Cold stress is one of the environmental stress that limits the growth and geographical distribution of forest trees across latitudinal and altitudinal gradients (Alberdi & Corcuera, 1991). In temperate climates, tree survival frequently depends on withstanding extremely low temperatures both before and after winter hardening (Sakai & Larcher, 2012). During autumn, plants are exposed to both chilling and freezing stress (Gao et al., 2020). Chilling stress can be defined as occurring above 2°C and is less harmful than freezing stress for plants. Models of climate simulation showed that despite global warming trends, extreme events ‒ such as early and late season frosts ‒ are more frequent (IPCC, 2013) due to increasing autumn and spring air temperatures that can delay dormancy at the end of the growing season or advance bud break at the start (Gu et al., 2008). Consequently, forest trees and seedlings may be increasingly exposed to chilling and freezing stress due to rapid temperature shifts (Rixen et al., 2012). Cold hardening is an important mechanism for many tree species to withstand low temperatures and cope with changing environmental conditions (Hamilton et al., 2016; Zeps et al., 2017), particularly given predicted shifts in species' geographic ranges to higher altitudes and latitudes especially across the Mediterranean areas (Valavi et al., 2019). Moreover, late flushing species like the oaks are less freezing resistant than earlier flushing species in temperate environments (Vitra et al., 2017), and global warming increases of 1–4 °C are changing the phenology of trees, negatively impacting the development of cold hardiness and autumn freezing resistance (Pagter & Arora, 2013; Zeps et al., 2017) due to insufficient physiological transitions in the fall (Hamilton et al., 2016).

Cold hardening is a gradual process that occurs in autumn when tree species are exposed to low temperatures and declining photoperiods (Beck et al., 2004). Cold hardening processes and freezing tolerance are accompanied by a series of biochemical and physiological changes (Alberdi & Corcuera, 1991; Rixen et al., 2012) and low photosynthetic productivity that causes growth cessation (Rossi et al., 2008; Bauerle et al., 2012). For example, osmotic adjustment through the accumulation of total carbohydrates (Morin et al., 2007) and proline (Hare & Cress, 1997) can allow plants to avoid intracellular freezing in cold hardening tissues of forest species trees (Delauney & Verma, 1993). Sodium and potassium can contribute to the regulation of osmotic potential under environmental stress, including chilling and freezing (Bogeat-Triboulot & Lévy, 1998). Also, chlorophyll fluorescence as a non-invasive tool also is appreciated for the study of tree species hardening such as oak species under low temperatures (Cavender-Bares et al., 2005; Corcuera et al., 2005; Cavender-Bares, 2007). Another reliable physiological indicator for testing the freezing tolerance and cold hardiness of seedlings and trees is electrolyte leakage (EL) (Flint et al., 1967; Koehler et al., 2012). This method is used extensively and shows membrane permeability and degradation of membrane integrity by cold stress (Moshtaghi et al., 2009).

The Zagros forests located in the west of Iran cover an area of about 6 million hectares, representing 44% of the country’s forested area (Sagheb-Talebi et al., 2013). Zagros forests are particularly important for water supply, soil conservation, environmental change, and the socio-economical balance of the whole country. Oak species (Quercus brantii Lindl., Quercus infectoria Oliver and Quercus libani Oliver) are dominate in these forests. Also, these species are native to large areas of Mediterranean forests and they are widely distributed in western Asia, namely in Iraq, Syria, Lebanon and Turkey (Mirhashemi et al., 2023).

Seedlings of oak during the first years of their life are very sensitive to environmental stress such as extreme summer drought and low temperatures. Therefore, these unfavorable climatic conditions can limit the distribution of oak species, and also most reforestation projects with these oak species have poor field performance. Some prior research has been conducted on physiological and biochemical responses to low temperatures stress in European or American oak species such as Quercus robur L., Quercus pubescens Willd., Quercus ilex L., Quercus virginiana, Quercus oleoides (Cavender-Bares et al., 2000; 2005; Cavender-Bares, 2007; Morin et al., 2007; Andivia, 2012b; Koehler et al., 2012), but there is no information for these three oak species (Q. brantii, Q. infectoria and Q. libani) despite the wide distribution of them in Mediterranean forests. On the other hand, understanding the physiological responses to low temperatures and the mechanisms involved in cold acclimation of species originating from altitudinal clines is very limited and it is critical to predicting how species will respond to climate changes (Cheaib et al., 2012). This information may also allow the selection of cold-tolerant ecotypes or populations for reforestation.

Therefore, the objectives of our studies were 1) to determine how cold hardening alters physiological responses to low temperature in these three oak species and the traits that contribute to low-temperature resistance after hardening or cold acclimation, and 2) to predict early season freezing resistance in these three species.

All PSII traits were restricted by progressing hardening, but the effect of species was only significant for NPQ (Tables 1and 2). Quercus libani had the largest value of NPQ compared to the other two species (1.22 vs 1.04). We found a significant interaction effect between hardening and species for chlorophyll content (Table 1). Thus that, despite the general decline of chlorophyll content by progressing hardening in the studied oak species, the rate of reduction was different in different species. Q. libani responded to simulated hardening quickly, whereas in Q. brantii and Q. infectoria chlorophyll declined at a slower rate (Fig. 3A).

The results of the three-way ANOVA test of the other physiological traits showed that the main effect of species was a significant factor in explaining variation in all traits except for K contents of the stem (Table 3). Also, RWC, and proline displayed highly significant differences among cold temperature treatments (Table 3). On the other hand, the interaction between species, cold temperature, and hardening for all of them except for leaf K was significant. The significant interaction between species and cold temperatures of leaf EL showed that Q. infectoria had the highest value under the chilling condition (4°C), while leaf EL of the other two oak species showed a larger increase under freezing treatments. The increase of leaf EL under freezing temperature was less pronounced in Q. libani and this species showed the lowest leaf EL value under freezing temperature (Fig. 3B). Also, the interaction of cold temperature and hardening showed that stem EL had a lower value under freezing temperature of hardening 2 compared to hardening 1 (Fig. 3C). There were no significant differences between species for stem EL in the hardening 1 treatment, but Q. libani performed significantly better than Q. infectoria in the hardening 2 treatment due to the higher value of stem EL in Q. infectoria (Fig. 3D). Leaf RWC showed the lowest value in the hardening 1 treatment in Q. brantii and declined with progressive cold hardening in both Q. brantii and Q. libani, while leaf RWC was stable between the hardening 1 and hardening 2 treatments in Q. infectoria (Fig. 3E). Also, stem RWC declined during hardening in Q. infectoria and Q. libani but did not change in Q. brantii. However, the RWC values were lower in Q. libani in both hardening treatments (Fig. 3F). Conversely, biochemical traits like proline and glucose increased during hardening treatments except for glucose, which decreased in Q. brantii. There was no significant change for glucose and proline with hardening treatment in Q. infectoria (Figs. 3G, H).

Comparison of means between species revealed that the leaf K content of Q. libani was significantly higher than the other two species (Fig. 4A). Leaf proline content did not change significantly from 4°C to -20°C (1 hour) and then increased after two hours of freezing at -20°C (Fig. 4B). Also, the effect of cold hardening on all traits was significant except for leaf EL and glucose (Table 3). The comparison of hardening 1 and 2 treatments showed that leaf and stem K content increased in the hardening 2 treatment (Figs. 4C, D), but the response of species to cold hardening was different for the other physiological and biochemical traits (RWC, EL, glucose, and proline) (Table 3).

Irrespective of the cold hardening and the freezing treatments, significant correlations among all measured traits were analyzed for each species, individually. Most of the traits in Q. libani and Q. brantii were significantly correlated while in Q. infectoria only leaf EL and leaf RWC had a weak direct correlation (r = 0.2).

In Q. libani, leaf EL was positively correlated with both leaf and stem RWC and negatively correlated with F_v/F_m. Also, K content showed a positive correlation with F_v/F_m. Glucose and proline content of Q. libani were highly positively correlated with each other, while there was no significant correlation between glucose and proline in Q. brantii. Stem RWC of Q. libani was also negatively correlated with leaf and stem K content and F_v/F_m, but the correlation between leaf RWC and F_v/F_min Q. brantii was positive (Fig. 5).

Principal component analysis (PCA) showed that the first two components explained 45.68% of the total variance. An ANOVA and t-test indicated that these two PCs were significant in different treatments. So, there were significant differences among the species and cold stress temperatures by only PC1. While both PC1 and PC2 were significant among hardening 1 and 2 (Table 4). The PC1 was contributed by the positive coefficient of physiological traits like EL and RWC content and a negative coefficient of K content (scores > 0.50). Q. libani seedlings, chilling stress (4°C), and some extent progressive hardening were aligned on the negative position of PC1 in comparison to the other two species (Q. brantii and Q. infectoria), freezing temperature (-20°C) and hardening 1. High osmolyte concentration, such as glucose and proline, was another characteristic that identified the second PC (Table 4and Fig. 6).

In temperate regions, cold hardening processes are critical for resistance to cold and freezing stress in forest trees (Burke et al., 2003). In this study, the autumn hardening was stimulated by decreasing photoperiods and temperature for oak seedlings during two months as previously described in other studies (Beck et al., 2004; Hamilton et al., 2016; Homayounfar et al., 2019). During progressive hardening, a reduction in chlorophyll content was observed in the three studied oak species. This result is consistent with other studies that found the preliminary chlorophyll degradation is mostly triggered by the shortening of daylight and further reduction occurrs by decreasing temperature (Fréchette et al., 2016). At the beginning of the hardening period, the amount of leaf chlorophyll content in Q . brantii and Q. libani seedlings diminished more in comparison to Q. infectoria which could provide frost tolerance in these two species by preventing photo-degradation (Oberschelp et al., 2020). All chlorophyll fluorescence traits also progressively declined during hardening, and in contrast, NPQ increased. In many tree species a restriction in photochemical quenching of photosystem II during cold acclimation has been observed (Cavender-Bares et al., 2000; Zhou et al., 2017; Homayounfar et al., 2019) due to a decline in metabolic sink capacity mediated by short photoperiod (Hamilton et al., 2016) or loss of photochemical efficiency of PSII and inhibition of the electron transport as a consequence of low temperature (Cavender-Bares et al., 2005). Furthermore, there were no significant variations across oak species in any of the chlorophyll fluorescence traits except of NPQ. The greater NPQ in Q. libani in comparison to the other two species during hardening suggested that Q. libani, which occurs at higher altitudes, has more efficient mechanisms of photo-protective energy dissipation (Murchie & Lawson, 2013). Consistent with this result, higher NPQ under chilling temperatures was observed in Q. virginiana found in temperate climates compared to Q. oleoides, which occurs in tropical climates that lack a cold winter period (Cavender-Bares, 2007).

Also, there was inter-specific variation in physiological responses between studied oak species to freezing stress after hardening. Freezing temperature can impair the integrity of cell membranes, causing an increase in EL (Rihan et al., 2017). We observed increases in leaf and stem EL with more negative freezing temperatures, as has been reported by many studies (Azzarello et al., 2009; Homayounfar et al., 2017). Also, freezing stress duration time had an increasingly negative effect on seedling leaves from one hour to two hours and this result is not surprising since the previous study has shown the leaves are more sensitive tissue to cold stress (Shao et al., 2013). Furthermore, the results of ANOVA and PCA demonstrated that different species had different levels of susceptibility to freezing stress. For example, Q. libani seedlings, which are native to colder regions and higher elevations than other species, displayed a lower increase in leaf EL during freezing stress. These results suggest that Q. libani is more freezing tolerant than Q. brantii or Q. infectoria. Also, a smaller increase in stem EL under freezing stress was observed by progressing the hardening period. These results revealed that freezing resistance is enhanced upon hardening, as expected in temperate species (Koehler et al., 2012; Homayounfar et al., 2017; Oberschelp et al., 2020). Yet the response of different oak species to hardening treatments was different, and freezing resistance increased earlier in the Q. libani. Prior observations also showed a more positive effect of cold acclimation on freezing responses in tree species from climates where freezing occurs earlier and is more intense (Cavender-Bares, 2007; Homayounfar et al., 2017). There was an inverse relationship between leaf RWC, EL, and Fv/Fm. Water status is known to affect cold hardening and cold resistance (Worland, 1996). Indeed, the presence of a large amount of RWC under cold stress caused a cellular injury and was accompanied by a reduction of photosynthesis activity (Balamurugan et al., 2018). We also found that decreasing RWC was accompanied by increasing in K+ content in Q. libani and Q. brantii. Increasing K⁺content, especially in Q. libani, could protect seedlings from low temperatures by osmotic adjustment of leaf cells upon hardening (Hsiao & Laüchli, 1986), consistent with the findings of Gleeson et al. (2004) and Andivia et al. (2012a). In addition, an increasing K⁺, proline, and glucose content, as osmo-protectants or scavengers, accompanied by progress of cold hardening and freezing, has been found by other researchers (Meng et al., 2015). Among species, we found highly significant variation in the accumulation of glucose and proline upon hardening. Because of the higher proline and sugar content of Q. libani, the osmotic adjustment ability of this species was stronger than Q. brantii and Q. infectoria, which originated from milder climates at lower altitudes. Similar results have been reported previously in some other Quercus species (Morin et al., 2007; Fernández et al., 2008; Andivia et al., 2012b). A reverse pattern of glucose in Q. brantii during progressive hardening and a lower value of PC2, which is associated with osmolyte content, demonstrated that glucose was excessively consumed as energy under conditions of severe low-temperature stress (Meng et al., 2015). Q. brantii was still able to increase cell osmotic to some extent concentration by increasing proline content in comparison to Q. infectoria.

Variation in cold tolerance in three studied oak species corresponds well with the habitat of their origin and provides evidence in support of adaptive plasticity in cold and freezing tolerance (Cavender-Bares, 2007; Koehler et al., 2012). Q. libani, which originates from colder climates and higher altitudes has evolved a survival strategy of higher tolerance to freezing stress through earlier and more effective cold acclimation and hardening than other species. The accumulation of osmotic metabolites such as glucose, proline, and K⁺appears to play an important role in seasonal increases in freezing tolerance in Q. libani seedlings. Also, NPQ measurement due to vast variation between species under cold stress is in agreement with the results Hussain et al. (2022) and can be usefully applied as a non-destructive tool to foresters for assessing the development of dormancy status and predicting freezing tolerance in oak species and selecting the genotypes or mother trees of oak species most likely to persist in a changing environment. Moreover, based on PC1 results, measuring EL and RWC could be the next choice for future evaluation, as already reported by other researchers as important indicators of cold tolerance (Wu et al., 2012; Balamurugan et al., 2018).

It appears that Q. infectoriaand Q. brantii are susceptible to early frost based on physiological data. In reality, though, Q. brantii is expected to die at a higher rate than Q. infectoria because early frost danger is higher in lower altitude species with wider geographic ranges. Fortunately, the PCA bi-plot analysis and other studies (Alikhani et al., 2014) suggested a high intraspecific variation within Q. brantii. Therefore, artificial screening for frost tolerance based on physiological responses to cold and freezing stress in Q. brantiifamilies and ecotypes is important for the reforestation of Zagros forests.