1. Introduction

Plants at high altitudes or alpine regions encounter freezing stress, particularly during winter. When the air temperature reaches −30 ℃ or lower, soil temperatures can drop as low as −10~−30 ℃ at the surface (4 cm) for 2–3 weeks (Gusta et al., 2006). Numerous studies have delved deeply into the survival strategies and cold tolerance of certain plant parts that overwinter above the snow, such as buds and seedlings (Kaya et al., 2020; Körner, 2012). However, little focus has been given to the survival strategies of seeds in response to low temperatures, mainly because seeds are often thought to maintain a warm micro-climate under snow cover (Williams, 2006). However, increased global warming in recent years has affected the duration and extent of snow cover, which may lead to direct exposure of seeds to extremely low temperatures in winter, further increasing survival risk (Jaganathan et al., 2015).

Many alpine seeds imbibe sufficient water and germinate in spring or summer when environmental temperatures rise and the snow melts (Marcante et al., 2012). Increasing moisture content will directly exacerbate frost damage in seeds at low temperatures (Jaganathan et al., 2020), as the ice nucleation formation within a seed can be extracellular or intracellular, with the latter generally being lethal (Maqbool et al., 2009). Due to the extreme climatic characteristics of the alpine region, especially the large difference between day and night temperatures, seeds can imbibe water from melting snow during the day and freeze at night (Körner, 2003), which poses a significant risk to their survival. Thus, the freezing tolerance of seeds of alpine species is a decisive factor in their survival and successful reproduction.

Jaganathan et al. (2020) conducted a comprehensive review of studies focusing on seed survival at low temperatures, summarizing three primary survival strategies identified in the existing literature. The first two strategies involve freezing avoidance: one is preventing ice formation within tissues through the physical barrier provided by an impermeable seed coat, and the other is the supercooling of seed tissues. Supercooling is the process in which water within seeds is maintained in a liquid state at sub-zero temperature, thus avoiding ice formation. Another documented strategy is freezing desiccation, which refers to ice formation in the extracellular spaces, desiccating freezable water from intracellular structures. The approach effectively reduces intracellular ice formation, minimizing the risk of freezing damage and enhancing the freezing tolerance of seeds (Jaganathan et al., 2020; Maqbool et al., 2009). Seed tolerance to freezing damage has been shown to be highly dependent on an ecologically meaningful cooling rate of < 4 ℃/h. When fully imbibed Lactuca sativa L. seeds were cooled at < 4 ℃/h, the seeds survived temperatures below −18 ℃ through freezing desiccation (Keefe and Moore, 1981, 1983). Furthermore, during cold acclimation, some plant species can synthesize and accumulate specific protective substances, including proteins, amino acids, and sugars. These substances play a crucial role in maintaining the integrity of membranes and macromolecules under cold conditions (Körner, 2012).

Poaceae is an important plant family in alpine grassland and meadows (Li et al., 2012; Yang et al., 2014), and it has been the focus of various studies on cold tolerance and acclimation capabilities (Höglind et al., 2010; Jia et al., 2014). Previously, studies have been done on transcription and identified transcription and expression of differential genes in response to low temperatures or cold acclimation, along with the associated changes in soluble sugar content (Pompeiano et al., 2015; Rao et al., 2011; Schubert et al., 2019; Yu et al., 2015). However, the frost tolerance of Poaceae seeds remains relatively underexplored. Existing studies have primarily focused on how seed characteristics (Li et al., 2015; Mondoni et al., 2011; Wu and Du, 2007), growing environments (Byars et al., 2007; Li et al., 2020), and climate change (Hoyle et al., 2013; Mondoni et al., 2012; Wang et al., 2020; Yang et al., 2014) influence seed germination and seedling growth. Consequently, it has not been determined if alpine Poaceae seeds possess specialized adaptations to withstand extreme cold temperatures.

Past research on seed survival at low temperatures has predominantly focused on lowland and boreal flora, encompassing seeds of staple crops such as Lactuca sativa (Bourque and Wallner, 1982; Jaganathan et al., 2016; Junttila and Stushnoff, 1977; Keefe and Moore, 1981; Stushnoff and Junttila, 1978), Brassica napus L. (Gusta et al., 2006), Triticum aestivum L., Oryza sativa L. (Ishikawa and Sakai, 1978), and various Pinaceae species (Hawkins et al., 2003). Marcante et al. (2012) studied the seed frost resistance of 12 species from the Austrian Central Alps and found that dry seeds exhibited the highest frost resistance, followed by wet seeds post-imbibition, with germinated seeds and seedlings showing the least resistance. Reducing frost resistance during germination could contribute to seed mortality during summer frosts. The survival mechanisms of alpine seeds, particularly imbibed seeds, in cold climates remain poorly understood, and further investigation is necessary to address this knowledge gap. We investigated seeds of three species in the Poaceae, which represent important plant groups in the alpine region. Our study aims to analyze the freezing response and survival mechanisms of the fully imbibed seeds of these species under different cooling rates and thus determine the adaptation strategies of Poaceae to the cold environments of alpine regions.

2. Materials and methods 2.1. Seeds collection

Elymus dahuricus Turcz. ex Griseb., Festuca elata Keng ex E.B. Alexeev, and Lolium multiflorum Lam. seeds were provided by Qinghai Normal University. Seeds were collected from June to August 2021 at an altitude of 3500–3700 m a.s.l. from Xining, Qinghai Province, China (100°52′–101°54′E, 36°13′–37°28′N). The collected seeds were stored at 15% RH and 20 ℃ for 2 months before experimentation.

2.2. Determination imbibition time and germination

Three sets of 25 seeds were placed in Petri dishes lined with moist filter paper for each species to determine the imbibition time of seeds at different temperatures for each species. These dishes were then incubated at temperatures of 5 ℃, 15 ℃, and 25 ℃ under a cycle of 12 h of darkness followed by 12 h of light at an intensity of 60 μmol m⁻² s⁻¹. Seed germination was observed and recorded every 6 h, and the time of visual observation of the first seed radicle protruding was used as the time when seeds started to germinate. The imbibition time was inferred according to the germination time. Four hours before beginning germination was taken as the imbibition time when seeds had fully absorbed water. In this stage, seeds have fully absorbed water but have not yet germinated. After determining the imbibition time, the seeds were continuously incubated at the respective temperatures. Germination was recorded daily until no further germination was observed for a week, and the final germination percentage was calculated. Additionally, four replicates of 25 fully imbibed seeds for each species were oven-dried at 103 ℃ for 17 h to determine the moisture content of fully imbibed seeds. The difference in seed mass before and after drying was used to calculate the moisture content on a fresh weight basis (%) (ISTA, 2020).

2.3. Programmed cooling and seed survival test

Fully imbibed seeds (in a 15 ℃ chamber) for each species were dried on soft tissues to remove water from the seed surface and placed in a cryotube (2 ml, Corning). Based on the previous studies on cooling rate (Jaganathan et al., 2017; Keefe and Moore, 1981), the seeds were cooled to different temperatures (fast cooling: −5, −10, −15, −20, and −25 ℃; slow cooling: −5, −10, −15, −17, −19, −20, −21, −23, and −25 ℃) using a programmable freezer (Kryo series Ⅲ), with an initial temperature of 15 ℃ and cooling rates of fast (−1 ℃/min) and slow (−0.05 ℃/min). Three replicates of 25 fully absorbed seeds for each species were used in fast and slow programmed cooling. After cooling, the cryotubes containing the seeds were placed directly into a 37 ℃ water bath for 1 min of rewarming, and then the seeds were incubated in Petri dishes with moistened filter paper at 15 ℃ to test survival.

2.4. Differential scanning calorimetry analysis (DSC)

Differential Scanning Calorimetry (DSC, PerkinElmer DSC8500, Shelton, CT) quantified the latent heat produced during the phase transition of fully imbibed seeds for each species. The DSC temperature scale was calibrated before the measurements, adhering to the protocol outlined by Mori et al. (2012), using indium and cyclohexane as standards. For the analysis, an intact, fully imbibed seed was placed in a specialized aluminum dish (dimensions: 5.4 mm × 2.6 mm), and the sample was sealed to prevent leakage by pressing the sample with the aluminum cover. The DSC temperature program involved initially maintaining the sample at 15 ℃ for 1 min, followed by cooling to −25.0 ℃ at a rate of either 0.05 ℃/min (slow cooling) or 1.0 ℃/min (fast cooling), and then held at this temperature for an additional minute. Subsequently, the sample was warmed back to 15.0 ℃ at a rate of 1.0 ℃/min, after which the program concluded. The heat release during the cooling process was measured and calculated to determine the onset temperature of ice crystal formation and melting, in addition to quantifying the enthalpy changes.

2.5. Tetrazolium test (TTC)

The viability of the seeds at the end of the freezing procedure was determined in a TZ test. For this, the seeds for each species were put in a 0.5% solution of tetrazolium at 15 ℃ for 24 h. After the seeds were cut lengthwise with a scalpel, the seed staining was observed under the microscope (PXS6555, Shanghai JV Optical Technology Co.) and recorded with a camera (EOS 60D, Canon, Japan).

2.6. Hematoxylin-eosin (HE) stained paraffin section

To determine the changes in the internal anatomical structure of the seeds for each species after freezing and the location of ice crystal formation, fresh imbibed seeds without cooling and imbibed seeds with fast/slow cooling to −20 ℃ were put directly into FAA fixative to be stored at 4 ℃ at least 1 week for fully fixed. The samples were dehydrated with 75, 85, 90, 95, and 100% ethanol (v/v) and then embedded with paraffin. The embedded samples were cut into 4 μm sections. After dewaxing and hematoxylin-eosin staining (HE staining), the samples were observed under a light microscope (Eclipse E100, Nikon, Japan).

2.7. Ultrastructural studies with transmission electron microscope (TEM)

From each treatment group of three species (fresh imbibed seeds without cooling, seeds subjected to slow cooling to −20 ℃, and seeds exposed to fast cooling to −20 ℃), five seeds were randomly selected as samples. These samples were initially fixed in a 2.5% glutaraldehyde solution for 12 h at 4 ℃. Subsequently, they were rinsed with 0.1 M phosphate buffer (pH 7.0), then post-fixed in 1% osmium tetroxide for 1–2 h, followed by three additional washes in 0.1 M phosphate buffer. The samples underwent a series of dehydration steps using graded concentrations of acetone (30, 50, 70, 80, 90, 95, and 100% v/v), after which they were embedded in synthetic resin. Ultrathin sections of 70 nm were prepared from these embedded samples, and these sections were double-stained with uranyl acetate and lead citrate. The prepared samples were then analyzed using a Hitachi HT7700 transmission electron microscope.

2.8. Electrolyte leakage analysis

Relative electrical conductivity (EC) was measured on seeds before and after the programmed cooling. EC was determined using the conductivity meter (DDS-307A, Shanghai INESA Scientific Instruments Co., Ltd, China) according to methods described by Mavi et al. (2014). The experiments were carried out with at least three repeats for each species.

2.9. Data analysis

The dataset encompassing the final germination percentages of seeds incubated at different temperatures (5 ℃, 15 ℃, and 25 ℃) was subjected to statistical analysis using one-way Analysis of Variance (ANOVA). Furthermore, the survival of each species, when cooled to varying temperatures at a consistent cooling rate, was statistically compared employing one-way ANOVA coupled with Duncan's posthoc test. The Differential Scanning Calorimetry (DSC) data, comparing the same species across two different cooling rates, and the electrical conductivity measurements for each species before and after programmed cooling (control and −20 ℃) were analyzed using the independent-sample t-test. All data analyses were conducted utilizing SPSS v.21 and OriginPro 2021 software.

3. Results 3.1. Seed imbibition and germination

As the imbibition temperature increased, the germination speed increased (Fig. 1). Seed germination was not affected by variations in temperature. Regardless of the incubation temperatures (5 ℃, 15 ℃, or 25 ℃), the final germination percentage for all three species seeds uniformly exceeded 80%, with no significant differences noted within each species across the different temperature conditions (Fig. 1). Seeds exhibited accelerated germination when incubated at elevated temperatures, particularly at 25 ℃. Within 1–2 days of incubation at 25 ℃, seeds from all three species began to show germination signs. Specifically, Elymus dahuricus and Lolium multiflorum achieved germination of 6.67 ± 0.94% and 2.67 ± 0.47%, respectively, within the first 24 h (Fig. 1a and c). In contrast, Festuca elata seeds demonstrated a slightly delayed germination at 25 ℃, reaching a germination of 4.0 ± 0% after 84 h (Fig. 1b). When incubated at lower temperatures of 5 ℃ and 15 ℃, the germination process was slower than 25 ℃. F. elata seeds began germinating at 1.33 ± 0.47%, requiring 84 h at 15 ℃ and 156 h at 5 ℃. Among three incubation temperatures, seeds at 15 ℃ incubation temperature absorbed and germinated at a moderate rate, and also in line with the temperature under the natural conditions of spring and summer in the alpine. Therefore, the incubation temperature of 15 ℃ was chosen for all subsequent experiments, and the 4 h before the first seed germinated (E. dahuricus, 44 h; F. elata, 80 h; L. multiflorum, 58 h) was taken as the time when seeds have fully absorbed water. The moisture content of fully absorbed water seeds was: E. dahuricus, 47.59 ± 0.56%, F. elata, 50.87 ± 0.75%, and L. multiflorum, 45.02 ± 0.63%.

3.2. Differential scanning calorimetry analysis

Thermal transitions of an intact, fully imbibed seed subjected to different cooling rates are shown in Table 1. Two distinct freezing events were identified in Lolium multiflorum seeds cooled at 1.0 or 0.05 ℃/min (Fig. 2). The onset temperature of the first exotherm, classified as the high-temperature exotherm (HTE), was observed at −6.11 ± 3.24 ℃ for fast cooling and −5.87 ± 3.55 ℃ for slow cooling, showing no significant difference (p > 0.05, Table 1). The second exotherm, termed the low-temperature exotherm (LTE), was noted at −11.71 ± 2.81 ℃ for fast cooling and −11.83 ± 3.04 ℃ for slow cooling (p > 0.05, Table 1). Contrastingly, Elymus dahuricus and Festuca elata seeds exhibited variations in the number of freezing events and onset temperatures compared to L. multiflorum. Specifically, E. dahuricus seeds demonstrated two freezing events during fast cooling (HTE: −4.47 ± 1.61 ℃ and LTE: −18.44 ± 1.18 ℃), while three events (HTE: −3.04 ± 0.64 ℃, medium-temperature exotherm (MTE): −5.84 ± 2.30 ℃, and LTE: −16.86 ± 0.89 ℃) were observed during slow cooling (Table 1, Fig. 2). In contrast, F. elata seeds exhibited three freezing transitions during slow and two during fast cooling, as detailed in Table 1 and Fig. 2.

3.3. Seed survival after programmed cooling

Germination of fully imbibed seeds from three distinct species decreased with decreasing temperatures and their responses to freezing varied depending on the cooling rate (Fig. 3). Elymus dahuricus and Festuca elata seeds exhibited no significant alteration in germination when rapidly cooled to −15 ℃, as compared to the germination of fresh seeds. However, germination of E. dahuricus and F. elata seeds cooled to −20 ℃ decreased to 44.0% and 29.3%, respectively (Fig. 3a). In contrast, Lolium multiflorum seeds experienced a steep decline in germination, plummeting to 45.33% when rapidly cooled to −15 ℃ (Fig. 3a). During slow cooling, E. dahuricus seeds gradually reduced germination, halving the initial germination at temperatures down to −20 ℃ (Fig. 3b). Both F. elata and L. multiflorum seeds exhibited a high sensitivity to slow cooling. Their germination significantly decreased to around 50% at temperatures ranging from −5 ℃ to −10 ℃, markedly lower than the germination observed under fast cooling (Fig. 3).

3.4. Tetrazolium test (TTC)

Fully imbibed seeds without any cooling treatment, the seed embryo was stained by TTC to a distinct dark red color, representing high seed viability. In contrast, after fast or slow cooling to the temperature of −20 ℃ for each species, respectively, the vast majority of seed embryos could not be stained red, and a few seed embryos, such as Elymus dahuricus, were partially stained light red.

3.5. HE stained the paraffin section

The morphological and anatomical features of the seeds, including the pericarp, seed coat, endosperm (encompassing the aleurone layer and starchy layer), and various components of the embryo (such as the scutellum, coleoptile, plumule, radicle, and embryonic axis), were distinctly visualized using Hematoxylin and Eosin (HE) stained paraffin sections (Fig. 5a1). The embryo cells exhibited a uniform and tight arrangement in fresh seeds, maintaining an intact morphological structure. However, the seeds exhibited considerable morphological changes post-freezing, particularly following slow cooling. The arrangement of the embryo cells became more sparse, intercellular spaces noticeably increased, and the structural integrity of the seed was significantly compromised (Fig. 5a3, b3, c3). In contrast, seeds subjected to fast cooling showed comparatively minor structural changes in the embryo cells, as observed through HE staining. Specifically, Lolium multiflorum seeds that underwent fast cooling (Fig. 5c2) displayed minimal differences from fresh, imbibed seeds (Fig. 5c1). While the arrangement of the seed cells was slightly looser than in the fresh state following fast cooling, there was also a slight increase in the cell gap.

The structural integrity was well-maintained in fresh embryos, characterized by a distinct and well-preserved nucleus, intact cell walls, and a dense cytoplasm (Fig. 6a). Embryos subjected to fast cooling displayed organized nuclei and cellular structures (Fig. 6b). In contrast, embryos exposed to slow cooling exhibited notable alterations in cellular morphology, including cell disruption and cytoplasmic retraction (Fig. 6c).

3.6. Ultrastructural studies

In fresh Elymus dahuricus seeds, the embryo cells exhibited a clear and intact structure with tightly arranged cells. Notably, nuclei, cell walls, and plastids containing well-defined starch grains were discernible. A profusion of vesicles within the cytoplasm indicated a well-developed plasma membrane association (Fig. 7a1). Similarly, the embryo cells of fresh Festuca elata and Lolium multiflorum seeds were compactly arranged with minimal intercellular spaces. These cells displayed irregular cell walls alongside distinctly visible plastids and starch grains (Fig. 7b1, c1).

Post fast cooling, Elymus dahuricus seeds maintained their cellular integrity, albeit with a slight increase in intercellular gaps. The cells continued to exhibit plastids with starch grains and numerous vesicles, some of which contained observable contents (Fig. 7a2). Lolium multiflorum seeds also retained a comparable cellular structure to their fresh counterparts, demonstrating good ultrastructural preservation post fast cooling (Fig. 7c2). Conversely, the ultrastructure of F. elata embryo cells displayed progressive deterioration following fast cooling. Although nuclei, mitochondria, and starch grains were still present, an increased number of vacuoles, absent in the fresh seeds, was noted (Fig. 7b2).

After slow cooling, the ultrastructure of seed embryos showed disruption. Festuca elata and Lolium multiflorum embryos showed vacuolation and vacuolar fusion, and there was a large cavity in the cytoplasm (Fig. 7b3, c3). Cell contents leaked into the cavity formed by the severe plasmolysis, apparently of many mitochondria and lipid bodies. The cellular ultrastructure was severely disrupted, including disruption of the nuclear envelope, releasing its contents into the cytoplasm (Fig. 7c3). The less dense cytoplasm dispersed within the cell, indicating a fundamental loss of cell membrane structure. But, the embryo cells of Elymus dahuricus were less disrupted in comparison, with no visible cavity within the cells; only the space between the cell wall and the cell membrane was enlarged due to severe plasmolysis and also showed mitochondria and lipid bodies (Fig. 7a3).

3.7. Electrolyte leakage analysis

The electrolyte leakage in seeds consistently increased as the freezing temperature decreased (Fig. 8). The electrical conductivity (EC) values for all three species were invariably lower under slow cooling conditions than fast cooling, as depicted in Fig. 8. The EC of Elymus dahuricus and Lolium multiflorum seeds, when subjected to both fast and slow cooling down to −20 ℃, was significantly higher than the EC of fresh seeds (p < 0.05). In contrast, the EC of Festuca elata seeds, following programmed cooling, exhibited an increase, albeit not statistically significant (p > 0.05).

4. Discussion

Seed germination was relatively consistent across various incubation temperatures for each species (Fig. 1). The higher germination temperatures within the experimental range (5, 15, 25 ℃) led to faster germination. Our results align with those of Yang et al. (2014) on seeds of Stipa species (Poaceae), which have a broad temperature range for germination. To emulate the conditions prevalent in natural alpine meadow soils during spring and summer, an imbibition temperature of 15 ℃ was selected for subsequent experiments. The choice is in harmony with the temperature data reported by Wang et al. (2020). Furthermore, Jaganathan et al. (2016) determined that the low-temperature survival mechanism in hydrated Lactuca sativa seeds was independent of cold acclimation and contingent on the cooling rate. Gusta et al. (2006) reached a similar conclusion in Brassica napus seeds. Therefore, a single incubation temperature was deemed sufficient for our testing procedures.

In our study, most seed embryos failed to exhibit red staining post-cooling (Fig. 4). The observation suggests that a significant proportion of the seeds lost their viability after undergoing programmed cooling, a conclusion corroborated by the survival percentage data. Seed viability post-cooling was evaluated using three established methods: seed germination, 2, 3, 5-triphenyl tetrazolium chloride (TTC) staining, and electrical conductivity (EC) in our study. In particular, TTC staining proved highly effective in determining seed viability (Fig. 4). Unlike seed germination assessment, which necessitates extensive monitoring over a period exceeding 2 months, TTC staining requires only 24 h to yield results, making it a particularly expedient measure for evaluating stress resistance in Poaceae seeds. On the other hand, EC measurements indicate freezing-induced damage in some seeds. With the gradual decrease in temperature, the EC values showed an increasing trend, whereas the changes in EC values before and after the freezing of F. elata seeds were not significant (Fig. 8).

Differential Scanning Calorimetry (DSC) thermograms of intact, imbibed seeds at a constant cooling rate can distinctly delineate the number of freezing events, as evidenced by the exothermic peaks. According to Bourque and Wallner (1982), multiple freezing events typically indicate that different seed parts undergo water freezing independently. The onset temperature in these thermograms suggests the commencement of water freezing in distinct seed components, such as the endosperm and pericarp. However, it has been shown that only two exothermic peaks were recorded in lettuce seeds, which have no endosperm (Keefe and Moore, 1983). In such cases, typically only two exothermic peaks were recorded during programmed cooling in thermal analysis: the high-temperature exotherm (HTE) corresponding to the seed coat and the low-temperature exotherm (LTE) linked to the embryo, reflecting the progression of ice formation from the outer to the inner layers of the seed (Bourque and Wallner, 1982; Jaganathan et al., 2016; Stushnoff and Junttila, 1978). In contrast, our study observed three distinct exothermic peaks in two species, marking a difference from the patterns in non-endosperm seeds.

Combining the seed anatomical structure results in the present study and Martin (1946) work, it can be concluded that Poaceae seeds possess a rich endosperm structure. Therefore, it is reasonable to deduce that the three exothermic peaks observed during the slow cooling in Elymus dahuricus, and fast cooling in Festuca elata seeds correspond to water freezing in different regions: HTE (pericarp), MTE (endosperm), and LTE (embryo). During the gradual cooling of E. dahuricus seeds, the seed coat and endosperm exhibited freezing points at −3.04 ± 0.64 ℃ (HTE) and −5.84 ± 2.30 ℃ (MTE), respectively. However, these initial freezing events did not significantly impact seed viability (Table 1 and Fig. 2). A marked decrease in seed viability was only observed when the temperature dropped below approximately −16 ℃ (LTE), at which point the embryo froze (Fig. 3). Similarly, the fast cooling of F. elata seeds demonstrated three exothermic peaks in the cooling curve. Notably, a substantial decline in seed survival was recorded after fast cooling to temperatures ranging from −15 to −20 ℃, with LTE observed at −18.99 ± 1.68 ℃. Our results align with the assertion of (Keefe and Moore, 1981) that seeds cooled to temperatures below the HTE can germinate, but seeds fail to survive when cooled below the LTE threshold. In addition, four other cooling procedure curves displayed two exothermic peaks (Fig. 2). In most cases (such as Lolium multiflorum slow cooling, fast cooling, and E. dahuricus fast cooling), the temperature at LTE also correlated closely with seed germination.

However, unlike the three exothermic peaks, the two exothermic peaks in seeds suggest that the endosperm may have frozen simultaneously with the pericarp or embryo. The HTE enthalpy of Lolium multiflorum seeds was lower than the LTE enthalpy, presumably due to the water in the endosperm freezing at LTE temperatures along with the embryo. We isolated the endosperm from seeds of three species for DSC analysis; the enthalpy onset temperature is attributed to the specific freezing temperature of the endosperm. However, in all the freezing curves of the isolated endosperm, only a sequential freezing peak was observed around −5 to −9 ℃ (data not shown). The sequential freezing peak started with a steep but brief decrease in heat flow, followed by one broad exotherm with some small shoulders, like that shown for cooling apple seeds (Nguyen and Kacperska, 1990). The difficulty in utilizing isolated endosperm for exclusive analysis of its freezing temperature is compounded by factors influencing ice nucleation, including the presence of cuts in the sample or the incompleteness of seed samples. Frequently encountered in DSC analysis, these issues often stem from the structural incompleteness of some seeds or the extrusion of seeds from the sample dishes due to their large size, resulting in a singular sequential freezing peak in the seed freezing curve.

The analysis of sequential freezing curves reveals that not all seeds undergo independent phase transitions during freezing. Slowly cooled seeds of Festuca elata exhibited a consistent decline in germination across a broad temperature region, ranging from −5 ℃ to −17 ℃. Yet, no distinct temperature thresholds corresponding to a marked reduction in seed viability were identified. The observation is further supported by the considerable standard deviation in the onset temperature of freezing (Table 1). Additionally, the narrow range between the HTE and LTE temperatures also mirrored variations in survival. The phenomenon was consistent with findings from previous studies on apple seeds conducted by Nguyen and Kacperska (1990), indicating that not all seeds within a seed lot are capable of undergoing discontinuous freezing, which explains the decrease in seed survival percentage at each stage of −5 ℃, −10 ℃, and −15 ℃. According to Keefe and Moore (1981), ice formation in one part of the seed can induce nucleation in another, leading to sequential water freezing. The process is detected as HTE (Jaganathan et al., 2016). Consequently, determining the freezing temperature of the endosperm based solely on the presence of two exothermic peaks is not feasible. However, the occurrence of LTE is a reliable indicator of seed viability loss.

Our experiments examining the changes in embryo cell structure before and after freezing, utilizing paraffin sectioning and ultrastructural analysis, elucidated distinct survival mechanisms under varying cooling rates. In the context of fast cooling, seed viability largely hinged on the strategy of freezing avoidance through supercooling. The ability to postpone ice nucleation within the embryo enabled seeds to withstand low temperatures. Specifically, we observed that the ultrastructure of the embryo cells was predominantly well-preserved, except for Festuca elata embryo cells (Fig. 6a2, b2, c2). A comparative analysis with fresh cells revealed a marginally looser arrangement of seed cells and a slight increase in intercellular space (Fig. 5a2,b2,c2), suggesting that fast cooling rapidly caused intracellular water freezing. Due to the accelerated pace of this process, intracellular water lacked adequate time to migrate extracellularly, resulting in minimal or no significant freeze dehydration.

The EC values for all three species were higher with fast cooling than slow cooling (Fig. 8), suggesting fast cooling rates do not allow desiccation of intracellular water to extracellular space through the cell membrane, leading to ice crystal growth that punctured the cell and resulted in a substantial leakage of electrolytes. Pukacki and Juszczyk (2015) noted an increase in conductivity and a rise in membrane permeability in tissues with high moisture content when frozen, which they attributed to ice crystallization. In contrast, slow cooling provided cells sufficient time for intracellular water transfer, reducing electrolyte leakage compared to fast cooling.

Contrastingly, seeds subjected to slow cooling predominantly underwent freeze-desiccation, supported by loosely arranged embryo cells with notably enlarged intercellular spaces (Fig. 5c1-3). The structural characteristic indicated pronounced freeze desiccation in seeds subjected to slow cooling, leading to the migration of intracellular water from embryo cells to the extracellular space within the embryo tissue that mitigated the formation of harmful intracellular ice (Fig. 5a3, b3, c3). Migration of this intracellular water reduced irreversible damage attributable to intracellular ice formation. However, it is important to note that only Elymus dahuricus seeds demonstrated an effective adaptation to the water migration, thereby exhibiting a higher tolerance to lower temperatures during slow cooling compared to Festuca elata and Lolium multiflorum seeds. In cells of the E. dahuricus embryo, the space between the cell wall and cell membrane was enlarged by severe plasmolysis (Fig. 7a3). In contrast, F. elata and L. multiflorum seeds experienced a significant loss of vigor before LTE and displayed cellular ultrastructural changes indicative of sensitivity to water migration. These changes included cavity formation within cells and disruption of the cell membrane structure (Fig. 7b3, c3), potentially attributable to excessive solute damage. The cooling process primarily resulted in vacuolation and cytoplasm retraction, a phenomenon analogous to that observed in cryopreserved Amaryllis belladonna (Sershen et al., 2012) and Bactris gasipaes embryos (Heringer et al., 2013).

The inability of Festuca elata and Lolium multiflorum embryos to achieve discontinuous freezing, leading to simultaneous freezing with the endosperm, might precipitate premature ice formation and hinder their ability to supercool to lower temperatures. Undoubtedly, the cooling rate significantly influences the survival strategies of seeds. This is supported by studies on lettuce, where embryos are more prone to supercooling at fast cooling rates of more than 5 ℃ per hour. Conversely, at slower rates of ≥ 4 ℃ per hour, seeds tend to undergo freeze desiccation (Keefe and Moore, 1981, 1983).

Acknowledgments

We thank Prof. Ma Yonggui of Qinghai Normal University for the seeds. This work was supported by National Science Foundation of China (NSFC) [No. 32001119].

CRediT authorship contribution statement

Jiajin Li: Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Investigation, Formal analysis, Conceptualization. Ganesh K. Jaganathan: Writing – review & editing, Visualization, Supervision, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization. Xuemin Han: Visualization, Resources, Formal analysis. Baolin Liu: Supervision, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.