The experiment was conducted under controlled conditions from March to June 2020 at the Fujian Institute of Tobacco Agricultural Sciences, Fujian, China. The seeds of the tobacco cultivar “Cuibiyihao” (Nicotiana tabacum L. cv. CB-1) were surface sterilized with 75% ethanol and 30% hydrogen peroxide (1:1) for 5 min. The seeds were then rinsed several times with sterilized water before being vernalized at 4°C for 12h in the dark (Forner et al., 2015; Anwar et al., 2018). In short, first, the vernalized seeds were sown in plastic trays, and seedlings were raised for 2 months under greenhouse conditions. Second, sixty healthy tobacco seedlings were selected and transplanted to hydroponic pots for adaptation to a nutrient solution environment at 25°C for 2 weeks. Third, 36 tobacco seedlings with uniform size and growth were selected and randomly divided into three equal parts and transferred to four different growth chambers, each with 12 hydroponic pots, and set the temperature of 8°C, 12°C, 16°C and 25°C respectively, because of frequent severe changes in temperature since last few years in the growing season of tobacco ( Figures 1B, C ). The black hydroponic pot with a lid, 12 cm long, 8 cm wide, and 10 cm high, with a volume of about 1000 mL, was used for the test. The black color was mainly used for shading, and the black lid had holes in the middle for the growth of plants.

The hydroponic pots of each growth chamber were divided into two sets (6 hydroponic pots for each). One set was treated with magnesium addition (+Mg) and four different temperatures, i.e., 8°C, 12°C, 16°C and 25°C, and named as +Mg8, +Mg12, +Mg16 and +Mg25, respectively. The other set without magnesium (–Mg) and the abovementioned temperatures are –Mg8, –Mg12, –Mg16 and –Mg25, respectively. It is important to investigate the effect of Mg because the main tobacco-producing areas of Fujian Province have an uneven distribution of exchangeable Mg in the soil, and some areas are severely deficient in Mg ( Figure 1A ). So, hydroponic pots of +Mg were supplemented with a modified Hoagland’s nutrient solution containing magnesium (pH 6.5-7.0; 995μM KNO₃, 931μM Ca(NO₃)₂·4H₂O, 397μM MgSO₄·7H₂O and 282μM KH₂PO₄. Micronutrients: 22.5μM Fe(II)-EDTA, 46.2μM H₃BO₄, 14.3μM MnCl₂·4H₂O, 0.76μM ZnCl₂, 0.32μM CuCl₂·2H₂O, 0.13Mμ Na₂Mo₄·2H₂O). While, hydroponic pots of –Mg were supplemented with modified Hoagland’s nutrient solution without magnesium (pH 6.5-7.0, 995μM KNO₃, 931μM Ca(NO₃)₂·4H₂O, 147μM Na₂SO_4, and 282μM KH₂PO₄; micronutrients: 22.5μM Fe(II)-EDTA, 46.2μM H₃BO₄, 14.3μM MnCl₂·4H₂O, 0.76μM ZnCl₂, 0.32μM CuCl₂·2H₂O, 0.13μM Na₂Mo₄·2H₂O). The growth chambers were set to photosynthetically active radiation of 700 μmol·photon m^-2·s^-1, the relative humidity is 80%, 12 hours: 12 hours for light: dark cycle. The nutrient solution was replaced weekly. The experiment was harvested after eight weeks of temperature and Mg treatment.

The tobacco plants were harvested and separated into shoots and roots. Shoot and root were weighed to obtain the fresh biomass of shoot fresh weight (SFW) and root fresh weight (RFW). To determine the shoot dry weight (SDW) and root dry weight (RDW), the plant material was oven-dried for 48h at 70°C (Zhang B. et al., 2020).

For different plant mineral analyses, plants’ roots and shoots were collected, oven-dried, ground, and passed through 60 mesh sieves. For nitrogen (N), phosphorous (P), and potassium (K) minerals analysis, 0.05 g of tobacco plant material of each root and shoot was digested in 5 mL of concentrated sulfuric acid (H₂SO₄) and 4.0 mL of hydrogen peroxide (H₂O₂) on digestion tubes. The digestion tubes were kept in the digestion block for approximately 2h, at a temperature of 200 ± 20°C until about 2ml of clear digestive fluid without granules was obtained. Then the deionized water was added to the digestive fluid and filtered using a 0.45-micron filter, and the extracts were stored at 4°C until nutrient determination (Huang et al., 2021). For magnesium (Mg) analysis, 0.1 g of tobacco plant material of each root and shoot was digested in a mixture of nitric acid (HNO₃): perchloric acid (HClO₄) in a ratio 5v:1v on the hot plate at 180°C until clear aliquot was obtained. The deionized water was added to the digestive fluid and filtered using a 0.45-micron filter, and the extracts were stored at 4°C for further analysis (Sharma et al., 2020). Total P and Mg in the samples were determined by an Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES). Total carbon (C) and total nitrogen (N) were measured by a Skalar Primacs SN100 (Netherlands, DUMAS). Total K was determined with a flame photometer.

On the day of harvest, 0.1 g of fresh leaves of tobacco seeding were taken from each replication to measure leaf total chlorophyll and carotenoid contents using an ultraviolet and visible spectrophotometer (UV-VIS). The samples were put into 25 mL 80% acetone solution. They were then kept in darkness at room temperature overnight to allow for the extraction of the leaf pigments. The mixture of chlorophyll and 80% acetone solution was examined at 663 nm, 645 nm, and 470 nm to determine chlorophyll-a, chlorophyll-b, and carotenoids, respectively. The calculation method was as follows:

Leaf gas exchange parameters, including net photosynthetic rate (Pn), stomatal conductance (Gs), and intercellular CO₂ concentration (Ci), were measured for each plant with a CIRAS-3 portable photosynthesis system (PP System, Maryland USA). Measurements were made on an upper, third fully expanded leaf of an individual plant representing each monoculture tank, typically between 09:00 and 11:00 local time, and were measured once Pn was stable after 2 to 4 min and under the following conditions: 400 μmol·mol^-1·CO₂, 1000 μmol·m^-2·s^-1 photosynthetically active radiation, 25°C leaf temperature (Lee et al., 2011; Koch et al., 2019).

For fluorescence, tobacco seedlings were kept in the dark for 30 min, and one leaf was obtained from the same rosette of each plant sample. The three to four fully expanded leaves were selected for chlorophyll fluorescence yield detection to compare the different treatment leaves. Chlorophyll fluorescence measurements were obtained using a Plant Efficiency Analyzer (PEA, Hansatech, England). The following equations were used to calculate the Fv/F_m (photosystemII- PSII) (Lichtenthaler et al., 2005; Liu et al., 2018);

Where F₀= initial fluorescence, F_v = variable fluorescence, F_m = maximum fluorescence, and F_v/F_m denotes the maximal quantum yield of PSII.

All the statistical analyses were performed using IBM-SPSS version 21.0 (IBM Corporation, New York, USA). Two-way analysis of variance (ANOVA) was performed to analyze the effect of Mg, temperature, and their interaction effect of tobacco morphological traits, minerals nutrients concentration in roots and shoots, photosynthetic traits, and quality parameters. All the measurements are presented as means with their standard errors (Mean ± SE). The statistically significant mean differences among various treatments were identified by using Duncan’s test at a significance level of P<0.05. The principal component analysis (PCA) was performed to evaluate the differences in biomass, minerals, nutrient content, photosynthetic traits, and quality parameters between the +Mg and –Mg under varying levels of low temperature (i.e., cold stress). The results were visualized with Origin Pro 2021 (Origin Lab Corporation, Northampton, USA).

Application of Mg significantly improved the tobacco morphological traits, including shoot fresh weight (SFW), root fresh weight (RFW), Shoot dry weight (SDW), and root dry weight (RDW) as compared to –Mg under cold stress ( Figure 2 ).

For SFW, the increasing trend was recorded with the increase in temperature, and a significant difference was found between different treatments. Increase in SFW under +Mg8, +Mg12, +Mg16 and +Mg25 was 12.9%, 35.0%, 7.7% and 18.2% compared to –Mg8, –Mg12, –Mg16 and –Mg25 treatments, respectively ( Figure 2A ). Similarly, RFW under +Mg treatment was significantly increased compared with –Mg treatments, while no significant difference was found at 8°C ( Figure 2A ). SDW was significantly increased under +Mg25 compared to –Mg25, while non-significant differences were recorded at 8°C and 12°C ( Figure 2B ). For RDW of the tobacco plant, significant differences were found among different treatments, and the maximum RDW was recorded at –Mg25 (1.65 g·plant^-1) and +Mg25 (2.07 g·plant^-1) ( Figure 2B ). The phenotypic differences for +Mg and –Mg were evident under different low temperatures. The results of two-way ANOVA showed the significant main effects of Mg and temperature and their interaction effect for RFW, while RFW showed significant effects only for Mg and temperature. In the case of SDW and RDW, Mg had no significant effect, while temperature had a significant effect ( Table 1 ).

Application of Mg under varying levels of temperature altered the nutrients uptake ( Figure 3 ). For example, carbon (C) contents of the shoot were significantly higher under +Mg treatment compared with –Mg at all temperatures, and the highest C content (6221.42 mg) was recorded at +Mg25 ( Figure 3A ). The increase in shoot C content under +Mg was 8.5%, 2.9%, 43,2% and 19.3% compared with –Mg at 8°C, 12°C, 16°C, and 25°C, respectively. In contrast, root C contents were non-significant at 8°C and 12°C, while significantly higher (849.07 mg) only at 25°C ( Figure 3A ). The results of Two-way ANOVA showed that shoot and root C contents were significantly affected by the main effects of Mg, and temperature, while for interaction effect significant results found for shoot C, and non-significant for root C ( Table 2 ).

Similarly, for nitrogen (N), a significant difference was observed for aboveground plant parts (i.e., shoot), and shoot N contents under +Mg treatment were significantly higher than –Mg at all temperatures and showed maximum N contents at 25°C ( Figure 3B ). For root, N contents were significantly higher at +Mg25 compared with other treatments ( Figure 3B ). Mg, temperature, and their interaction effect had a significant effect on shoot N content, while for root N, only temperature showed the significant effect ( Table 2 ).

The potassium (K) contents of the shoot were not significantly affected among +Mg and –Mg treatments at 8°C, 12°C, 16°C, and 25°C temperature, while temperature had a significant effect. However, it significantly increased with increasing temperature. Maximum shoot K content (683.2 mg) was found at +Mg25 ( Figure 3C ). The root K content was significantly increased with increasing temperature and higher under +Mg application compared with –Mg under all temperatures, and maximum root K content was recorded under +Mg25 ( Figure 3C ). The Mg and temperature had significant effect on root K content. In contrast, the interaction effect for both root and shoot was non-significant ( Table 2 ).

The +Mg significantly increased the shoot phosphorous (P) contents at all four temperature levels compared with –Mg. Shoot P contents significantly increased with increasing temperature, and maximum shoot contents (117.5 mg) were observed at +Mg25 ( Figure 3D ). Hence, Mg and temperature significantly affected shoot P. In contrast, Mg supplementation did not significantly affect P content in tobacco roots. While the increasing temperature also increased the root P contents, and the highest root P contents were recorded at +Mg25 (24.4 mg) ( Figure 3D ). Hence, the temperature had a significant effect on the root P content of tobacco plants ( Table 2 ).

Mg treatment had a significant effect on the Mg contents of the tobacco plant. For instance, shoot Mg contents were significantly higher under +Mg treatment than –Mg under all temperature levels. The maximum shoot Mg was found under +Mg treatment at 25°C ( Figure 3E ). Similar results were recorded for root Mg contents ( Figure 3F ). Shoot and root Mg contents were maximum at +Mg25 about 173.34 mg·plant^-1 and 26.68 mg·plant^-1, respectively. The increase rates were 91.9% and 88.5%, respectively, compared with –Mg8. Thus, shoot and root Mg contents were significantly affected by Mg, temperature, and their interaction effect ( Table 2 ).

The gas exchange parameters, including stomatal conductance (Gs), intercellular carbon dioxide concentration (Ci), transpiration rate (Tr), and photosynthetic rate (Pn), were significantly affected by the Mg application under cold stress. The Gs for both +Mg and –Mg decreased with the decrease in temperature level. The Gs under +Mg tobacco plants increased 36.9%, 62.0%, 9.0% and 29.2% compared to –Mg treatment at 8°C, 12°C, 16°C and 25°C, respectively. The highest value of Gs was recorded at 25°C, i.e., 420.67 mmol·m^-2·s^-1 under +Mg and 325.33 mmol·m^-2·s^-1 for –Mg treatment ( Figure 4A ). For Ci, there was an opposite result, the highest value was found under –Mg treatment about 359.00 μmol·mmol^-1 at 25°C ( Figure 4B ). Whereas, the highest value of Tr and Pn was found under +Mg treatment about 6.72 mmol H₂O·m^-2·s^-1 and 18.27 mmol·m^-2·s^-1, respectively, at 25°C ( Figures 4C, D ). Results of two-way ANOVA showed that Mg and temperature significantly affected Gs, Ci, and Pn ( Table 3 ).

The Mg application had a significant effect on chlorophyll and carotenoid contents. Chlorophyll-a (Chl-a) content was significantly affected by Mg at 12°C, 16°C and 25°C, and maximum Chl-a content was recorded at +Mg25 ( Figure 5A ). A similar trend was observed for chlorophyll-b (Chl-b) ( Figure 5B ) and total chlorophyll contents (Chl-a+b) ( Figure 5C ). The carotenoid content (Cart) was increased under +Mg about 17.6%, 25.0%, 17.6% and 20.8% compared to –Mg at 8°C, 12°C, 16°C and 25°C, respectively ( Figure 5D ). Moreover, Chl-a, Chl-b, Chl-a+b and Cart contents were significantly affected by Mg, temperature, and their interaction effect ( Table 4 ).

Mg supplementation had a significant effect in increasing the sucrose in tobacco plants. Similarly, the starch content increased with increasing temperature, and the maximum value was observed at 25°C 24.7% and 30.6% for both +Mg and –Mg treatment, respectively. It was increased by 7.7%, 7.5%, 33.8% and 24.1% under +Mg treatment compared to –Mg at 8°C, 12°C, 16°C and 25°C, respectively ( Figure 6A ). Hence, the sucrose contents reached a maximum about 17.3% and 16.9% at 25°C for +Mg and –Mg treatment, respectively. It was decreased by 40.2%, 8.5%, 32.1% and 2.5% under +Mg treatment compared to –Mg at 8°C, 12°C, 16°C and 25°C, respectively ( Figure 6B ). Results of two-way ANOVA have been shown that sucrose and starch were significantly affected by Mg and temperature ( Table 5 ).

The above results indicated that the photosynthetic system of tobacco plant had changed to some extent. With the increase of temperature, the chlorophyll of tobacco plants increased and the carbohydrate produced by the photosynthetic system also increased. In order to further clarify the relationship between chlorophyll (Chl-a, Chl-b, Chl-a+b), gas exchange parameters (Ci, Gs, Tr), carbohydrate (sucrose, starch reducing sugar) and photosynthetic rate (Pn), correlation analysis was conducted between chlorophyll (Chl-a, Chl-b, Chl-a+b), gas exchange parameters (Ci, Gs, Tr) carbohydrate (sucrose, starch reducing sugar) and photosynthetic rate (Pn). As shown in Figure 7 , photosynthetic rate (Pn) was positively related to Ci, Gs, Tr, Chl-a, Chl-b, Chl-a+b, sucrose, starch and reducing, respectively.

The JIP-test parameters were used to analyze further tobacco’s photosynthetic behavior in various treatments ( Figure 8 ). It was found that, under 25°C, the +Mg treatment did not significantly affect the photosynthetic behavior of the tobacco, but cold stress significantly affected JIP-test parameters ( Figure 8A ). With decreasing temperature, the TR_O/RC, ET_O/RC, ABS/CSm, TR_O/CSm, ET_O/CSm, PIabs, DFabs indices were significantly reduced with a decrease in temperature ( Figures 8B, C ). Cold stress significantly decreased the PI (abs), and PSII functional activity index based on standardized ABS absorption, and the lower the cold stress, the greater the decline ( Figure 8D ). Mg application significantly inhibited the decrease in PI (abs). Cold stress significantly decreased the DF (abs), and PSII driving force based on standardized ABS absorption.

Simultaneously, photosynthesis’s energy transfer in the electron transport chain was analyzed. As the results showed, after cold stress, the light energy absorbed (ABS/RC) per active reaction center (RC) and the heat dissipation (DIo/RC) were increased, suggesting that after partial RC deactivation caused by cold stress, the energy dissipation efficiency of the remaining active RC improved, which increased the burden of active RC, to resist cold environmental stress. However, +Mg treatment significantly alleviates this procedure. Additionally, cold stress decreased the excitation energy used to reduce QA (TRo/RC) and the electron transfer capacity (ETo/RC) of a single RC in tobacco, while +Mg treatment significantly alleviated the decline of ETo/RC, causing more energy to enter the electron transfer, which was conducive to maintaining the normal operation of the PSΠ linear electron transfer and increasing the tobacco’s resistance to adversity, especially cold stress.

The phenomenological energy fluxes per the exciting cross sections (CS) of the grapevine leaves during the growing season are presented in Figure 9 . 25°C measurements, the leaves were characterized by relatively high energy fluxes, which significantly decreased at 8°C. The rapid decrease of absorbed energy (ABS/RC) and electron transport per cross-section (ET/CS) was observed between 8-25°C ( Figure 9 ). Then, the photosynthetic apparatus was deactivated gradually. The dissipated energy (DI/CS) did not change significantly at different temperatures but gradually increased after magnesium was added. The presented results indicate that the change of temperature (chilling stress) is the main factor of photosynthetic damage in tobacco leaves, and adding magnesium can improve the resistance to photosynthetic damage to some extent.

The principal component analysis (PCA) clustered the input variables of temperature and Mg into different groups ( Figure 10 ). The first two components accounted for 82.1% (68.3% for PC1 and 13.8% for PC2) of the total variation for morphological traits (SFW, RFW, SDW, RDW), nutrients contents (shoot C (S.C), root C (R.C), shoot N (S.N), root N (R.N), shoot K (S.K), root K (R.K), shoot P (S.P), root P (R.P), shoot Mg (S.Mg), root Mg (R.Mg)), photosynthetic traits (Gs, Ci, Pn, Chl-a, Chl-b, Chl-a+b, Cart, ABS/RC, DI_O/RC, TR_O/RC, ET_O/RC, ABS/CSm, DI_O/CSm, TR_O/CSm, ET_O/CSm, PIabs, DFabs), and quality attributes (sucrose and starch), and overall these showed the best performance in under +Mg at 25°C.