Relationship between Freezing Tolerance and Leaf Growth during Acclimation in Winter Wheat

Abstract

Winter hardiness is influenced by many environmental factors, and freezing tolerance is among the main ones, rendering the phenotypic selection of winter wheat (Triticum aestivum L.) under field conditions a difficult task due to the irregular occurrence or absence of winter damage in field trials. Plant growth in response to low temperatures during the acclimation period might be used as an indirect approach to assess freezing tolerance. Thirteen winter wheat cultivars were investigated for autumn and spring growth and winter hardiness under field conditions for two growing seasons. Additionally, a precise and non-destructive technique was applied to study leaf growth at a high temporal resolution accompanied by a freezing tolerance test under laboratory and semi-field conditions. The results of the study revealed variations in thermal growth patterns among the 13 winter wheat cultivars. The cultivars with the lower base temperature (T_b) values, in particular ‘Lakaja DS’ and ‘Sedula DS’, grew slower and, thus, had a lower response to temperature increases (Slp_LER-T) than the fast-growing cultivars, such as ‘Simano” and ‘KWS Ferrum’, whose Slp_LER-T values were stronger and whose T_b values were higher. A correlation analysis of the investigated traits showed a clear association between leaf growth parameters and freezing tolerance, indicating a certain level of genetic adaptation to growth cessation under low temperatures, and which confirmed that these are important factors for explaining the freezing tolerance of different cultivars. The evaluated freezing tolerance (LT₃₀) showed a strong negative correlation (r = −0.82 ÷ −0.89, p = 0.01) to winter hardiness scores from the field experiment, supporting the essential contribution of growth rate patterns to winter hardiness. The findings provide novel information for the development of winter-hardy wheat cultivars that are adapted to the future environments.

1. Introduction

The observed and projected climate change trends show the increase in annual average temperatures, prolonged growing seasons and milder winters at temperate latitudes. The increase in precipitation, especially in winter, has also been observed [1]. Low temperatures and short growing seasons are one of the main limiting factors, thus climate change may offer new possibilities for crop production in Northern Europe countries (see [2], as reviewed in [3]). On the other hand, a prolonged growing season and higher autumn and winter temperatures challenge herbaceous crops, such as perennial forage grasses or biennial winter cereals, with new types of stresses. Multiple winter stresses, including frost, oxygen deficiency and snow mould, affect plants, and their ability to tolerate these effects determines their winter hardiness [4]. Though it is a complex trait, freezing temperatures have the highest impact on plants, and thus are often considered to be a measure of winter hardiness [5,6,7].

Freezing tolerance is largely determined by cold acclimation (or hardening), which develops in autumn as a response to low, non-freezing temperatures [8,9,10], causing alterations at the physiological, genomic, metabolomic and transcriptomic levels [11,12,13,14,15]. Plants start acquiring freezing tolerance at temperatures below +10 °C, and it accelerates when the temperature decreases to +5 °C or even 0 °C [16,17,18]. Proper acclimation increases the freezing tolerance level and, thus, the chances for the plant to minimize the damage incurred to the vegetative tissues, which leads to higher yield potential [19,20]. However, cold acclimation resulting in freezing tolerance might be difficult to achieve for the non-native herbaceous plant species, as they do not halt growth in autumn and enter winter at a larger size and later developmental stage. The main challenge for these plants in autumn is to reduce growth early enough, and later to completely cease their growth, so that they can successfully acclimate [21].

Leaf growth dynamics in monocots have been described in several studies, showing temperature to be the main driver [22,23,24]. However, the way that the plant senses and responds to temperature varies among species. Cold-adapted winter cereals and forage grasses maintain leaf growth via cell division under the temperatures close to freezing, or so-called base temperature [25]. Nonetheless, the temperature response of the leaf elongation rate (LER) is known to be linear under optimal conditions, though the LER at base temperatures is not so clear [22]. Some reports show an exponential response, while others advocate using an exponential function at the temperatures just above base temperatures and a linear function at optimal temperatures [26].

Autumn growth, namely, the leaf elongation rate, has been characterized on a physiological level in ×Festulolium [27], oilseed rape (Brasicca napus L.) [28] and perennial ryegrass (Lolium perenne L.) [21,29]. However, research on how the leaf growth of winter wheat (Triticum aestivum L.) responds to the autumn temperature changes at higher latitudes is scarce [30]. Despite the gained knowledge about temperature’s effect on wheat growth and phenology, little is known about the genotypic response pattern to varying temperature conditions during the acclimation phase in autumn. Furthermore, using non-destructive in situ methods is essential for gaining a deeper insight into the physiological mechanisms underlying growth in the field and the environmental factors affecting it, and combining these with a series analysis of the same organ would increase the resolution. Therefore, the measurement of leaf elongation under suboptimal and unfavourable growing conditions has been used as a valuable tool in physiological and agronomical studies [21,22].

Winter wheat (Triticum aestivum L.) is one of the most important crops in temperate areas. In these regions, frost injury during winter (winterkilling) and early spring can be particularly destructive. Wheat breeders are continuously searching for new cultivars with greater winter hardiness; however, limited success has been gained in developing freezing tolerant cultivars that would be more compatible with the changing climate [31,32]. This can be due, in part, to the complex tolerance mechanism, which involves various different freezing tolerance genes, as well as to the unpredictable nature of frost injury under field conditions where snow and sloped ground both create microclimates. The identification of mechanisms by which wheat copes with LT stress is crucial for breeding new frost-resistant cultivars and decreasing the risk of crop failure in cold areas. However, the combined effects of predicted climate change on the autumn growth and winter survival of winter wheat under low irradiance–low temperature conditions are still largely unexplored. The aim of the study was to investigate the relationship between leaf growth during the acclimation phase and low-temperature tolerance in winter wheat.

2. Materials and Methods

2.1. Plant Material, Experimental Set-Up and Phenotyping Procedures

Field experiments at the mini-plot scale were designed with three replications per genotype using 2.0 × 1.1 m size plots and conducted in the winter seasons of 2018/2019 and 2019/2020. The experiments with a set of thirteen winter wheat genotypes comprising 8 modern European cultivars (Capone, Galerist, Hanswin, KWS Ferrum, Nordkap, Proteus Simano and SW Magnifik) supplemented with 5 Lithuanian cultivars (Ada, Gaja DS, Lakaja DS, Sedula DS and vaxy-type Minija DS) were conducted at the LAMMC Institute of Agriculture in Akademija, Lithuania (55°23′ N, 23°57′ E). The soil of the experimental site was light loam Endocalcari-Epihypogleyic Cambisol, supplemented with 15 kg N ha⁻¹, 50 kg P₂O₅ ha⁻¹, 100 kg K₂O ha⁻¹ of fertilizers before sowing and 100 kg N ha⁻¹ in spring. The wheat was sown in mid-September.

The autumn growth was evaluated after the cessation of vegetation in the season of 2018/2019 and 2019/2020, measured on a 1–5-score scale, where 1 denotes the lowest growth. Winter hardiness, which includes freezing tolerance, disease and waterlogging resistance, was evaluated after the resumption of vegetation on a 1–9-score scale, where 1 denotes the lowest hardiness. The spring growth type was evaluated at the beginning of intensive growth in spring on a 1–5-score scale, where 1 denotes the lowest growth.

The microplot-scale experiments were conducted using the 0.25 × 0.25 m plots to monitor genotypic leaf growth patterns. The leaf elongation of 4 replications per variety was monitored in 5 consecutive weeks from 24 October 2019 to 1 December 2019 under field conditions. Leaf growth measurements and image analysis were conducted using the methods described by Yates et al. [24]. The leaf tip of the youngest growing leaf was attached with a hair pin to a string and kept taut while plastic beads were threaded onto the strings and placed on the growth array to provide artificial landmarks for image-based marker tracking (Supplementary Figure S1). The leaf elongation was registered in ten-minute intervals. Images of the growth array were taken with a LupusNET HD camera with a 2.1 pixel mm⁻¹ resolution (LUPUS-Electronics^® Gmbh, Landau, Germany) installed at an approximate distance of 1.2 m. Image sequences were analysed with the LLT software.

2.2. Meteorological Conditions

The environmental stratification of Europe defines Lithuania as a nemoral zone with a cool, temperate climate and a quite short growing season of 190–195 days [33]. Daily meteorological data (mean, minimal and maximal temperatures, precipitation and snow cover) over the period of 2018–2020 were recorded at the meteorological station in Akademija, Lithuania (55°23′ N, 23°57′ E). The winter period was considered to start from the first day when T_min ≤ −10 °C, cold days were considered when T_min is ≤ −15 °C, while the warm days were when T_mean ≥ 0 °C [34,35]. The autumn of 2018 was warm and dry, and the same temperature trends were observed for the autumn of 2019, though this season featured more precipitation (Figure 1). The winter of 2018/2019 was characterized as long, lasting for 127 d., with a snow insulation and few cold days. In contrast, the winter of 2019/2020 was extremely warm, with temperatures below 0 °C recorded just for 7 days, and no snow cover was observed.

2.3. Freezing Tolerance Test under Controlled Environment

Freezing tolerance tests under controlled conditions were performed on the same set of winter wheat as in the field experiments. Ten seeds of each variety were sown in 4 × 7 cell packs where each cell was filled with 125 mL of peat substrate. In total, six replications per variety were used for each target temperature and grown in the phytotron set at 18 °C, 180 μmol m⁻² s⁻¹ PAR, with a 12 h photoperiod and 80% relative air humidity until seedlings reached the 3-leaf stage. Afterwards, plantlets were moved to the PlantMaster growth chambers (CLF Plant Climatics GmbH, Germany) and acclimated under artificial conditions at 2 °C, 200 μmol m⁻² s⁻¹ PAR, with a 8 h photoperiod and 80% relative air humidity and cold acclimated for 14, 28, 42 and 56 days. Additionally, one set of plants was moved to the semi-field setup and acclimated for 56 days (5 October 2019 to 29 November 2019) under natural conditions prior to one of the freezing tests. The plants were irrigated weekly with a nutrient solution containing macro- and micronutrients. Just prior to freezing, plants were counted, and the substrate was saturated with cold water. Freezing test were carried out in a PE 2412 UY-LX freezing chamber (Angelantoni Industrie, Italy). The substrate temperature in each cell pack was measured at a crown depth every 2 min using temperature probes and recorded with a KD7 recorder (Lumel, Poland). The plants were kept at 2 °C for 1 h, whereupon the temperature was reduced from 2 °C to −6 °C over a 6 h period and held for 72 h. The temperature was then gradually lowered at a rate of 1.2 °C to a target temperature of −8 °C, −10 °C, −12 °C, −14 °C, −16 °C and −18 °C and held at this temperature for 11 h. Afterwards, trays were moved to the growth chamber set to 2 °C with an 8 h photoperiod for 24 h. After the freezing test and thawing at 2 °C, leaves were cut back, and plants were moved to the phytotron at 18 °C. Plant recovery was rated (alive vs. dead) after 3 weeks, and a lethal temperature, where 30 percent of plants are killed (LT₃₀), was calculated for each genotype.

2.4. Statistical Analysis

The statistical analysis was implemented in the open-source R statistical environment (version 4.0.2; [36]). The relative growth rate was calculated as described by [22] as a difference in leaf length at a given time interval:

$$\mathrm{RGR}=\frac{\Delta \mathrm{L}}{\Delta \mathrm{t}},$$

(1)

where ΔL is leaf extension in mm and Δt is the time period in h. The obtained relative growth rates of each measurement interval were set in relation to the environmental variable air temperature (T) which was collected from Dotnuva weather station at 15 min intervals, but which were summarized hourly and used for the calculations. The following linear regression model was employed to evaluate the LER in relation to temperature of each of cultivar:

$$\mathrm{LER}=\mathrm{a}+\mathrm{bT}+\mathsf{\epsilon },$$

(2)

where T is temperature, a describes the intercept (LER at 0 °C; Int_LER-T,) and b is the thermally corrected LER or slope (Slp_LER-T), showing an LER increase if T is increased by 1 °C and $\mathsf{\epsilon }$ is the residual error. Curves representing the development of single-leaf LER over time were extracted for individual plant. The coefficient of determination (R²) was calculated to determine the goodness of fit. The minimum temperature at which growth stops is denoted as T_b, and this was estimated for each individual plant based on Equation (2). The calculation was performed using R package “lmList” [37].

Analysis of variance and post hoc Tukey HSD tests were conducted using R package ‘agricolae’ [38], and relationships between the traits were evaluated by calculating the Pearson correlation coefficients for each trait pair. A generalized linear model using a binomial distribution to fit a curve was employed, while the MASS package [39] for R was used to calculate a fractional dosage value of 0.5 and to determine the LT₃₀ value for each cultivar.

3. Results

3.1. Growth during Acclimation and Winter Hardiness under Field Conditions

A great variation in autumn growth, winter hardiness and spring growth was observed for all winter wheat varieties grown under field conditions (Figure 2). The ANOVA showed the significant effect of the environment on autumn growth (p < 0.05) as well as on winter hardiness (p < 0.001) and spring growth (p < 0.05). In the whole set of winter wheat varieties, the autumn growth scores varied within a range of 1.0–5.0, with a mean score of 2.68 for 2018/19 and 3.25 for 2019/20. Significant differences (p < 0.001) for autumn growth rates among varieties were observed, where ‘Simano’ (mean of 4.73) and ‘Minija DS’ (mean of 4.30) were the most vigorous, while the least vigorous varieties were ‘Sedula DS’ and ‘Lakaja DS’, with a mean score of 1.20 and 1.65, respectively. The winter of 2019/20 was very mild, with just a few days of frost and no snow cover, resulting in a very good winter hardiness scores of all varieties, with a mean value of 8.64. Most of the varieties scored the highest values, except ‘Simano’, which was slightly damaged (the mean score was 6.3). In contrast, the winter of 2018/19 was typical to the winters of this latitude and lasted for 127 d, with a snow insulation and few cold days when the mean daily temperature dropped to −5 °C and below. These conditions revealed a higher differentiation in tolerance to winter stress when winter hardiness scores ranged from 1.0 to 6.0. The best performing genotypes were ‘Lakaja DS’ and ‘Sedula DS’, while ‘Hanswin’, ‘SW Magnifik’ and ‘Gaja DS’ were the least tolerant to winter stress. The spring growth scores for most of the genotypes were lower (p < 0.05) in the 2018/19 season than in the 2019/20, with a mean score of 3.59 and 3.97, respectively. The spring growth of Hanswin was the highest in both seasons, with a mean score of 4.83, while the plants of Lakaja DS exhibited the poorest spring growth, and their mean score was 2.00 in 2018/19 and 3.16 in 2019/20.

3.2. Leaf Elongation Measurements

The leaf elongation of thirteen winter wheat cultivars was monitored in 10 min interval during the hardening phase in the autumn of 2019. Over a period of 4 weeks, approximately 4300 time series measurements were made for each of the genotypes in three to four replicates to reveal a genotypic response to temperature. As is represented in Figure 3a, the leaf elongation rate (LER) strongly fluctuated, obvious following daily temperature. Plotting LER against the thermal calendar (Figure 3) instead of calendar time revealed more even leaf elongation, nevertheless having a large variation among the cultivars studied.

At the temperatures below 3 °C, the cultivars grew less than 0.20 mm h⁻¹, while at the range of 3 to 5 °C, the leaf growth of the cultivars in response to temperature started to differ and, for some of the cultivars, reached up to 0.40 mm h⁻¹. For the days when temperatures rose above ~5 °C, the LER increased rapidly and the genotypic response to temperature was even more clearly pronounced. As the leaves of wheat cultivars grew exponentially, we extracted linear fits of all genotypes studied by relating the LER of the different cultivars to temperature. The analysis defined the genotypic differences among cultivars for the growth rate response to temperature and distinguished them as slow-, moderate- and fast-growing cultivars (Figure 3b,

Table 1. Origin of cultivars and mean intercept and slope of the response of leaf elongation rate (LER) to temperature (Int_LER-T, Slp_LER-T) as well as derived base temperature (T_b) and coefficient of determination R².

Cultivar	Country of Origin	IntLER-T	SlpLER-T (mm.h−1. °C)	Tb (°C)	R2
Ada	Lithuania	−0.016	0.042	0.4	0.695
Capone	Germany	−0.051	0.058	0.9	0.696
KWS Ferrum	Germany	−0.054	0.063	0.9	0.787
Gaja DS	Lithuania	−0.028	0.052	0.5	0.438
Galerist	Germany	−0.021	0.046	0.5	0.754
Hanswin	Switzerland	−0.064	0.061	1.1	0.715
Lakaja DS	Lithuania	−0.007	0.037	0.2	0.748
SW Magnifik	Sweden	−0.026	0.045	0.6	0.765
Minija DS	Lithuania	−0.053	0.061	0.9	0.768
Nordkap	Germany	−0.040	0.051	0.8	0.642
Proteus	France	−0.024	0.048	0.5	0.674
Sedula DS	Lithuania	−0.017	0.036	0.5	0.763
Simano	Switzerland	−0.068	0.060	1.1	0.750

).

‘KWS Ferrum’ was the most vigorous cultivar, growing on average 0.0633 mm.h⁻¹.°C, whereas cultivars ‘Sedula DS’ and ‘Lakaja DS’ grew almost twice as slow and showed the lowest vigour with a mean of 0.0363 and 0.0375 mm.h⁻¹. °C (Figure 4, Table 1). The mean base temperature T_b, the temperature at which plants start their growth, also differed among cultivars, and ranged from 0.2 °C for cultivar ‘Lakaja DS’ to 1.1 °C for cultivar ‘Simano’.

3.3. Freezing Tolerance Test

Thirteen winter wheat cultivars were evaluated for freezing tolerance, estimated as the LT₃₀ value, after acclimating plants under controlled and natural conditions. One set of studied cultivars was cold-acclimated for 56 days following the natural autumn season temperature fluctuations in the semi-field setup (Figure 1), while, at the same time, the second set was acclimated under controlled conditions at 2 °C. A strong correlation (r = 0.97, p < 0.01) was found between the LT₃₀ (temperature at which 30% of the plants are killed by freezing stress) of cultivars acclimated for 56 days at natural and controlled conditions.

Six cultivars, namely, ‘Gaja DS’, ‘Lakaja DS’, ‘SW Magnifik’, ‘Minija DS’, ‘Ada’ and ‘Sedula DS’, showed high freezing tolerance and formed one homogenous group with a mean temperature of LT₃₀ ranging from −14.20 °C to −15.77 °C for naturally acclimated plants and −14.37 °C to −15.32 °C for plants acclimated under controlled conditions (Figure 5). In contrast, cultivar ‘Simano‘ was the least tolerant, with estimated LT₃₀ temperatures of −8.38 °C after acclimation under controlled conditions and −7.54 °C at natural conditions. The rest of the cultivars were of intermediate tolerance and formed a cluster consisting of ‘Capone’, ‘Hanswin’, ‘KWS Ferrum’, ‘Proteus’, ‘Nordkap’ and ‘Galerist’ (Figure 4). A series of cold acclimation periods—14, 28, 42 and 56 days of acclimation at 2 °C—were conducted to determine the LT₃₀ of every cultivar at each acclimation period (Figure 5). Plants of all cultivars began to acclimate rapidly and already, after 14 days of acclimation, they exhibited tolerance to freezing stress, showing three different levels of acclimation rate. Cultivars ‘Simano’ and ‘KWS Ferrum’ were the least tolerant, while ‘Lakaja DS’, ‘SW Magnifik’ and ‘Sedula DS’ were the fastest ones to gain tolerance, and the rest of the cultivars were moderately tolerant. After an initial burst, the rate of freezing tolerance continued to increase for all cultivars except ‘Simano’, whose tolerance to low temperatures even declined and remained at approximately the same level despite the duration of the acclimation period. Moreover, different tolerance accumulation patterns of the cultivars were revealed. Cultivars with the fastest accumulation of freezing tolerance, namely, ‘Lakaja DS’, ‘SW Magnifik’ and ‘Sedula DS’, and a few cultivars from the group of moderate tolerance—‘Capone’, ‘Hanswin’, ‘KWS Ferrum’, ‘Proteus’ and ‘Nordkap’—reached their maximum LT₃₀ already after 28 days of acclimation at 2 °C. An acclimation period longer than 28 days resulted in losing their freezing tolerance. In contrast, cultivars ‘Gaja DS’, ‘Galerist’, ‘Minija DS’ and ‘Ada’ prolonged their freezing tolerance accumulation and reached their lowest LT₃₀ and, thus, the maximum freezing tolerance after acclimation for 42 days. Even though continuous acclimation for 56 days resulted in a tolerance decrease, cultivars remained clearly grouped according to their freezing tolerance level.

To evaluate the interrelations between the measured traits describing growth rate in autumn and tolerance to freezing, Pearson correlation coefficients were calculated for each trait pair (Figure 6). If not indicated otherwise, the reported correlations were significant (p < 0.01). A strong, negative correlation (r = −0.94) was observed between the intercept (Int_LER-T) and slope (Slp_LER-T) of the leaf growth in response to temperature, indicating that genotypes growing slower or showing no growth at the temperature around 0 °C demonstrated more vigorous growth when the temperatures rose above 5 °C. Slp_LER-T showed a strong, positive interrelation with T_b which, in turn, correlated with autumn growth under field conditions in both experimental years (AG_2018/19, AG_2019/20). Spring growth under field SG_2019/20 was highly correlated with T_b, but for SG_2018/19, correlations were not significant.

Strong positive correlations (0.76 < r < 0.92) were found among freezing tolerance scores after hardening plants for fourteen (FT₁₄), twenty-four (FT₂₈), forty-two (FT₄₂) and fifty-six (FT_56Cont) days under controlled conditions, as well as for freezing tolerance after acclimation for fifty-six days under natural conditions (FT_56Nat), suggesting the reproducibility of both acclimation methods (Figure 7). FT₁₄, FT₂₈ and FT₄₂ positively and strongly (0.71 < r <0.86) correlated with relative growth under field conditions in the autumn of both experimental years (AG_2018/19, AG_2019/20), but was negatively associated with Int_LER-T (−0.88 < r <−0.93). There were no significant relations between winter hardiness WH_2018/19 and FT in the first experimental year, while winter hardiness in the second year WH_2019/20 was negatively associated with FT₂₈, FT₄₂, FT_56Cont and FT_56Nat, confirming the results of the freezing tolerance test in the climate chambers and supporting a corroboration between the laboratory and field environments.

4. Discussion

In the future, winters are expected to be milder at higher latitudes, thus making the selection of winter wheat a difficult task due to the irregular occurrence or absence of winter damage in field trials. Thus, in keeping pace with the new challenges the changing climate brings, breeders are forced to search for new breeding tools and strategies. In this study, we aimed to monitor the leaf growth of winter wheat during cold acclimation and reveal the relationship between growth rate in autumn and freezing tolerance.

The precise, non-destructive and largely automated leaf growth evaluation approach used in the study enabled us to solve the problem of limited temporal resolution in the field experiment and allowed us to determine different growth responses. Many automated image-based platforms have been used to study plant growth on a single-plant level under stressed and optimal conditions [40,41,42,43]; however, most of these studies were conducted under controlled conditions and could not fully mimic complex environmental scenarios, such as fluctuations of temperature over the days or a combination of high light intensity with low temperatures. Furthermore, the obtained results often could hardly be translated into the field experiments under naturally fluctuating conditions and from a single pot-to-plot level. Studying the effects of changing environmental conditions on growth requires experiments to be reproducible for a given genotype and independent in time. This was accomplished in our study by measuring leaf elongation in a wide range of temperatures, fluctuating from +12 to −3 °C (Figure 2). Relatively strong R² values demonstrate the clear relationship of LER to the temporal course of temperature. The growth response to temperature, in other words thermal growth, is known to have a genotypic manner and was evidenced in monocots by measuring wheat canopy cover [32], stem elongation [44] and leaf elongation in wheat and ryegrass [22,43]. Additionally, in the previous study of Grieder et al. [32], the base temperature (T_b), at which plants start/stop their growth, was shown to be a useful parameter for describing the genetic variation in LER–temperature-based relations among different wheat genotypes, though Seefeldt et al. [45] reported no differences in T_b among spring wheat cultivars. The developmental stage at which measurements were taken and statistical methods are very important factors for analysing the data and, thus, giving different results. The growth response to the range of optimal temperatures is way better described using an Arrenius-type function [46], while a simple linear model was proven to have high statistical power for explaining wheat growth in relation to low range temperatures and early developmental stages in spring [22] and autumn [32], as well as in our study. The measuring period is also a very important factor, as short-term temperature fluctuations might result in different thermal growth patterns when compared to long-term responses [47]. The results of our study revealed 13 winter wheat cultivars exhibiting different thermal growth patterns and were in line with the previous findings by Grieder et al. [32]. The cultivars with lower T_b values, in particular ‘Lakaja DS’ and ‘Sedula DS’, grew slower and, thus, had lower Slp_LER-T values compared to the fast-growing cultivars, such as ‘Simano” and ‘KWS Ferrum’, whose response to temperature increases (Slp_LER-T) was stronger, and their T_b values were higher. These parameters might serve as a reliable, non-invasive approach for obtaining indirect information on plants’ freezing tolerance potential. A correlation analysis of the investigated traits showed a clear connection between leaf growth parameters, namely, Slp_LER-T and T_b values, and freezing tolerance after hardening plants for fourteen (FT₁₄), twenty-four (FT₂₈), forty-two (FT₄₂) and fifty-six (FT_56Cont) days under controlled conditions, as well as under natural conditions (FT_56Nat). This observation indicates a certain level of genetic adaptation to growth under low temperatures and was confirmed in our study to be an important factor for explaining the freezing tolerance of different cultivars.

Freezing tolerance is of pivotal importance for survival over winter and is one of the multitudes of physiological factors influencing plant winter hardiness. Winter hardiness is a complex trait regulated by multiple genes and interactions among different genetic systems, shaping plants’ adaptations to specific environments [48]. Previous studies have shown that genotypes exhibiting delayed transition from vegetative to reproductive stage at temperatures in the acclimation range were more tolerant to freezing, as the plants accumulated a higher content of water-soluble carbohydrates, which ensured better hardening [49]. However, these findings do not explain all variations among genotypes in terms of growth rate. The estimated growth parameters Slp_LER-T and T_b reveal the individual and precise growth patterns of the genotypes studied, even when the differences are not clear or hardly visible by other means. Furthermore, when linked to the freezing tolerance LT₃₀, they allow us to uncover the freezing tolerance potential of the genotypes at the early developmental stages and in the absence of the freezing conditions.

Testing for the winter hardiness of wheat cultivars generally includes field trials based on visual ratings, semi-controlled experiments in snow-out shelters [50] or laboratory experiments under controlled conditions in the climate chambers, allowing us to assess the lethal temperature at which 50% or 30% of the plants die due to freezing (LT₅₀ or LT₃₀; [51]). In this study, we have evaluated the LT₃₀ and found a strong negative correlation (r = −0.82 ÷ −0.89, p = 0.01) between freezing tolerance measured in a controlled environment and winter hardiness scored in the field experiment of 2019/20, while no correlation was found for the season of 2018/19. This might be explained by the damage induced by the snow mould (data not shown) and not the freezing temperatures, as the plants were covered by snow, resulting in a fungus-favourable microclimate. These results confirm the complex nature of the winter hardiness trait and stress the importance of selection for disease resistance along with the freezing. On the other hand, the mild winter of 2019/20, with the lowest temperature of −5 °C and no snow cover, affected several of the most freezing-sensitive cultivars, causing sufficient damage and, thus, resulting in a negative correlation between LT₃₀ and WH_2019/20.

The milder winters projected in the future require shifting the breeding focus to certain beneficial traits that would facilitate plant resilience for future climatic adversities. This study combines the evaluation of leaf growth during acclimation, employing a high temporal resolution phenotyping approach in a semi-field experiment, with freezing tolerance assessed under controlled environments. To the best of our knowledge, this is the first time that these traits have been linked for characterizing plant tolerance to extreme low temperatures as a consequence of its growth pattern during early developmental stages. Moreover, the leaf growth phenotyping method not only describes the growth pattern of the genotype, but also serves as a tool for dissecting a complex trait into fundamental components. The method also has strong practical applications, as it enables the identification of genotypes with lower T_b and restrained growth in response to rising temperatures Slp_LER-T, resulting in a superior freezing tolerance. The relationship between growth response, base temperature and the degree of cold acclimation deepens our understanding of complex plant × winter environment interaction, and the knowledge gained could supplement traditional breeding techniques, boosting the effectiveness of future plant breeding.