419–427 Zuther E, Schulz E, Childs LH, Hincha DK (2012) Clinal variation in the non-acclimated and cold-acclimated freezing tolerance of Arabidopsis thaliana accessions. Plant Cell Environ 35:1860–1878

13