Freezing tolerance and tolerance to de-acclimation of European accessions of winter and facultative barley

Due to global warming, winter hardiness may seem to become less important for plant survival and yield. However, this is a superficial assumption, as probably only the most important factors locally affecting plant overwintering will change. For example, the frequency, degree, and length of extreme winter warming events may increase, leading to de-acclimation of plants. This study aimed to investigate existing variability in de-acclimation tolerance in Polish winter barley breeding materials and European winter and facultative barley cultivars, and to identify accessions with the highest and the lowest tolerance to de-acclimation by means of visual estimation of regrowth after freezing, measurements of electrolyte leakage and chlorophyll fluorescence, and LT50 assessment. The results of this study showed that freezing tolerance and tolerance to de-acclimation are independent traits, and even highly freezing tolerant plants can be susceptible to de-acclimation. Our results highlight the role of photosynthetic apparatus in de-acclimation, proving that chlorophyll fluorescence parameters, especially ET0/CS, can be useful indicators of tolerance to de-acclimation. This study also confirmed that although the mechanisms of response to de-acclimation seem to be common for susceptible barley accessions, the mechanisms of tolerance are different, and may be related to the accession’s origin.


Freezing tolerance assessment
After 3 weeks of cold acclimation, the plants were sampled for freezing tolerance assessment.Survival of the plants after freezing was assessed using three methods: regrowth after freezing (FT-R, similar to 16 ), measurement of transient parameters of chlorophyll fluorescence (Table 2) after freezing 17 , and electrolyte leakage (EL), as described in detail by 18 .Before testing, the leaves were cut about 1 cm above the soil.Plastic boxes (containing the rest of the plants after cutting) were put in a programmed freezer.Temperature decreased at a rate of 2 °C/h from 0 to -8 °C, -10 °C, and -12 °C, and after 12 h increased at the same rate.The cut leaves were placed into plastic bags, and put together in a programmed freezer with the batch of plastic boxes meant for freezing at -12 °C.Afterward, the plants were placed in an unheated glasshouse, and the leaves were used for chlorophyll fluorescence measurements.A visual estimation of FT-R was done after 10 days and after 21 days.Each accession in FT-R estimation was represented by 15 to 40 replicates (individual plants growing in four different boxes, 10 seeds/box, diverse germination rate).Induction of chlorophyll a fluorescence signal in the defrosted leaves was measured after 30 min of the leaf dark adaptation with the Handy PEA fluorimeter (Hansatech, Kings Lynn, UK).The measurements were performed in 10 replicates (10 leaves from different plants).The EL was measured using two about 2 cm-long fragments cut from the middle part of the second leaf of 10 plants from each cultivar/line and each freezing temperature.The leaf fragments were submerged in deionized water (which conductivity, W, was measured prior to the experiment), and gently shaken for 24 h.Then, the first measurement of conductivity of the water mixed with leaf sap was performed (EL1), followed by putting the leaf fragments in liquid nitrogen to destroy cell walls and membranes, and placing them back in the same liquid they were in before.After 24 h of shaking the destroyed leaf fragments, conductivity of the water mixed with leaf sap was measured again (EL2).EL was calculated as follows: EL = (EL1 -W)/(EL2 -W)100%.

De-acclimation tolerance assessment
Out of the 58 tested lines and cultivars, 20 that proved to be the most freezing tolerant at -10 °C and -12°C (8 from the group of the Polish breeding lines and 12 of the European cultivars, including one Polish), were further tested for their tolerance to de-acclimation.After 3-week acclimation to cold, the plants were subjected to de-acclimation (7 days at 12/5 °C, day/night), and sampled afterwards for freezing tolerance assessment.Plant survival after freezing was assessed in the same manner as described above, at the freezing temperature of -10 °C.

LT50 assessment
LT50, as the temperature killing 50% of the plants, was assessed for four lines/cultivars selected as tolerant to de-acclimation, and four selected as susceptible.The seeds were sown, and the plants were acclimated to cold and de-acclimated in the same manner as described above.Plant survival was tested for cold-acclimated and de-acclimated plants, using FT-R method.There were six freezing temperatures: -2, -4, -6, -8, -10, and -12 °C.The plants were frozen for two hours at these temperatures.The plant FT-R was assessed on a 0-9 scale 16 , where 0 referred to plants showing no signs of life, and 9 referred to those showing full regrowth.The score of 4.

Results
Freezing tolerance assessment Diversity in the freezing tolerance of the tested accessions was visible after freezing at all temperatures, with the highest mortality rate at − 12 °C (Figure S1).Freezing at − 8 °C poorly diversified the tested lines and cultivars in either series, while the temperature of − 12 °C seemed to be too severe (Figure S1).The temperature of -10 °C provided the optimal freezing conditions for the tests in this group of diverse barley breeding lines and cultivars (Fig. 1).
The results of FT-R, chlorophyll fluorescence, and EL measurements after freezing of the tested lines and cultivars showed considerable diversity in their freezing tolerance, especially at − 10 °C (Figs. 1, 2, 3) and − 12 °C (Figure S1b, S2b).Most of the non-Polish European cultivars displayed lower regrowth rate after freezing than the Polish breeding lines and cultivars (Fig. 1).
In the case of EL, the values measured after freezing at − 8 °C and − 10 °C were rather similar and low (below or around 10%) for ca 75% of the accessions (Figs. 2 and S2a).Greater diversity with regard to this parameter was observed only after freezing at -12 °C (Figure S2b).In most cases, the Polish breeding lines and cultivars performed better than the non-Polish European varieties with regard to EL after freezing at -12 °C (Figure S2b).
The most diversifying chlorophyll fluorescence parameters were ET 0 /CS and DI 0 /CS, while F V /F M proved to be the least informative as to the level of freezing tolerance among the tested accessions (Fig. 3).Three out of four displayed chlorophyll fluorescence parameters showed better performance of the non-Polish European varieties exposed to freezing (Fig. 3).Only the results of TR 0 /CS were in line with those for EL and FT-R after freezing at − 12 °C (Figure S1b and S2b, Fig. 3b).

De-acclimation tolerance assessment
Twenty selected lines and cultivars were also subjected to a week in de-acclimating conditions, preceding their freezing tolerance tests.As compared with the control (cold-acclimated) conditions, the de-acclimated accessions showed lower FT-R (Table 3).The biggest difference was observed in EL, which was several times higher in the de-acclimated plants (Table 3).The PCA showed that the parameters of chlorophyll fluorescence after de-acclimation were grouped opposite to the same parameters measured in the cold-acclimated plants (Fig. 4), proving that they reflected progress of de-acclimation, and were suitable for selecting barley accessions tolerant to de-acclimation.EL values for the cold-acclimated and de-acclimated plants were also grouped at a distance from each other, but not completely opposite (Fig. 4).FT-R, on the other hand, seemed to identify the plants the same way, regardless of their coldacclimation status, as the values of this parameter in de-acclimated and non de-acclimated plants were situated near each other on the PCA graph (Fig. 4).This observation was also confirmed by the correlation coefficient for the survival rate of the cold-acclimated and de-acclimated plants, which amounted to 0.616 (data not shown).
All of the Polish new breeding lines, as well as the only Polish cultivar among the tested European varieties, cv.'Gloria' , were grouped together with a visible distance from the rest of the studied objects in the PCA graph (Fig. 4).Those lines and cv.'Gloria' showed the best performance after de-acclimation followed by freezing in most of the measured parameters (Table 3).Also the worst overall performing cultivars, namely ' Aydanhanim' , ' Astartis' , ' Avcii 2002' , 'Carola' , 'Frost' , and 'Mellori' were grouped together in the PCA, suggesting a similar mechanism of response to de-acclimation (Fig. 4, Table 3).S1.
Table 3. Rankings (RNK) of freezing tolerance after de-acclimation measured as plant appearance FT-R (0 -9 score), % of EL after freezing at -10°C and selected chlorophyll fluorescence parameters measured after freezing of detached leaves.Accession numbers are explained in Table 1.Means are given together with confidence interval limits for P = 0.05 according to ANOVA analysis.Lower (-95%) and upper (+ 95%) limit is presented in the case when the highest frost resistance corresponds to a higher or lower value of the parameter, respectively.As a consequence means further down the ranking which are below or above the limit, respectively, did not differ statistically at P = 0.05 from the mean which this limit accompanies.Among three cultivars originating from Turkey, cvs.' Avci 2002' and ' Aydanhanim' were grouped together in the PCA, suggesting a similar response to de-acclimation, while cv.' Aday-4' was located in a different quarter of the graph (Fig. 4).These results are in line with FT-R and chlorophyll fluorescence values after de-acclimation of those cultivars, as cvs.' Aydanhanim' and ' Avci 2002' were among the worst, while cv.' Aday-4' was among the best objects in terms of FT-R and some fluorescence parameters (Table 3).Cv. ' Aday-4' was also the most distinct among all the 20 tested objects.It demonstrated the second best FT-R and the worst EL at the same time (Table 3), and in the PCA graph it was located the nearest to the group of cultivars with the poorest performance after de-acclimation (Fig. 4).Two German cultivars, 'Pamina' and 'Vincenta' , were located quite far from each other in the PCA graph, suggesting their partly different response to de-acclimation.According to the PCA, regarding its reaction to de-acclimation, cv.'Vincenta' seemed to have more in common with the Polish lines and cv.'Gloria' (Fig. 4).
Cvs. 'Pamina' and 'Bruker Stamm II' were the only cultivars with facultative growth habit tested for their de-acclimation tolerance in this study.They were both located in the same part of the PCA graph, but not close to each other (Fig. 4), suggesting a partially different mechanism of response to de-acclimation.Cv. 'Pamina' performed much better than cv.'Bruker Stamm II' in terms of FT-R and some chlorophyll fluorescence parameters, especially ET 0 /CS (Table 3).Most of the lines and cultivars that performed the best in terms of their FT-R after de-acclimation were also among those with the highest ET 0 /CS (Table 3).

LT50 assessment
LT50 for de-acclimated plants was significantly higher than for cold-acclimated ones (Fig. 5), which confirmed that all studied lines successfully de-acclimated during one week of de-acclimation.The results of LT50 assessment were in line with those of the physiological measurements and FT-R only in the case of two Polish breeding lines (Figs. 4 and 5), confirming their superior freezing and de-acclimation tolerance.On the other hand, the results of LT50 assessment after de-acclimation showed no significant differences between the cultivars selected as susceptible (' Astartis' , ' Aydanhanim' , 'Carola' , and 'Mellori'), and two of the accessions selected as tolerant (' Aday-4' and 'DS.1028/16').The susceptible accessions performed slightly worse than 'DS.1022/16' , which showed the lowest LT50 after de-acclimation.Cv. 'Pamina' , selected as tolerant to de-acclimation in previous analyses, showed the highest LT50 in the conditions of this experiment (Fig. 5).

Discussion
Although the problem of mid-winter de-acclimation is becoming increasingly crucial for overwintering of herbaceous plants, and future predictions show that its role will only grow in the next decades 1,2,19 , there is still very little known about the mechanisms of response to this stress, both on the phenotypic and genetic level.The existing studies focus mostly on dicotyledonous species [20][21][22][23] , especially Arabidopsis thaliana [24][25][26][27][28] .What is more, the studies up to now represent a whole spectrum of different de-acclimation times, temperatures, and day lengths, as well as different freezing conditions 20,21,[23][24][25][28][29][30][31] . All ofthe above make mid-winter de-acclimation studies, and comparisons between them, very challenging.Our study used a wide spectrum of physiological indices in order to investigate freezing tolerance in control conditions (cold-acclimated plants), and active de-acclimation tolerance defined here as freezing tolerance of de-acclimated plants.In the freezing tolerance assessment, the results of only one of the fluorescence parameters, namely TR 0 /CS, were in line with the results of EL and FT-R, and indicated the same accessions as the most and least tolerant, while the results for ET 0 /CS and DI 0 /CS were opposite.F V /F M , which is one of the most widely used chlorophyll fluorescence parameter for assessing freezing tolerance [32][33][34][35] , did not differentiate the accessions used in this study, and thus could not serve as a freezing tolerance indicator.Limited usefulness of F V /F M as an indicator of freezing damage was reported previously 36 .
On the other hand, F V /F M was among the parameters that were affected by active de-acclimation in our study, and showed diversity of the investigated barley accessions, which was consistent with the results of FT-R after de-acclimation.The chlorophyll fluorescence parameter that was the most reliable in pinpointing the objects the most and least tolerant to de-acclimation, was ET 0 /CS.These results confirm that ET 0 /CS is one of the most useful chlorophyll fluorescence parameters, usually the best in terms of correlating with survival after freezing in laboratory and field conditions 37,38 .De-acclimation affected mostly EL from the leaf tissues, when compared with the chlorophyll fluorescence parameters and FT-R.Dramatically increased EL in de-acclimated plants, as compared with the cold-acclimated ones, indicated that the changes in response to de-acclimation appear sooner in the leaves (but not directly in the photosynthetic apparatus) than in the other parts of the plant.This type of reaction was especially visible in one of the most de-acclimation tolerant cv.' Aday-4' , which displayed the highest EL among all the studied accessions.Apart from EL, also the chlorophyll fluorescence parameters related to the membrane integrity, namely DI 0 / CS and RC/CS M , showed the greatest changes after de-acclimation in most of the studied accessions.A possible reason for high survival rate in de-acclimated plants with apparent severe membrane damage in the frozen leaves, could be their ability to regenerate from the tillering nodes, as suggested previously 39,40 .
However, there are some accessions tolerant to de-acclimation, e.g.'DS.1022/16' and 'DS.2026/16' , in which de-acclimation seems to affect EL and FT-R to a similar degree.This may mean that the mechanisms of deacclimation tolerance are different in different barley accessions.A similar conclusion was drawn in our previous study 9 , where it was proposed that the response to de-acclimation in the susceptible accessions is probably similar, while the mechanisms of tolerance to de-acclimation are difficult to identify due to their diverse nature.Common mechanisms of response to de-acclimation in the susceptible barley accessions seem to be also confirmed by the PCA, in which the most susceptible accessions were grouped together despite their different origin.
The results of LT50 assessment for cold-acclimated and de-acclimated plants correlated with those for freezing and de-acclimation tolerance only to some extent, namely in the case of the Polish breeding lines selected as tolerant.A direct comparison of LT50 and both types of freezing tolerance assessment used in this study is of course impossible.That is mainly due to different methods used.The freezing tolerance assessments, both for the cold-acclimated and de-acclimated plants, included EL, chlorophyll a fluorescence measurements, and FT-R tests, and the overall freezing and de-acclimation tolerance assessment took all those results into account.On the other hand, in LT50 assessment, only FT-R test was performed.That is also an argument in favor of the hypothesis that the de-acclimation-related changes appear first in the leaves of a frozen plant.The observed differences might also be caused by a slightly different freezing protocol used in those experiments.In the case of LT50 assessment, the freezing time was shorter than in the preceding experiments, as we needed to test many more freezing temperatures in a single experiment.The differences in freezing tolerance of the cold-acclimated and de-acclimated plants, observed between the LT50 test and the other tests, confirmed the role of freezing exposure time in inducing plant damage [41][42][43] .
The Polish breeding lines and cultivars performed better regarding FT-R and EL both after cold acclimation and de-acclimation.However, the other European cultivars displayed better photosynthetic performance after freezing, as measured by chlorophyll fluorescence parameters in cold-acclimated state.After de-acclimation, the Polish accessions showed higher tolerance of the photosynthetic apparatus to direct influence of freezing, as evidenced by the values of chlorophyll fluorescence parameters.These results show extreme vulnerability of the photosynthetic apparatus to direct freezing in the de-acclimated state, which is in line with the previous suggestions on the role of the photosynthetic apparatus in de-acclimation signal perception 21 , as well as the role of the antioxidant system in the response to active de-acclimation 9 .The differences in overall performance and the fact that the Polish breeding lines and cv.'Gloria' form a distinct group in the PCA of the performance after de-acclimation, might indicate that the origin of barley accessions determines their reaction to de-acclimation.Different origin, as well as breeding and cultivation area, might have created an unintended selection pressure, of which the breeders were unaware.Thus, the genetic background of the response to mid-winter active deacclimation might be different in the cultivars of various descent.These alleged differences might be responsible for the tolerance mechanisms observed in this and our previous study 9 .
In conclusion, the differences in freezing tolerance and tolerance to de-acclimation of the tested accessions support the hypothesis that these two traits are determined separately, and even highly freezing tolerant plants can be susceptible to de-acclimation.Our results highlight the role of the photosynthetic apparatus in the deacclimation process, proving that the chlorophyll fluorescence parameters, especially ET 0 /CS, can be useful indicators of tolerance to de-acclimation, and thus help in the selection process.This study also confirmed our previous findings, that although the mechanisms of response to de-acclimation seem common for the susceptible barley accessions, the mechanisms of tolerance are different and may be related to the plant origin.

Figure 3 .
Figure 3. Chlorophyll fluorescence parameters measured after freezing at − 10 °C on leaves of 58 barley accessions (Polish breeding lines, marked in orange and European cultivars, marked blue) cold-acclimated for 3 weeks at 4/2 °C (day/night).Means and confidence intervals for P = 0.05 (a mean, which is in a range of error bar did not differ statistically at P = 0.05).Homogeneity groups (HSD test) are shown in TableS1.

Figure 4 .
Figure 4. Biplot showing PCA analysis results of variables measured in 20 barley accessions cold-acclimated for 3 weeks at 3 weeks at 4/2 °C, day/night (open, blue dots) and de-acclimated 7 days at 12/5 °C, day/night (closed, green dots) plants as well as the distribution of the accessions (triangles).Polish breeding lines and cultivar are indicated in orange and European cultivars in blue.

Figure 5 .
Figure 5.Freezing tolerance expressed as LT 50 (temperature which killed 50% of the plants) of eight barley accessions (Polish breeding lines and European cultivars) selected for their diverse tolerance to de-acclimation.Tolerant lines were marked with solid and susceptible with open bars.Freezing tolerance was studied after cold acclimation (3 weeks at 4/2 °C, day/night) and de-acclimation (7 days at 12/5 °C, day/night).50% of plants killed by freezing temperature corresponds to 4.5 score on 0-9 scale of FT-R assessment.Means are given together with 95% confidence limits according to dependent variable prediction in 'Multiple regression' module of Statistica 13 (Dell, Round Rock, TX).Means, which are below or above the vertical bar limit did not differ statistically at P = 0.05.

Table 1 .
Characteristics of plant material used in the study.

Table 2 .
Formulae and glossary of terms used by the OJIP-test in the present study (modified after Strasser et al. 2004).Subscript "0" indicates that the parameter refers to the onset of illumination.