Since the industrial revolution, global population growth and deforestation have promoted climate change. Global warming caused by abnormal weather is turning living habitats into deserts (Naumann et al. 2018). These environ-mental changes simultaneously cause abiotic stressful environments, and especially in the case of plants, they face great trials. Plants respond to physiological and genetic changes to survive abiotic stresses caused by adverse environments, such as high-intensity light, drought, heat, cold and salinity (Bray 1997). Among them, drought is the abiotic stress factor that most affects plant life. Drought stress aggravates the symptoms of abiotic stress, such as salt and cold, or may appear simultaneously (Cruz de Carvalho 2008). Plants have different drought tolerances depending on species or varieties, but if the intensity or duration is prolonged, it induces dehydration, causing fatal damage to growth, development and reproduction (Fang and Xiong 2015). A large number of studies have been reported on the effect of drought stress and adaptation of plants until recently. For example, there are changes in crop yield, growth, pigment synthesis, photosynthetic activity, membrane integrity, osmotic adjustment, stomatal opening, cell division and accumulation of reactive oxygen species (ROS) in plants (Sairam and Srivastava 2001; Shinozaki and Yamaguchi-Shinozaki 2007; Praba et al. 2009).

Populus is ecologically and economically important because it is a rapidly growing tree and has a good biomass that plays a good role in bioenergy production and carbon sequestration (Brunner et al. 2004). Populus is considered a model tree for woody plants, especially in the molecular biology of trees, because it has a complete genome se-quence and is capable of genetic manipulation, rapid growth, vegetative propagation and a short rotation cycle (Jansson and Douglas 2007). Therefore, poplar is an important material for the study of phenotype, metabolism and genetic changes caused by abiotic stresses (Marron et al. 2002). Populus davidiana Dode (Salicaceae) (P. davidiana) is a native species of the Korean Peninsula and is one of the most widely distributed species in Eastern Asia and East Russia. P. davidiana has strong tolerance to drought, cold, and barren soils and thus grows well on slopes, ridges and gullies (Hou et al. 2018). P. davidiana not only adapts well to the mountainous terrain of Korea but is also valuable in commercial timber production (Latva-Karjanmaa et al. 2006). Consequently, P. davidiana is a good parental candidate when trying new hybrid varieties with good adaptability (Noh et al. 1989). Populus alba L. (P. alba) is native to Eurasia, commonly called silver poplar, and is widely distributed in floodplain forests of central Europe and central Asia (Brundu et al. 2008). P. alba is differenti-ated from P. tremula in response to abiotic stresses (tole-rance of drought, wind, salinity, and high temperatures) (Li and Kakubari 2001; Beritognolo et al. 2007). It is not only an effective tree on windy shores due to its good root extension and good tolerance to salinity but also efficient for timber and pulp production due to its rapid growth (Paolucci et al. 2010).

To prepare for the increase in drought caused by climate change, it is very important to accurately understand poplar seedlings’ initial response to drought stress. Therefore, ex-ploring the natural differences in poplar species that differ phenotypically in drought tolerance may be useful in explor-ing the complex mechanisms of the response to drought stress. In our preliminary experiment, we tested drought stress treatment on several species and their poplar hybrids. Among them, we selected P. davidiana, which had the best phenotypic tolerance to drought stress, and P. alba × P. davidiana, which was most affected by drought stress. In this study, we intend to report the difference in physio-logical and molecular analyses of these two species in response to drought stress.

Growth conditions and drought treatment

Poplar (P. alba × P. davidiana and P. davidiana) seedl-ings were grown on top-soil and sand (3:1) mixture in pots containing the appropriate soil moisture in the greenhouse. The temperature of the greenhouse was maintained at 18-25℃. Three replicates were used in the experiments. For each replicate, leaves of 6 plants were pooled and used for analysis experiment. These poplar clones were used as solitary maternal plants. In May 2020, the tissues were cultured and transferred to pots six weeks later in a greenhouse. Seventeen weeks after tissue culture, healthy and uniform plantlets were randomly selected from each treatment group. Plants were watered to a soil moisture level of 45% one day before the drought treatment. The plants were then deprived of water for 6 days and observed during this time. Soil moisture was measured every other day using a moisture probe (ICT International Pty. Ltd., Armidale, NSW, Australia). The experiment was con-ducted for 18 weeks at the National Institute of Forest Science in Suwon, Korea (37°15′04″N, 136°57′59″E), under semicontrolled conditions. Poplar leaves were har-vested at the same time in the morning in September 2020. The upper and lower leaves were harvested and mixed at the same stage, and the samples used in all experiments were the same. Samples were stored at ‒80℃ in a cryotic refrigerator until the experiment.

Four-month-old poplars encountered drought treatment, in which the soil relative water content (RWC) was reduced from 45%. Control plants were kept under the same conditions, except that the soil RWC was maintained at 40%. Control and treated plants were grown in suitably sized pots, and each pot had a tray.

Measurement of chlorophyll fluorescence and content

The maximum efficiency of photosystem Ⅱ (PSⅡ) che-mistry (F_v/F_m) and the potential activity of PSⅡ were measured by the Kautsky induction method using a portable Handy Fluorcam (Photon System Instruments Ltd., Brno, Czech Republic) (Kautsky 1931). After the start of the experiment, six plants at similar stages were measured every 2 days. To induce chlorophyll fluorescence, leaves were adapted to the dark by blocking light for 15 minutes followed by irradiation with 1,500 mmol∙m^‒2∙s^‒1 for 5 seconds. The fluorescence variables F_o, F_m, F_v/F_o, and F_v/F_m were measured and analyzed.

Photosynthetic pigments were determined according to the method of Sibley et al. 1996. The absorbance of the supernatants was measured at 470 nm, 647 nm and 664 nm with a Biospectrometer (Eppendorf, Hamburg, Germany). The contents of photosynthetic pigments and their means were calculated as follows for each plant and treatment:

• Chlorophyll a = 12.7A₆₆₄ ‒ 2.79A₆₄₇.

• Chlorophyll b = 20.7A₆₄₇ ‒ 4.62A₆₆₄.

• Carotenoids = (1000A₄₇₀ ‒ 1.82Chl a ‒ 85.02Chl b)/198.

(A, absorbance; pigment concentration in mg/g fresh weight (FW)).

Extraction and measurement of soluble sugar

Glucose, fructose, and sucrose were extracted and analyzed following the protocol of (Lu and Sharkey 2004). The sugar concentrations were determined enzymatically using a method described by (Stitt et al. 1989) and a Biospectrometer (Eppendorf, Hamburg, Germany). The contents of soluble sugars were expressed as mmol g^‒1 FW. To analyze starch content, the resulting sediments from aqueous ethanol extractions were rinsed with distilled H₂O and then autoclaved for 3 hours in distilled H₂O. Subse-quently, the supernatant was enzymatically digested to glucose according to the method described by (Walters et al. 2004). The sugar concentrations were determined and expressed in the same way as above. Total soluble sugars were extracted from leaves by 80% ethanol by a modified method of (Irigoyen et al. 1992). Total soluble sugar content was determined at 620 nm by a Biospectrometer (Eppendorf, Hamburg, Germany) using glucose as the standard. The content of total soluble sugar was expressed as mg g^‒1 FW. The detailed experimental methods used in the above experiments are described by (Kim et al. 2021).

Measurement of MDA, proline, and hydrogen peroxide (H₂O₂)

Malondialdehyde (MDA) content was determined accord-ing to the method of (Heath and Packer 1968) and expressed as mmol g^‒1 FW. The absorbance of the supernatant was read at 532 nm using a Biospectrometer (Eppendorf, Hamburg, Germany). Proline was extracted from a sample of 0.5 g fresh leaves in 3% (w/v) aqueous sulfosalicylic acid and estimated using the ninhydrin reagent according to the method of (Bates et al. 1973). The absorbance was read at a wavelength of 520 nm. The proline concentration was determined using a calibration curve and expressed as mmol g^‒1 FW. Hydrogen peroxide (H₂O₂) was measured spectrophotometrically after reaction with KI. The absorbance was measured at 390 nm. The amount of hydrogen peroxide was calculated using a standard curve prepared with known concentrations of H₂O₂ and expressed as mmol g^‒1 FW (Alexieva et al. 2001). The detailed experimental methods used in the above experiments are described by (Kim et al. 2021).

Determination of soluble protein and antioxidant enzyme activities

Leaves (0.2 g) were ground in liquid nitrogen followed by homogenization in 1.5 mL of potassium phosphate buffer (0.1 M, pH 7.5, containing 0.5 mM EDTA) in superoxide dismutase (SOD) and catalase (CAT) to prepare the extract. The homogenates were centrifuged at 4℃ for 20 minutes at 13,000 × g. The supernatant was collected and used for the assays of enzymatic activities. All steps in the preparation of the enzyme extract were carried out at 4℃. The Bradford assay method and bovine serum albumin as a standard were used to estimate the concentration of soluble proteins (Bradford 1976). The absorbance at 595 nm was determined by a Biospectrometer (Eppendorf, Hamburg, Germany). A superoxide dismutase (SOD) assay kit (Sigma-Aldrich Crip, Mi, USA) was used to determine SOD activity in homogenized leaf samples. All procedures were performed according to the kit instructions. The optical density was measured immediately at 450 nm after preparation using an automated plate reader (SpectraMax M2, Molecular Devices, San Jose, CA, USA). An EZ-Catalase (CAT) assay kit (cat. DG-CAT400, Dogen, Seoul, Korea) was used to determine CAT activity in homogenized leaf samples. All procedures were performed according to the kit instructions. The optical density was measured immediately at 560 nm after preparation using an automated plate reader (SpectraMax M2, Molecular Devices, San Jose, CA, USA).

Gene expression analysis by qRT-PCR

Leaf samples were frozen in liquid nitrogen, ground and stored at ‒80℃. RNA from three biological replicate plants for each treatment was extracted separately for cDNA synthesis. cDNAs for qRT-PCR were each synthesized from all samples. Total RNA was isolated using a Ribospin^TM Plant (Geneall, Seoul, Korea). For real-time quantitative RT-PCR (qPCR) analysis, first-strand cDNA was synthesized from 1 mg of DNase-treated total RNA using RNA to cDNA EcoDry^TM Premix (TaKaRa, Shiga, Japan). All reactions were performed using IQ^TM SYBR Green Supermix (BIO-RAD, CA, USA) and carried out in a CFX96 Touch Real-Time PCR Detection System (BIO-RAD, CA, USA) according to the manufacturers’ instructions. The gene-specific primers used for qPCR are listed in Supplementary Table S1. The reaction cycle was 1 cycle of 95℃ for 3 minutes, followed by 40 cycles of 95℃ for 15 seconds and 58℃ for 30 seconds. Three independent biological replicates and three technical replicates for each biological replicate were run. Relative quantification was performed to calculate the expression levels of target genes in different treatments using the 2^‒ΔΔCt method (Livak and Schmittgen 2001). The expression levels of ACTIN and UBQ7 were used for the normalization of qPCR results (Pettengill et al. 2012; Zhang et al. 2014).

Statistical analysis

Each measurement had three biological replicates. Analyses were carried out using one-way ANOVA with multiple comparisons using Tukey’s HSD. P values < 0.05 were considered significant. Values are presented as the means with SD.

Effect of drought stress on growth rate, leaf chlorophyll fluorescence and content, and carbohydrates

We examined changes in the growth rate, photosynthetic rate, and photosynthetic pigment content in seedlings of two species of poplar with strong or weak drought tolerance under drought stress conditions. After 4 days of drought treatment, the soil moisture content decreased to less than 10% and decreased to less than 2% after 6 days (Fig. 1A, B). We observed the phenotypes of shoots and measured the physiological changes in poplar leaves. Although there were no statistically significant differences observed between the control and drought stress condi-tions, the height and diameter of the two poplars tended to decrease when grown under water deficit conditions (Fig. 1C, D). P. alba × P. davidiana had a more severe decrease in growth rate.

Figure 1. Representative phenotype of P. alba × P. davidiana and P. davidiana under water deficit conditions. (A) Morphological differences between P. alba × P. davidiana and P. davidiana under drought stress conditions. (B) Volumetric water content in the soil of the control- and drought-treated plant pots. The soil moisture was measured every 2 days for 6 days. The control plants were watered throughout the experiment. (C, D) The effect of drought stress treatment on the shoot growth and diameter of poplar seedlings. The values are the means ± SD (n = 6).

Chlorophyll fluorescence reflects damage to photo-synthetic systems or photoprotection-related effects under drought stress (Morales et al. 2008). The maximal photochemical efficiency of photosystem II (PSII) was measured in an F_v/F_m state, and both poplars significantly decreased after 6 days of drought treatment (Fig. 2). The F_v/F_m of P. alba × P. davidiana and P. davidiana reached the minimum value at 6 days, which was 12.9% lower than that on the first day (0 days) under drought and 5.7% lower than that at 0 days. The potential activity of PSII can be seen through the ratio of F_v/F_o, and the values of both poplars also significantly decreased after 6 days of drought treatment. The F_v/F_o of P. alba × P. davidiana and P. davidiana reached the minimum value at 6 days, which decreased by 27.9% compared with that at 0 days under drought treatment and decreased by 10.8% compared with that at 0 days. The potential activity of PSII (F_v/F_o) showed a similar pattern to that of F_v/F_m, but more dramatic changes were observed in P. alba × P. davidiana. There is a connection between chlorophyll content and photosynthetic activity, and chlorophyll levels are affected by water deficit conditions. The contents of chlorophyll a and chlorophyll b in both poplars significantly decreased after 4 days and 6 days compared with 0 days or 2 days (Fig. 3). In addition, the total chlorophyll content of the two poplars was also significantly reduced at 4 days and 6 days. In the case of carotenoids, only P. alba × P. davidiana showed a change compared with the 2-day treatment. In both P. alba × P. davidiana and P. davidiana, there was no significant change in the ratio of chlorophyll a/b under drought stress conditions. The levels of photosynthetic pigments were overall higher in P. davidiana than in P. alba × P. davidiana under water deficit conditions. Taken together, these changes in phenotypic and chlorophyll fluorescence and photosynthetic pigments indicated that P. alba × P. davidiana was under drought stress more than P. davidiana.

Figure 2. Changes in chlorophyll fluorescence under drought stress. (A) F_v/F_o represents the ratio of variable to potential activity of PSII (the number and size of active photosynthetic centers). (B) F_v/F_m represents the ratio of variable to maximal chlorophyll fluorescence (maximal PSⅡ photochemical efficiency). Error bars represent standard error. Different letters on lines indicate significant differences at P < 0.05 (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD).

Figure 3. Effects of drought stress treatment on photosynthetic pigments in poplars. (A) Chlorophyll a. (B) Chlorophyll b. (C) Total chlorophyll. (D) Carotenoids. (E) Chlorophyll a/b. (F) Total chlorophyll/carotenoids. Values are means ± SDs of three independent measurements. Different letters indicate significant differences, and “n.s” indicates not significant (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD, P < 0.05).

Soluble sugars play an important role in maintaining the osmotic balance of plants and act as signal regulators under drought stress (Wahid et al. 2007; Rosa et al. 2009; Sami et al. 2016). The major carbohydrates detected in the leaves of poplars were sucrose, fructose, and glucose. The amounts of glucose, fructose, sucrose, total soluble sugar, and starch in the poplars were measured to determine the influence on carbon partitioning under drought stress. Both poplars showed significant increases in glucose and fructose contents after 6 days of drought treatment; P. alba × P. davidiana increased by approximately 3.6-fold, and P. davidiana increased by approximately 2.6-fold (Fig. 4). In addition, total-soluble sugars in P. alba × P. davidiana exhibited a significant increase 6 days after drought treatment, while those in P. davidiana gradually and slightly increased. The starch content of both poplars rapidly decreased on the 6-day drought treatment, and it decreased by almost half compared with the peak. However, statistically meaningful differences were not observed in the sucrose of P. alba × P. davidiana. The variation pattern of glucose, fructose, sucrose, total soluble sugar, and starch along with drought stress was similar for both poplars, but the change in P. alba × P. davidiana was larger.

Figure 4. Effects of drought stress treatment on carbohydrate contents in poplars. (A) Glucose. (B) Fructose. (C) Sucrose. (D) Total soluble sugars. (E) Starch. Values are means ± SDs of three independent measurements. Different letters indicate significant differences, and “n.s” indicates not significant (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD, P < 0.05).

Effect of drought stress on lipid peroxidation, proline, H₂O₂, and antioxidant activities

We measured MDA content, which efficiently indicates cell membrane damage caused by drought stress and is a classic marker of lipid peroxidation. As drought treatment progressed, the MDA content of P. alba × P. davidiana exhibited slightly increasing trends and showed a signi-ficant increase after 6 days of treatment (Fig. 5A). After 6 days of drought stress, the MDA content increased approximately 1.2-fold compared with the MDA content at 0 days. In contrast, the MDA content in P. davidiana showed no statistically significant change.

Figure 5. Changes in MDA, H₂O₂, proline, SOD, CAT, and soluble protein levels in response to drought stress. (A) MDA. (B) H₂O₂. (C) Proline. (D) Soluble protein. (E) SOD. (F) CAT. Values are means ± SDs of three independent measurements. Different letters indicate significant differences, and “n.s” indicates not significant (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD, P < 0.05).

Given that drought induced H₂O₂ production, experiments were carried out to examine the change in the level of H₂O₂ in leaves under drought stress (Fig. 5B). The level of H₂O₂ in both poplars increased at 4 days and 6 days after drought treatment. Overall, it was observed that P. alba × P. davidiana had approximately twice the H₂O₂ content than P. davidiana.

Proline accumulation is a common adaptive response to drought stress in plants. The proline content in P. alba × P. davidiana rapidly increased starting 6 days from the beginning of the experiment, reaching values 2.4-fold higher than those at 0 days (Fig. 5C). The proline content in P. davidiana gradually increased little by little.

Soluble protein plays a critical role in maintaining the osmotic balance of plants under drought stress (Wahid et al. 2007; Sami et al. 2016), and this can be an indicator to identify plants that lack water (Bano et al. 2013). The concentration of soluble protein in P. alba × P. davidiana and P. davidiana decreased by approximately 33% and 20%, respectively, after 6 days of drought treatment (Fig. 5D). The response patterns were similar, but the concent-ration of overall soluble proteins was at least twice as high in P. davidiana than in P. alba × P. davidiana.

Antioxidant enzymes alleviate oxidative stress caused by drought stress (Sharma and Dubey 2005). Therefore, we measured the activity changes of the representative antioxidants SOD and CAT (Fig. 5E, F). The change in SOD activity in the two poplars showed a different trend in our study. SOD activity in P. alba × P. davidiana was significantly decreased after drought stress, while SOD activity in P. davidiana showed no significant change. The CAT activity of P. davidiana decreased sharply 6 days after drought treatment. In the case of P. alba × P. davidiana, the CAT activity continued to decrease very rapidly after drought treatment. Except for proline, it was confirmed that cell damage and antioxidant-related indicators were worse in P. alba × P. davidiana.

Comparison of the expression of drought marker genes under drought stress conditions

We compared the expression levels of the reported drought marker genes, including PtP5CS, PtSUS3, PtLTP3 and PtDREB8, at 6 days on both poplars with those at 0 days (Fig. 6). PtP5CS, PtSUS3, PtLTP3 and PtDREB8 encode proteins participating in proline synthesis, sucrose synthesis, lipid transfer and dehydration-responsive element binding (DREB) transcription factors, respectively, which have been reported to be upregulated by drought stress and to play pivotal roles in the drought stress response of poplars (Tang et al. 2013; Wang et al. 2014). In our study, increases in PtSUS3 and PtDREB8 were induced by drought treatment in both poplars, whereas PtP5CS and PtLTP3 were increased only in P. davidiana. These results showed that some drought stress-related genes respond more sensitively to earlier in P. davidiana than in P. alba × P. davidiana.

Figure 6. Quantitative RT–PCR of four drought marker genes. (A) PtP5CS. (B) PtSUS3. (C) PtDREB8. (D) PtLPT3. Transcript levels are normalized to ACTIN and UBQ7, and error bars represent standard error (n = 3). Different letters indicate significant differences, and “n.s” indicates not significant (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD, P < 0.05).

Comparison of the expression of electron transfer- and H₂O₂ production-related genes under drought stress conditions

To observe drought stress-induced transcriptional chan-ges, we analyzed the relative expression of eight genes, including four genes related to the electron transfer rate (PETA, PETB, PETM and ATPA) and four genes related to H₂O₂ production (GOX1, GOX2, GOX3 and PGLP1), through qRT-PCR (Figs. 7 and 8). PETA, PETB, PETM and ATPA encode cytochrome a, b, m and ATPase alpha subunits, respectively (Song et al. 2014a). These genes were differentially expressed under drought stress. Although PETA was slightly downregulated, upregulation of PETA, PETB, PETM and ATPA showed that ATP synthesis was induced. Both poplars were gradually downregulated for PETA and were upregulated for PETB on the 6-day drought treatment. In addition, the expression level of PETM in P. davidiana was upregulated, and the expression level of ATPA in P. alba × P. davidiana was upregulated at 6 days of treatment. In the case of H₂O₂ production-related genes, GOX1, GOX2, GOX3 and PGLP1 in both poplars showed a similar trend. In both poplars, PGLP1, GOX1 and GOX2 were upregulated, and GOX3 was downregulated. Notably, the increase in the expression level of PGLP1 in P. alba × P. davidiana was larger than that in P. davidiana. Overall, the qRT-PCR results of the two poplars showed similar changes in the expression levels of genes related to the electron transfer rate and H₂O₂ production under drought treatment, but ATPA and PGLP1 showed larger changes in P. alba × P. davidiana.

Figure 7. Quantitative RT-PCR of four candidate genes related to electron transfer. (A) PETA. (B) PETB. (C) PETM. (D) ATPA. Transcript levels are normalized to ACTIN and UBQ7, and error bars represent standard error (n = 3). Different letters indicate significant differences, and “n.s” indicates not significant (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD, P < 0.05).

Figure 8. Quantitative RT-PCR of four candidate genes related to H₂O₂ production. (A) PGLP1. (B) GOX1. (C) GOX2. (D) GOX3. Transcript levels are normalized to ACTIN and UBQ7, and error bars represent standard error (n = 3). Different letters indicate significant differences (a lower-case letter for P. alba × P. davidiana and a capital letter for P. davidiana based on ANOVA with Tukey’s HSD, P < 0.05).

As the initial growing environmental conditions of seedlings greatly affect growth, development and survival, we conducted an early response study to physiological, biochemical and genetic changes under water deficit conditions. In general, the optimum soil moisture content for plant growth is 20-60%. In the experiment, it was measured to be less than 20% after the 2-day drought stress treatment, and it was less than 10% on the 4-day treatment, so it is thought that the seedlings were affected by the water shortage (Fig. 1). In the preliminary experiment of several species and their hybrids, the phenotypically most stressed P. alba × P. davidiana and the least stressed P. davidiana were selected. On the 6-day drought stress treatment, the difference in wilting between the two species was clearly observed. P. alba × P. davidiana has a larger stature due to the stronger P. alba characteristics, so these differences may also have affected stress under drought conditions. Plant growth is affected by drought stress, and the response depends on the adaptation to the duration and severity of the water deficit state (Yang and Miao 2010). In this study, we found that P. alba × P. davidiana possessed a better height growth rate under good water conditions than did P. davidiana, whereas P. davidiana maintained a better height growth rate than did P. alba × P. davidiana under drought stress (Fig. 1). Under drought stress conditions, this decrease in growth rate should take into account the effect plant wilting to some extent. Nevertheless, this result suggests that P. alba × P. davidiana was more sensitive to drought stress than P. davidiana and that P. davidiana had stronger drought tolerance.

We explored the photochemical efficiency of the PSII response to drought stress in two poplar leaves through photosynthesis-related spectroscopy analysis. It is well known that water shortage is one of the main causes of the decline in PSII activity (Sagardoy et al. 2009). In F_v/F_o and F_v/F_m, the decreasing trend of both species was similar, but P. alba × P. davidiana decreased more than P. davidiana (Fig. 2). The significant declines in F_v/F_o in P. alba × P. davidiana and P. davidiana at 6 days after drought stress treatment. The parameter F_v/F_o means the ratio of photo-chemical and nonphotochemical de-excitation fluxes of excited chlorophyll, so these results indicated a change in the rate of electron transport from PSⅡ to the primary electron acceptors with the density and size (Lu and Zhang 2000). F_v/F_m represents the maximum yield of primary photochemistry and photosynthetic capability of the entire PSII. When the ratio of F_v/F_m is < 0.8, the PSII reaction core is inactivated or damaged. The decrease in F_v/F_m in P. alba × P. davidiana and P. davidiana reflects damage to the photosynthetic apparatus in response to drought stress.

Although it is greatly affected by the variety, duration, and stage of the plant, the analysis of the chlorophyll content of leaves is one of the most effective methods for measuring drought tolerance. In this study, there were significant decreases in chlorophyll a, chlorophyll b and total chlorophyll after drought treatment (Fig. 3). The decline in chlorophyll content under drought stress is a common feature that has been considered a typical symptom of oxidative stress and chlorophyll destruction (Smirnoff 1993). The chlorophyll content in leaves is affected by stress, and the chlorophyll content is correlated with photosynthetic activity (Anjum et al. 2011). These reports are also consistent with the similar trend and degree of reduction of chlorophyll content and photochemical efficiency in both poplars in our study (Figs. 2 and 3). Carotenoids in both poplars reduced slightly or retained their contents, although it has been reported that the capacity to scavenge singlet oxygen and lipid peroxy- radicals and to inhibit lipid peroxidation and superoxide generation under dehydrate force (Deltoro et al. 1998). Overall, the levels of chlorophyll a, chlorophyll b, total chlorophyll and carotenoids under drought stress conditions were higher in P. davidiana than in P. alba × P. davidiana. These results suggest that P. davidiana was less affected by drought stress than P. alba × P. davidiana.

Glucose and sucrose play a role in substrate as osmolytes, and fructose is related to the synthesis of secondary metabolites (Gupta and Kaur 2005). The levels of glucose, fructose and total soluble sugars significantly increased 6 days after drought treatment, except sucrose and starch (Fig. 4). In general, drought induced an increase in soluble sugars, which is consistent with our results (Rosa et al. 2009). There was a sudden drop in the level of starch in both poplars 6 days after drought treatment. Although many plants have been reported to have increased sugar contents during the degradation of starch under drought stress (Fischer and Höll 1991), the sucrose content in both poplars remained or increased slightly in this experiment. It was confirmed that P. alba × P. davidiana react more sensitively than P. davidiana to the change in soluble sugars with respect to drought stress.

In general, a decrease in the concentration of soluble protein in plants is a representative feature under drought stress, although protein may increase rapidly in an immediate response to stress (Fischer and Höll 1991). In this experiment, the concentrations of soluble proteins in both poplars slightly decreased (Fig. 5). Referring to the report that lowered levels of protein are a characteristic symptom of oxidative stress (Moran et al. 1994), it is thought that there is a correlation between the increase in H₂O₂, the decrease in SOD, and the decrease in CAT and the decrease in soluble protein in our experiment (Fig. 5B, E, F). The concentration of soluble protein was notably different between the two species, and the concentration of soluble proteins in P. davidiana was maintained more than as high as that in P. alba × P. davidiana. In some cases, it was reported that the accumulation of soluble proteins was enhanced in plants under drought stress and conferred drought tolerance (Chen and Plant 1999). We speculate that this intrinsic difference in the soluble protein con-centration of the two species may have affected the drought stress tolerance of the two species as an osmolyte. In addition, the decrease in soluble protein is correlated with the accumulation of free amino acids (Iqbal et al. 2011).

The accumulation of proline content in plants under drought stress is a well-known drought tolerance response, as proline acts as an osmolyte and radical scavenger (Yin et al. 2005). Both poplars showed a gradual increase in proline content up to the 6-day drought treatment but did not show a strong response to drought stress (Fig. 5). P5CS, Δ¹-pyrroline-5-carboxylate synthetase, is a gene encoding a rate-limiting enzyme for proline synthesis, which catalyzes the NADPH-dependent reduction of glutamate to g-glutamate-semialdehyde (GSA) (Liang et al. 2013). At the RNA level, the expression level of PtP5CS, a gene encoding the key enzyme for proline synthesis, was slightly downregulated in P. alba × P. davidiana and increased approximately 4-fold in P. davidiana at 6 days (Fig. 6). These results show that proline production has not yet been induced explosively even though the proline content slightly increased until 6 days of drought treatment and that P. davidiana reacts more sensitively at the RNA level.

The osmotic balance in plants is maintained via the accumulation of compatible osmoprotectants by stabilizing cellular membranes and maintaining turgor, such as soluble sugars, soluble protein, and free proline (Moran et al. 1994; Yang and Miao 2010; Sami et al. 2016). In the two poplars, the amounts of soluble sugars (glucose, fructose, and total soluble sugar) and free proline started to increase at 6 days, and it was confirmed that those in P. alba × P. davidiana increased relatively more than those in P. davidiana (Figs. 4 and 5). This result suggested that P. alba × P. davidiana possessed a better osmotic adjustment, although it had lower drought tolerance than P. davidiana. It can be inferred from the experimental results that proline and soluble sugars (glucose, fructose, sucrose and total soluble sugar) do not play a pivotal role in drought stress tolerance in P. alba × P. davidiana and P. davidiana. This is because, despite their role as functional osmolytes and antioxidants for osmotic stress and ROS detoxification responses, higher levels of accumulation were observed in drought-sensitive species.

ROS produced excessively due to stress disrupt the balance between antioxidants and ROS and cause oxidative damage to membrane lipids (Apel and Hirt 2004). The byproduct produced at this time is MDA, which is a representative lipid peroxidation marker (Møller et al. 2007). Although P. alba × P. davidiana showed a significantly more wilted phenotype as drought treatment progressed compared to P. davidiana, the difference in MDA level between the two species was not significant (Figs. 1 and 5). While P. davidiana had no statistically significant change, P. alba × P. davidiana had a gradual but definite increase in MDA levels at the end. This result also suggested that P. davidiana has better drought tolerance than P. alba × P. davidiana. This response of the MDA level is considered to be correlated with the result that P. alba × P. davidiana, which will be described later, main-tains a relatively high level at H₂O₂, and the antioxidants (SOD, CAT) rapidly decrease (Fig. 6).

Plants maintain the balance of ROS and antioxidants through sophisticated and complex control under normal conditions. Overproduction of ROS, such as hydrogen peroxide (H₂O₂), superoxide (O₂^⋅‒), hydroxyl radical (OH^⋅) and singlet oxygen (¹O₂^⋅), is enhanced under abiotic stresses, which can cause oxidative damage to nucleic acids, proteins, lipids, and carbohydrates (Hossain et al. 2015). Compared to other ROS, its relatively weak toxicity, long lifespan (~1 ms), and small size allow H₂O₂ to play a stress-response signaling role as a secondary messenger (Cruz de Carvalho 2008). Thus, an increase in H₂O₂ levels is also a measure of the stress the plant is under, along with MDA induced by oxidative damage. As a major scavenger of superoxide anion radicals, SOD catalyzes the dismutation of superoxide (O₂^⋅‒) to molecular oxygen and hydrogen peroxide (H₂O₂) (Cruz de Carvalho 2008). Subsequently, CAT, APX and other antioxidant enzymes scavenge H₂O₂. Among them, CAT has a lower affinity for H₂O₂ than APX and is located only in the peroxisome to perform limited functions. However, CAT is suitable as an indicator of drought tolerance because it only removes the overproduced H₂O₂ due to severe drought stress. Reports on SOD and CAT activities under drought stress are heterogeneous, with different responses depending on plant species, age and tolerance as well as the duration and intensity of stress treatment. (Türkan et al. 2005; Cruz de Carvalho 2008). In general, increased SOD activity provides oxidative stress tolerance (Pan et al. 2006). The level of H₂O₂ was increased on day 4 in both species, and CAT activity was also decreased at the same time (Fig. 5B, F). However, because P. alba × P. davidiana had a relatively high intrinsic H₂O₂ content and the decrease in CAT activity was also more severe, it would have been subjected to more oxidative stress. In addition, in P. alba × P. davidiana, the lack of superoxide treatment due to the rapid decrease in SOD activity may also have affected the oxidative stress (Fig. 5E). Collectively, these results are consistent with the trend of change in MDA levels (Fig. 5). We observed that P. alba × P. davidiana showed a sharper decrease in the levels of antioxidant enzymes (SOD, CAT) than P. davidiana as drought treatment progressed, although the intrinsic H₂O₂ concentration of P. alba × P. davidiana continued to be high (Fig. 5). There was a higher-level increase in lipid peroxidation (MDA) and H₂O₂ in P. alba × P. davidiana than in P. davidiana, suggesting a further decreased ability to modulate peroxide anion free radicals and H₂O₂ in P. alba × P. davidiana. The ability to maintain the activity of antioxidant enzymes to reduce cell damage under drought stress may be an important characteristic of P. davidiana. Therefore, P. davidiana has a better antioxidant capacity to counteract drought stress.

In our study, increases in PtSUS3 and PtDREB8 were induced by drought treatment in both poplars, whereas PtP5CS and PtLTP3 were increased only in P. davidiana (Fig. 6). In P. davidiana, the expression of PtP5CS was upregulated at 6 days of drought treatment, but it seems that it has not yet been reflected in proline synthesis. In addition, the upregulation of LPT3 at 4 days was very rapid and increased more than threefold, showing a very distinct response difference from P. alba × P. davidiana. LTP3 is directly regulated by MYB96, and it participates in the plant’s tolerance to drought and freezing stresses (Guo et al. 2013).

At the same time, chlorophyll fluorescence, chlorophyll content and physiological analysis indicated that the electron transfer rate significantly decreased and H₂O₂ significantly increased in the 6-day drought treatment. These results suggested that changes in electron transfer and H₂O₂ might be related to a reduction in photosynthesis under drought stress. Thus, we selected four genes (PETA, PETB, PETM and ATPA) associated with the electron transfer rate and four genes (PGLP1, GOX1, GOX2 and GOX3) associated with H₂O₂ production for gene expression analysis (Figs. 7 and 8). Genes related to H₂O₂ production in both poplars showed similar changes. Of the two poplars, the expression of PGLP1, especially in P. alba × P. davidiana, was significantly upregulated 6 days after drought treatment. It may alleviate photosynthetic inhibition due to the scavenging of 2-PG, which is known as a potent inhibitor of enzymes in photosynthetic carbon metabolism (Kelly et al. 1976). It is known that environ-mental stress can damage PSII (Song et al. 2014b). In these results, the expression of PETB was upregulated at 6 days, and PETB plays an important role in PSII repair and ATP synthesis (Song et al. 2014a). In addition, ATPA, which plays an important role in promoting the conversion of ADP to ATP, was also upregulated at 6 days. Including the above PGLP1 results, all showed that P. alba × P. davidiana was more upregulated than P. davidiana. It is thought that the upregulation of genes for maintaining ATP synthesis in P. alba × P. davidiana when chlorophyll fluo-rescence degradation became more severe. These results confirmed that both poplars recognized the photosynthesis degradation caused by drought stress at 6 days and res-ponded at the gene level. The GOX gene family plays an important role in converting glycolate to glyoxylate and in H₂O₂ production (Rojas and Mysore 2012). Although GOX3 of both poplars was slightly downregulated, PGLP1, GOX1 and GOX2 were upregulated, and antioxidant CAT activity was reduced, resulting in increased H₂O₂ pro-duction.

This paper presented a study focused on measuring several physiological, biochemical and molecular analyses of two poplar species with different drought tolerances. In conclusion, our study provided an extensive analysis of the biochemical and physiological early response characteristics of two poplar seedlings to drought stress. When poplars were exposed to progressive drought, changes appeared early and severely in height growth inhibition, reduction of chlorophyll fluorescence and photosynthetic pigments, and increment of MDA and H₂O₂ in P. alba × P. davidiana compared to P. davidiana. In contrast, P. davidiana showed less growth inhibition and wilting, and the reduction of soluble protein and enzymatic antioxidants such as SOD and CAT was relatively slow and less affected. In addition, the concentration of soluble protein was maintained at a higher level than that of P. alba × P. davidiana. Sur-prisingly, although the changes in proline and soluble sugar contents under drought stress were higher in P. alba × P. davidiana than in P. davidiana, they did not have a significant effect on drought tolerance. In addition, the changes in the genes for electron transport and H₂O₂ production do not appear to be fully activated in the early stages of drought stress. In this study, the strength of P. davidiana compared to P. alba × P. davidiana is its ability to maintain enzymatic and nonenzymatic antioxidants such as SOD, CAT and carotenoids and strong expression of the LTP3 gene. Our data suggest that the lower drought tolerance of P. alba × P. davidiana might be correlated with oxidative damage by diminishing antioxidant systems. The physiology, biochemistry and gene expression analyses of this study have expanded our understanding of the early response to drought stress and provide new clues for further studies.

This study was supported by National Institute of Forest Science, Republic of Korea (FG0402-2022-01).

The authors declare that they have no conflict of interest.