E ﬀ ect of Drought Stress at Reproductive Stages on Growth and Nitrogen Metabolism in Soybean

: This study aims to determine variability among soybean ( Glycine max (L.) Merr.) cultivars under drought conditions and how nitrogen metabolites, metabolism-related enzymes, and gene expression vary during soybean growth. Three soybean cultivars, Shennong17 (CV.SN17), Shennong8 (CV.SN8), and Shennong12 (CV.SN12), were grown in pot culture and subjected to drought stress at reproductive stages for 45 days. The results showed that long-term drought stress decreased biomass allocation to reproductive organs, weakened antioxidant capacity, and reduced seed weight, e ﬀ ects that were less pronounced in CV.SN12 compared with those in CV.SN8 and CV.SN17. Drought stress decreased the concentrations of nitrogen and soluble protein but increased nitrate concentration in leaves. This was related to the signiﬁcantly reduction of nitrogen metabolism e ﬃ ciency, including decreased activities of nitrogen metabolism enzymes, and downregulated expression of GmNR , GmNiR , GmGS , and GmGOGAT . Drought stress increased the concentrations of free amino acid, proline, and soluble sugar in leaves to enhance the osmotic adjustment ability. Furthermore, soybean seed weight showed signiﬁcantly correlation ( p < 0.05) with nitrogen-metabolism-related parameters. Based on the performance of growth, nitrogen metabolism, and yield attributes, CV.SN12 showed the highest tolerance to drought, followed by CV.SN8 and CV.SN17. In addition, these nitrogen-metabolism-related parameters could be used in soybeans to select for drought tolerance.


Introduction
Drought stress is one of the most unfavorable environmental stress factors and can severely reduce crop yield and quality [1,2]. Drought stress can interfere with numerous physiological processes, including oxidative stress [3], membrane integrity [4], and enzyme activity [5], all of which could lead to plant growth inhibition. During reproductive stages, source organs (mainly mature leaves) transport carbohydrates and energy to sink organs (mainly seeds) to promote the seeds development and growth. Drought stress occurring at reproductive stages could negatively affect seed size, weight, and composition as well as final crop yield [6,7]. Therefore, understanding the effect of drought stress at reproductive stages on physiological and molecular mechanisms of crops is beneficial to grain yield improvement.
Plants have evolved various mechanisms, including changing morphological structure, regulating water transport, regulating nutrient synthesis and redistribution, superoxide anion scavenging mechanism, osmotic substance regulation, and hormone regulation, to adapt to adverse conditions [8,9]. It has been extensively reported that drought stress induces reactive oxygen species (ROS) accumulation. Excessive ROS results in oxidative damage to cell structures, cell metabolism, and macromolecules, such as membrane lipids, nucleic acids, and proteins [1,10]. Thus, in order to keep the balance between and GOGAT are also related to drought stress response and they are often regarded as important metabolic indicators of drought tolerance [26,27]. For example, Nagy et al. [26] showed that drought stress decreased the activities of glutamine synthetase isoenzymes in sensitive wheat, while no change was found in those tolerance genotypes.
N assimilation products not only act as energy sources but also function as signaling molecules in a variety of physiological metabolism processes and play important roles in plant growth and development [24,28]. Proline is one of important products in nitrogen metabolism and plays an important role in maintaining the integrity of the cytomembrane and protein structure. Furthermore, proline and some other nitrogen compounds are important osmotic adjustment (OA) substances involved in various metabolic pathways in response to abiotic stress [29][30][31], including maintaining cell turgor and subcellular structures, quenching free radicals, and buffering cellular redox potential [29]. Accumulation and synthesis of compatible solutes (amino acids and soluble sugars) and ions (NO 3 − ) are considered as main contributors to OA in plants under drought stress [29,32]. Moreover, previous studies have shown that nitrogen assimilation products, including NO 3 − , NH 4 + , and proline, can regulate the selective absorption and transport of ions in plants and play an important role in maintaining the dynamic balance of physiological metabolism in plants under abiotic stress [33][34][35]. Soybean (Glycine max (L.) Merr.) is an important legume cash crop and is widely cultivated around the world [36,37]. The balance between N accumulation and distribution during soybean reproductive stages is important to grain yield and quality [20]. During the reproductive stages, various energy substances such as amino acids and sugars are rapidly transported, synthesized, and stored in seeds. This process requires sufficient water supply. Previous studies have shown that drought stress at reproductive stages significantly decreases soybean grain yield or even results in no harvest [38][39][40].
In this study, three soybean cultivars with similar maturity, Shennong17 (CV.SN17), Shennong8 (CV.SN8), and Shennong12 (CV.SN12), were used to (1) elucidate the relationships between nitrogen metabolism and seed weight and (2) investigate the physiological, molecular, and enzymatic performance of these cultivars as a response to drought stress during reproductive stages.

Measurement of Soluble Protein, Free Amino Acid, Proline, and Soluble Sugar Concentrations
The soluble protein concentration was determined by the Brilliant Blue G-250 regent with bovine serum albumin (BSA) as a standard according to Bradford [43].
Concentrations of free amino acid and proline were analyzed by the methods according to Bates et al. [44]. Frozen leaf tissues (0.5 g) were homogenized in 10 mL of 3% aqueous sulfosalicylic acid. After filtering the homogenate, 2 mL of filtrate was mixed with 2 mL glacial acetic acid and 2 mL acidic ninhydrin. The reaction mixture was incubated at 100 • C for 1 h and then cooled on the ice for 20 min before extraction with 4 mL toluene. The concentrations of free amino acids and proline were measured with a spectrophotometer at A580 nm and A520 nm, respectively.
Total soluble sugar concentration was determined by the methods according to Hendrix and Donald [45]. Dried leaf tissues (0.1 g) were added in 5 mL of 80% (v/v) ethanol and incubated in a water bath at 80 • C for 30 min, followed by centrifugation at 10,000 × g for 10 min. The residue was extracted two more times using 80% ethanol. The three supernatants were combined and 80% ethanol was added to a total volume of 5 mL. The soluble sugar concentration was determined spectrophotometrically at A620 nm wavelength.
The reaction buffer to measure NR activity contained 0.4 mL nicotinamide adenine dinucleotide (NADH), 1.2 mL of 0.1 M KNO 3 , and 0.4 mL extract. The reaction was initiated by incubating the enzyme at 25 • C for 30 min and terminated using 1 mL sulfanilamide. Then, 1 mL of 1% N-1-naphthyl ethylene diamine dihydrochloride was added into the samples. The samples were incubated at room temperature for 15 min and then centrifuged at 12,000× g for 10 min. NR activity was determined colorimetrically at A540 nm, as in Hageman and Reed [47], based on the reduction of nitrate to nitrite at 25 • C for 30 min. NR activity was calculated from the standard curve of nitrite nitrogen.
The reaction buffer to measure NiR activity contained 0.4 mL of 0.1 M potassium phosphate buffer (pH 7.4), 0.1 mL of 15 mM sodium nitrite, 0.2 mL of 5 mM methyl viologen, 0.2 mL of 86.15 mM sodium dithionite in a 190 mM NaHCO 3 and 0.1 mL extract. The reaction was terminated by a violent agitation on vortex after 30 min incubation. Then, the absorbance of supernatant was measured at A540 nm according to Singh et al. [34]. NiR activity was expressed in terms of µmol NO 2 − reduced g −1 FW h −1 .
The reaction buffer to measure GS activity contained 0.2 mL of 0.3 M Na-glutamate, 0.3 mL of 0.25 M imidazole-HCl (pH 7.0), 0.1 mL of 0.5 M MgSO 4 , 0.2 mL of 0.03 M Na-ATP (pH 7.0), and 0.6 mL extract. The reaction was initiated by incubating the enzyme at 25 • C for 5 min, and then 0.1 mL of 1.0 M hydroxylamine was added. After incubating the mixture at 25 • C for 5 min, the reaction was stopped using 0.4 mL mixed reagent (10% FeCl 3 ·6H 2 O, 50% (v/v) HCl, and 24% (w/v) trichloroacetic acid). The mixture was incubated at 25 • C for 20 min to change color and then centrifuged at 15,000× g for 10 min. The absorbance of supernatant was measured at A540 nm according to Zahoor et al. [32]. GS activity was calculated from the standard curve of c-glutamyl hydroxamate.
The reaction buffer to measure SOD activity contained 50 mM PBS (pH 7.8), 0.1 mM EDTA, 13 mM methionine, 75 µM NBT, 2 µM riboflavin, and 0.1 mL extract. Then, the absorbance of supernatant was measured at A560 nm according to Giannopolitis and Ries [50]. One unit of SOD activity was considered as the amount of enzyme inhibiting 50% of NBT photoreduction.
The reaction buffer to measure POD activity contained 100 mM PBS (pH 6.0), 20 mM of guaiacol, 40 mM of H 2 O 2 , and 0.1 mL extract. Then, the absorbance of supernatant was measured at A460 nm according to Chance and Maehly [51]. One unit of POD activity was defined as change of absorbance of 0.01 units min −1 .
The reaction buffer to measure CAT activity contained 19 mM H 2 O 2 in 50 mM PBS (pH 7.8) and 0.1 mL extract. Then, the absorbance of supernatant was measured at A240 nm according to Havir and McHale [52]. One unit of CAT activity was considered as the amount of enzyme catalyzing the H 2 O 2 decomposition per minute.
The MDA level was determined by 2-thiobarbituric acid (TBA) test as described by Havir and McHale [52].

RNA Extraction and Complementary DNA Synthesis
Total RNA was isolated from each soybean leaf sample using a MiniBEST Universal RNA Extraction Kit (Takara, Kusatsu, Japan) according to the manufacturer's protocol. RNA quantity and integrity were checked by optical density at 260 nm and 1.0% agar gel electrophoresis, respectively. Single-stranded cDNA was synthesized from 1 µg of total RNA using a PrimeScriptTM RT Reagent Kit (Perfect Real Time) (Takara, Kusatsu, Japan).

Gene Expression by qRT-PCR
For each sample, 1 µL of a reaction mixture was used for qRT-PCR in a 20 µL total reaction volume using TransScript ® Top Green qPCR SuperMix (TransScript, Beijing, PR China). Two reference genes, GmEF1a and GmEF1b, were selected for examination in this study [53]. The following thermal cycle conditions were used: 95 • C for 1 min, followed by 39 cycles of 95 • C for 5 s, 58 • C for 20 s, and 60 • C for 20 s. The relative expression was determined according to the 2 − Ct method. The gene-specific primers are listed in Table 1.

Statistical Analysis
The data were analyzed using a two-way analysis of variance (ANOVA) for dry matter accumulation and seed yield (Tables 2 and 3) and a three-way ANOVA for physiological parameters Table 4), as implemented in SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). The differences between the means were compared using Tukey's multiple range test (at a p < 0.05 significance threshold), with significant differences indicated by different letters above the bars in each figure. The correlations between seed weight with physiological parameters in soybean leaves were assessed via Pearson's product-moment correlation [54]. The data presented are means (± SD) of three independent experiments.

Effect of Drought Stress on Dry Matter Accumulation and Seed Yield
Two-way ANOVA analysis revealed that treatment (T), genotype (G), and their interaction significantly affected the dry matter accumulation and seed yield ( Table 2). Three soybean cultivars showed different yield and yield component performances both under control and drought stress conditions. Drought stress during reproductive stages significantly inhibited the accumulation of biomass in root, stem, pod wall, and seed and reduced pod and seed number; more reduction was observed in CV.SN17 and CV.SN8 than CV.SN12. As shown in Table 2, pod number, seed number, and seed weight were reduced under drought stress by 64.60%, 63.55%, and 68.60%, respectively, in CV.SN17 as compared to control. Pod number, seed number, and seed weight were reduced under drought stress by 44.55%, 51.72% and 66.43%, respectively, in CV.SN8 as compared to control. Pod number, seed number, and seed weight were reduced under drought stress by 36.25%, 60.41% and 57.03%, respectively, in CV.SN12 as compared to control. Compared with the control, drought stress significantly reduced the 100-seed weight of CV.SN17 and CV.SN8, but had no significant effect on the 100-seed weight of CV.SN12. Drought stress had different effects on the distribution ratio of dry matter accumulation in different organs (Table 3). Drought stress significantly decreased the distribution ratio of dry matter in seed and pod wall of three soybean cultivars, but increased the distribution ratio of dry matter in root and stem, except root biomass distribution ratio in CV.SN17 root. Seed biomass/total biomass ratio was more reduced in CV.SN8 (decreased by 13.24%) due to drought but less reduction was recorded both in CV.SN17 (11.37%) and CV.SN12 (11.92%).

Effect of Drought Stress on Antioxidant Enzymes Activities and MDA Contents in Leaves
The activities of SOD, POD, and CAT and the contents of MDA in soybean leaves significantly increased with plant growth under well-watered growth conditions ( Figure 1). Compared with the control, SOD activities were significantly increased in leaves during 15-30 DAF in all cultivars under drought stress, while significantly decreased at 45 DAF ( Figure 1A). POD activity was increased by 17.62%-90.74% in CV.SN17, 21.91%-246.08% in CV.SN8 and 110.88%-249.03%2 in CV.SN12 under drought stress as compared to control during 15-45 DAF ( Figure 1B). CAT activity was increased by 32.06%-34.29% in CV.SN17, 30.85%-41.37% in CV.SN8 and 38.64%-48.09% in CV.SN12 during 15-45 DAF ( Figure 1C). MDA concentration increment during drought stress was 52.86%-90.11% in CV.SN17, 22.46%-45.45% in CV.SN8 and 22.25%-28.76% in CV.SN12 as compared to control during 15-45 DAF ( Figure 1D). Plant growth stage (S), treatment (T), genotype (G), and their interaction significantly affected activities of SOD, POD, and CAT and concentration of MDA in leaves, except G and S × T × G for MDA content, according to the three-way ANOVA (Table 4).

276
The activities of SOD, POD, and CAT and the contents of MDA in soybean leaves significantly Nitrogen

Effect of Drought Stress on Nitrogen and Nitrate Contents in Leaves
Plant growth stage (S), treatment (T), genotype (G) and their interaction significantly affected the concentrations of nitrogen and nitrate in leaves, except S × T × G for nitrogen concentration by the three-way ANOVA ( Table 4). The concentrations of nitrogen and nitrate in soybean leaves slightly decreased with plant growth under well-watered growth conditions (Figure 2). Drought stress had significant effects on the nitrogen concentration and nitrate nitrogen concentration (p < 0.05, Table 3). Nitrogen concentration was reduced under drought stress by 25.11%, 13.21%, and 13.05% in CV.SN17, CV.SN8, and CV.SN12, respectively. However, compared with the control, drought stress increased the nitrate concentration by 132.90%, 78.57%, and 54.93% in CV.SN17, CV.SN8, and CV.SN12, respectively.

Effect of Drought Stress on Nitrogen and Nitrate Contents in Leaves
Plant growth stage (S), treatment (T), genotype (G) and their interaction significantly affected the concentrations of nitrogen and nitrate in leaves, except S × T × G for nitrogen concentration by the three-way ANOVA ( Table 4). The concentrations of nitrogen and nitrate in soybean leaves slightly decreased with plant growth under well-watered growth conditions (Figure 2). Drought stress had significant effects on the nitrogen concentration and nitrate nitrogen concentration (p < 0.05, Table 3). Nitrogen concentration was reduced under drought stress by 25.11%, 13.21%, and 13.05% in CV.SN17, CV.SN8, and CV.SN12, respectively. However, compared with the control, drought stress increased the nitrate concentration by 132.90%, 78.57%, and 54.93% in CV.SN17, CV.SN8, and CV.SN12, respectively.

Effect of Drought Stress on the Activities of Enzymes Related to Nitrogen Metabolism in Leaves
To better understand how drought affects nitrogen utilization and partitioning of nitrogen, the activities of four nitrogen-metabolism-related enzymes were analyzed, including NR and NiR for N assimilation and GS and GOGAT for ammonia assimilation. The activities of NR, NiR, GS, and GOGAT in soybean leaves significantly decreased with plant growth under well-watered growth conditions, except the NiR activities in CV.SN8 and CV.SN12 during 15-30 DAF (Figure 3). Compared with the control, drought stress caused significant decreases in the activities of NR, NiR, GS, and GOGAT by 32.06%, 23.10%, 34.61%, and 23.23%, respectively, in CV.SN17. Compared with the control, drought stress caused significant decreases in the activities of NR, NiR, GS, and GOGAT by 28.60%, 20.83%, 23.41%, and 20.21%, respectively, in CV.SN8. Compared with the control, drought stress caused significant decreases in the activities of NR, NiR, GS, and GOGAT by 29.01%, 18.66%, 18.86%, and 19.51%, respectively, in CV.SN12. Plant growth stage (S), treatment (T), genotype (G), Figure 2. Changes of (A) nitrogen content and (B) nitrate content in leaves of CV.SN17, CV.SN8, and CV.SN12 under control and drought stress conditions. Data show the means ± standard deviation of three independent samples. Different letters above vertical bars indicate significant differences between means at a p < 0.05 level.

Effect of Drought Stress on the Activities of Enzymes Related to Nitrogen Metabolism in Leaves
To better understand how drought affects nitrogen utilization and partitioning of nitrogen, the activities of four nitrogen-metabolism-related enzymes were analyzed, including NR and NiR for N assimilation and GS and GOGAT for ammonia assimilation. The activities of NR, NiR, GS, and GOGAT in soybean leaves significantly decreased with plant growth under well-watered growth conditions, except the NiR activities in CV.SN8 and CV.SN12 during 15-30 DAF (Figure 3). Compared with the control, drought stress caused significant decreases in the activities of NR, NiR, GS, and GOGAT by 32.06%, 23.10%, 34.61%, and 23.23%, respectively, in CV.SN17. Compared with the control, drought stress caused significant decreases in the activities of NR, NiR, GS, and GOGAT by 28.60%, 20.83%, 23.41%, and 20.21%, respectively, in CV.SN8. Compared with the control, drought stress caused significant decreases in the activities of NR, NiR, GS, and GOGAT by 29.01%, 18.66%, 18.86%, and 19.51%, respectively, in CV.SN12. Plant growth stage (S), treatment (T), genotype (G), and their interaction significantly affected the activities of NR, NiR, GS and GOGAT in leaves, except T × G and S × T × G for NR activities, S × G for NiR activities, S × T, T × G, and S × T × G for GOGAT activities, according to the three-way ANOVA (Table 4). and their interaction significantly affected the activities of NR, NiR, GS and GOGAT in leaves, except T × G and S × T × G for NR activities, S × G for NiR activities, S × T, T × G, and S × T × G for GOGAT activities, according to the three-way ANOVA (Table 4).

Effect of Drought Stress on Expression Levels of Genes Involved in Nitrogen Metabolism in Leaves
To investigate the mechanisms of nitrogen metabolism pathway in leaves under drought stress, a qRT-PCR expression analysis was performed on four genes, namely GmNR, GmNiR, GmGS, and GmGOGAT. The expression levels of GmNR, GmNiR, GmGS, and GmGOGAT in soybean leaves significantly decreased with plant growth under well-watered growth conditions (Figure 4). Compared with the control, drought stress significantly decreased the expression levels of GmNR, GmNiR, GmGS, and GmGOGAT by 63.26%, 57.68%, 51.20%, and 44.92%, respectively, in CV.SN17.

Effect of Drought Stress on Expression Levels of Genes Involved in Nitrogen Metabolism in Leaves
To investigate the mechanisms of nitrogen metabolism pathway in leaves under drought stress, a qRT-PCR expression analysis was performed on four genes, namely GmNR, GmNiR, GmGS, and GmGOGAT. The expression levels of GmNR, GmNiR, GmGS, and GmGOGAT in soybean leaves significantly decreased with plant growth under well-watered growth conditions (Figure 4). Compared with the control, drought stress significantly decreased the expression levels of GmNR, GmNiR, GmGS, and GmGOGAT by 63.26%, 57.68%, 51.20%, and 44.92%, respectively, in CV.SN17. Compared with the control, drought stress significantly decreased the expression levels of GmNR, GmNiR, GmGS, and GmGOGAT by 43.60%, 72.41%, 51.55%, and 37.34%, respectively, in CV.SN8. Compared with the control, drought stress significantly decreased the expression levels of GmNR, GmNiR, GmGS, and

Effect of Drought Stress on Soluble Protein, Free Amino Acid, Proline, and Soluble Sugar Contents in Leaves
Under well-watered growth conditions, the concentrations of soluble protein and free amino acid in soybean leaves showed a slightly decrease with plant growth, while proline concentrations

Effect of Drought Stress on Soluble Protein, Free Amino Acid, Proline, and Soluble Sugar Contents in Leaves
Under well-watered growth conditions, the concentrations of soluble protein and free amino acid in soybean leaves showed a slightly decrease with plant growth, while proline concentrations were increased ( Figure 5). Drought stress had significant effects on the concentrations of soluble protein, free amino acid, proline, and soluble sugar (p < 0.05, Table 3). Soluble protein content was reduced under drought stress by 27.55%, 21.82%, and 18.14% in CV.SN17, CV.SN8 and CV.SN12, respectively ( Figure 5A). However, compared with the control, drought stress increased the concentrations of free amino acids, proline, and soluble sugar by 17.11%, 33.23%, and 12.13%, respectively, in CV.SN17. Compared with the control, drought stress increased the concentrations of free amino acids, proline, and soluble sugar by 27.66%, 75.17%, and 12.02%, respectively, in CV.SN8. Compared with the control, drought stress increased the concentrations of free amino acids, proline, and soluble sugar by 42.28%, 65.97%, and 13.04%, respectively, in CV.SN12 (Figure 5B-D). Plant growth stage (S), treatment (T), genotype (G), and their interaction significantly affected the concentrations of soluble protein, free amino acid, proline, and soluble sugar in leaves, except S × T for soluble content and T × G and S × T × G for soluble sugar content, according to the three-way ANOVA (Table 4).
Agronomy 2020, 10, x FOR PEER REVIEW 4 of 22 were increased ( Figure 5). Drought stress had significant effects on the concentrations of soluble protein, free amino acid, proline, and soluble sugar (p < 0.05, Table 3). Soluble protein content was reduced under drought stress by 27.55%, 21.82%, and 18.14% in CV.SN17, CV.SN8 and CV.SN12, respectively ( Figure 5A). However, compared with the control, drought stress increased the concentrations of free amino acids, proline, and soluble sugar by 17.11%, 33.23%, and 12.13%, respectively, in CV.SN17. Compared with the control, drought stress increased the concentrations of free amino acids, proline, and soluble sugar by 27.66%, 75.17%, and 12.02%, respectively, in CV.SN8. Compared with the control, drought stress increased the concentrations of free amino acids, proline, and soluble sugar by 42.28%, 65.97%, and 13.04%, respectively, in CV.SN12 ( Figure 5B-D). Plant growth stage (S), treatment (T), genotype (G), and their interaction significantly affected the concentrations of soluble protein, free amino acid, proline, and soluble sugar in leaves, except S × T for soluble content and T × G and S × T × G for soluble sugar content, according to the three-way ANOVA (Table 4).

Correlations of Seed Weight with Physiological Parameters
The correlations of seed weight with physiological parameters during reproductive phases (15-45 DAF) were analyzed (Table 5). Seed weight showed a significant positive correlation (p < 0.05) with nitrogen concentration at 30 DAF and 45 DAF and a significant negative correlation with nitrate content during 15-45 DAF. Seed weight was positively correlated with soluble protein (p < 0.05), but negatively correlated with free amino acid, proline, and soluble sugar (p < 0.05). The activities of nitrogen-metabolism-related enzymes and expression levels of genes involved in metabolism all showed a significant positive correlation with seed weight (p < 0.05), with the exception of seed weight and GOGAT activity at 15 DAF, which was not significantly correlated. The activities of SOD, POD and CAT, and MDA content all showed a significant negative correlation with seed weight (p < 0.05), with the exception of seed weight and SOD activity at 45 DAF, which showed a significant positive correlation (p < 0.05). Table 5. Pearson's correlation of seed weight with physiological parameters during reproductive phases.

Discussion
Drought stress has significantly negative effects on soybean growth, development, and yield [38,39]. Drought stress inhibits the production of photosynthetic products by decreasing leaf photosynthetic capacity [8,55] and the accelerating leaf senescence and oxidative stress [56]. Our study showed that drought stress during the reproductive stages decreased biomass of soybean stem, root, and seed ( Table 2) and reduced biomass allocation to seed compared to the control (Table 3), which finally led to seed weight loss. The production and mobilization of assimilates to developing seeds decreased and considerable yield losses occurred in plants undergoing drought stress. Three soybean cultivars showed differential behaviors depending on their differential drought stress sensitivities under drought stress, with CV.SN12 being the most tolerant, followed by CV.SN17 and CV.SN8 as the most susceptible. It obvious from this research that soybean cultivars under drought stress displayed different physiological and molecular responses for regulating growth, gene expression levels, and enzyme activities involved in nitrogen-assimilates production and metabolism.
Drought-stress-induced excessive production of ROS can cause oxidative stress, resulting in membrane lipid peroxidation, metabolic disorders, and serious plant growth inhibition [57]. In plants, antioxidant enzymes, including SOD, POD, and CAT, play an important role in scavenging excess ROS and maintaining ROS in a dynamic equilibrium. The activities of antioxidant enzymes in plants can directly reflect the ability of scavenging ROS. Previous studies showed that drought stress increased the SOD, POD, and CAT activities in soybean [58] and alfalfa [4]. In this study, short-term drought stress (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) increased the activities of antioxidant enzymes, including SOD, POD, and CAT ( Figure 1A-C), indicating that the antioxidant enzyme system was involved in the early drought resistance response of soybean. However, once the ROS produced by plants exceed the scavenging capacity of antioxidant enzymes, the antioxidant system will be destroyed [15,59]. Our results showed that long-term drought stress (45 DAF) significantly reduced SOD activity in leaves, but increased POD and CAT enzyme activities compared to the control ( Figure 1A-C). Previous studies showed that SOD and CAT activities were significantly decreased under prolonged drought stress in Kentucky bluegrass [12,60]. Thus, although the changes of antioxidant enzyme activities vary among species under long-term stress, antioxidant capacity to scavenge ROS was weakened, leading to ROS-mediated oxidative damage. This was further supported by the fact that drought stress increased MDA concentration in soybean leaves, especially at 45 DAF ( Figure 1D). POD and CAT were higher in CV.SN12 than CV.SN17 and CV.SN8 under drought stress, which might contribute to retarded leaf senescence and reduced seed weight. Therefore, drought stress enhanced the peroxidation degree of membranes and weakened the scavenging ability of ROS, which caused damage to cellular components and basic metabolism and finally inhibited plant growth and reduced yield.
N is an essential element for soybean growth [19,61]. Effective N fixation and utilization can promote crop growth and increase yield [20,62]. Zapata et al. [63] suggested that maximum N fixation occurs during R3 and R5 stages of soybean development. A number of previous studies have suggested that N application at reproductive stages increased soybean yield, especially in high-yielding environments [20,[64][65][66]. However, when N supply does not meet requirements of growth and development, soybean will remobilize N accumulated in leaves to the grain, accompanied by diminishing photosynthetic capacity, thus decreasing yield potential [19]. Thus, coordination between N supply and N utilization is crucial to plant growth and development [17,67]. During seed-filling stages, N concentration in leaves decreased with the development of soybean growth process (Figures 2A and 6), indicating that N was transferred from leaf to grain to meet the demand of seed development and growth. Drought stress reduced N concentration, but increased nitrate concentration in leaves, especially during 30-45 DAF (Figure 2). Nitrate is an important signaling element for various molecular and physiological processes in plants which integrates second messengers to early signaling and nitrate perception, leading to physiological and metabolic changes [68,69]. The accumulated NO 3 − not only acts as an important nutrient for cells but also as osmoprotectant against injuries to the membrane [70][71][72]. Wilkinson et al. [71] indicated that NO 3 − in leaves contributed to regulate stomatal opening through its effect on depolarization of guard cells. Thus, the accumulation of NO 3 − under drought stress might be a strategy for soybeans to adapt to drought stress. Previous studies showed that drought stress regulated N metabolism balance through changes in N metabolic enzyme activities [17,18,32]. Drought stress decreased activities of NR and GOGAT enzymes in both leaves and roots of Populus simonii seedlings [67] and activities of NR, GS, and GOGAT enzymes in cotton leaves [32]. The changes of enzyme activities can directly affect the efficiency of N uptake and utilization. Drought stress significantly decreased NR and NiR activities ( Figure 3A,B) and the expression levels of GmNR and GmNiR ( Figure 4A,B) in soybean leaves, indicating that drought stress reduced the ability of NO 3 − reduction to NH 4 + and resulted in a decrease in the efficiency of nitrogen assimilation. This can be further supported by the fact that drought stress induced NO 3 − accumulation in leaves ( Figures 2B and 6). Efficient GS/GOGAT metabolism cycle is conducive to the rapid NH 4 + detoxification [73]. In this experiment, GS and GOGAT activities were decreased ( Figure 3C,D) and the expression levels of GmGS and GmGOGAT were downregulated under drought stress ( Figure 4C,D). This result indicated that drought stress inhibited the GS/GOGAT metabolism cycle, leading to weakened NH 4 + assimilation and causing toxicity to cells. Furthermore, seed weight was positively correlated with nitrogen content and nitrogen-metabolism-related parameters, especially during late seed filling stages (30-45 DAF) ( Table 5). The lower NO 3 − concentration and less decreased activities of nitrogen-metabolism-related enzymes and expression levels of genes related to metabolism in CV.SN12 than in CV.SN17 and CV.SN8 suggested that maintaining high N utilization and metabolism was responsible for enhancing drought stress adaptation.
Agronomy 2020, 10, x FOR PEER REVIEW 7 of 22 . Figure 6. A model of the nitrogen regulation in soybean leaf in response to drought stress. Here, drought stress triggers a nitrogen-mediated tandem reaction in response to drought stress. Drought stress might change the expression of key regulatory metabolic genes and the activities of nitrogen metabolism enzymes, which regulate nitrogen distribution and utilization to adapt to environmental stress. NR, nitrate reductase; NiR, nitrite reductase; GS, glutamine synthetase; GOGAT, glutamate synthetase; Glu, glutamate; Gln, glutamine; 2-OXG, 2-oxoglutarate. Increased items under drought stress are marked with upward red arrows and decreased items are marked with downward green arrows.
Previous studies showed that drought stress regulated N metabolism balance through changes in N metabolic enzyme activities [17,18,32]. Drought stress decreased activities of NR and GOGAT enzymes in both leaves and roots of Populus simonii seedlings [67] and activities of NR, GS, and GOGAT enzymes in cotton leaves [32]. The changes of enzyme activities can directly affect the efficiency of N uptake and utilization. Drought stress significantly decreased NR and NiR activities ( Figure 3A and B) and the expression levels of GmNR and GmNiR ( Figure 4A and B) in soybean leaves, indicating that drought stress reduced the ability of NO3 − reduction to NH4 + and resulted in a decrease in the efficiency of nitrogen assimilation. This can be further supported by the fact that drought stress induced NO3 − accumulation in leaves ( Figure 2B and Figure 6). Efficient GS/GOGAT metabolism cycle is conducive to the rapid NH4 + detoxification [73]. In this experiment, GS and GOGAT activities were decreased ( Figure 3C and D) and the expression levels of GmGS and GmGOGAT were downregulated under drought stress ( Figure 4C and D). This result indicated that drought stress inhibited the GS/GOGAT metabolism cycle, leading to weakened NH4 + assimilation and causing toxicity to cells. Furthermore, seed weight was positively correlated with nitrogen content and nitrogen-metabolism-related parameters, especially during late seed filling stages (30-45 DAF) ( Table 5). The lower NO3 − concentration and less decreased activities of nitrogen-metabolismrelated enzymes and expression levels of genes related to metabolism in CV.SN12 than in CV.SN17 and CV.SN8 suggested that maintaining high N utilization and metabolism was responsible for enhancing drought stress adaptation.
Synthesis and accumulation of soluble metabolites such as soluble sugars and amino acids, especially proline, mediate OA during drought stress [31,74]. Drought stress increased the contents of free amino acid and proline ( Figure 5B and C), indicating that N metabolism was involved in OA in response to drought stress. These results were in accordance with previous studies, which reported that proline accumulation was a part of the drought-stress response [75,76]. Furthermore, proline not only functions as an osmoprotectant, but it can also efficiently scavenge ROS, thus protecting plant cells from oxidative damage [77,78]. The increments of free amino acid and proline could have Figure 6. A model of the nitrogen regulation in soybean leaf in response to drought stress. Here, drought stress triggers a nitrogen-mediated tandem reaction in response to drought stress. Drought stress might change the expression of key regulatory metabolic genes and the activities of nitrogen metabolism enzymes, which regulate nitrogen distribution and utilization to adapt to environmental stress. NR, nitrate reductase; NiR, nitrite reductase; GS, glutamine synthetase; GOGAT, glutamate synthetase; Glu, glutamate; Gln, glutamine; 2-OXG, 2-oxoglutarate. Increased items under drought stress are marked with upward red arrows and decreased items are marked with downward green arrows.
Synthesis and accumulation of soluble metabolites such as soluble sugars and amino acids, especially proline, mediate OA during drought stress [31,74]. Drought stress increased the contents of free amino acid and proline ( Figure 5B,C), indicating that N metabolism was involved in OA in response to drought stress. These results were in accordance with previous studies, which reported that proline accumulation was a part of the drought-stress response [75,76]. Furthermore, proline not only functions as an osmoprotectant, but it can also efficiently scavenge ROS, thus protecting plant cells from oxidative damage [77,78]. The increments of free amino acid and proline could have resulted from N assimilation (mainly by amino acid synthesis) or proteolysis [79]. In this study, the accumulation of free amino acids and proline ( Figure 5B,C, Figure 6) were more likely resulting from proteolysis, for the decreased protein content ( Figure 5A) and N-metabolism-related enzyme activities ( Figure 3). However, Zhong et al. [17] suggested that the short-term water stress increased the contents of free amino acids and proline due to the synthesis of amino acids, because the protein content in rice remained unchanged and the activity of N assimilation enzymes increased. Therefore, drought stress regulation of amino acids and proline metabolism may vary among species, plant growth stages, treatment degree, and treatment duration. Under drought stress, soluble sugar accumulated in leaves ( Figure 5D) also participated in osmotic adjustment, which was beneficial to maintain leaf turgor. The accumulated soluble sugars can cooperate with amino acids to prevent dehydration of cells and maintain the structure and function of cell membranes under drought stress [17,80]. Furthermore, seed weight was negatively correlated with the contents of free amino acids, proline, and soluble sugar ( Table 4), indicating that efficient OA may contribute to reduce seed weight loss under drought stress. The increment of free amino acids, proline, and soluble sugars concentrations were higher in CV.SN12 than CV.SN17 and CV.SN8 under drought stress. Taken together, these results suggested that N metabolism cooperated with carbon metabolism to modulate OA in leaves under drought stress.

Conclusions
This study showed that drought-stress-reduced soybean seed weight was due to the reduced biomass accumulation of organs and decreased biomass allocated to seed. Long-term drought stress decreased SOD activity, which induced ROS-mediated oxidative damage and MDA accumulation. Soybean growth under drought stress is regulated by nitrogen distribution and metabolism ( Figure 6). Drought stress reduced the activities of nitrogen metabolic enzymes and downregulated the expression of genes such as GmNR, GmNiR, GmGS, and GmGOGAT in soybean leaves. These changes declined the utilization efficiency of nitrogen, resulting in the decreased nitrogen content and increased nitrate content. N assimilation products, including free amino acids and proline, cooperated with soluble-sugar-coordinated osmotic adjustment in leaves under drought stress. Our results showed that soybean seed weight was significantly (p < 0.05) correlated with nitrogen-metabolism-related parameters. It was evident that the combination of different changes in gene expression and enzymatic activities could be responsible for drought tolerance. Based on the performance of growth, nitrogen metabolism, and yield attributes, CV.SN12 showed the highest tolerance to drought, followed by CV.SN8 and CV.SN17. Our results indicated that these nitrogen-metabolism-related parameters could be used as selection traits for drought tolerance in soybean.