Accumulated Expression Level of Cytosolic Glutamine Synthetase 1 Gene (OsGS1;1 or OsGS1;2) Alter Plant Development and the Carbon-Nitrogen Metabolic Status in Rice

Maintaining an appropriate balance of carbon to nitrogen metabolism is essential for rice growth and yield. Glutamine synthetase is a key enzyme for ammonium assimilation. In this study, we systematically analyzed the growth phenotype, carbon-nitrogen metabolic status and gene expression profiles in GS1;1-, GS1;2-overexpressing rice and wildtype plants. Our results revealed that the GS1;1-, GS1;2-overexpressing plants exhibited a poor plant growth phenotype and yield and decreased carbon/nitrogen ratio in the stem caused by the accumulation of nitrogen in the stem. In addition, the leaf SPAD value and photosynthetic parameters, soluble proteins and carbohydrates varied greatly in the GS1;1-, GS1;2-overexpressing plants. Furthermore, metabolite profile and gene expression analysis demonstrated significant changes in individual sugars, organic acids and free amino acids, and gene expression patterns in GS1;1-, GS1;2-overexpressing plants, which also indicated the distinct roles that these two GS1 genes played in rice nitrogen metabolism, particularly when sufficient nitrogen was applied in the environment. Thus, the unbalanced carbon-nitrogen metabolic status and poor ability of nitrogen transportation from stem to leaf in GS1;1-, GS1;2-overexpressing plants may explain the poor growth and yield.


Introduction
Nitrogen is an essential macronutrient required for plant growth and development, and it is a major limiting factor in determining plant productivity and crop yield [1][2][3][4]. Carbon is crucial for plants to perform their routine and fundamental cellular activities. Carbon compounds include various carbohydrates; in particular, sucrose, glucose and organic acids provide both the energy and the carbon skeletons for ammonium (NH 4 + ) assimilation during amino acid biosynthesis. Amino acids and the resulting proteins, in particular enzymes, are essential for nearly all cellular activities [5]. In addition to their independent utilization, the coordination and optimal functioning of the metabolic pathways for nitrogen and carbon assimilation in plants are critical for determining plant growth and, ultimately, biomass accumulation [6,7]. In addition, maintaining an appropriate balance or ratio of carbohydrates to nitrogen metabolites in the cell, which is referred to as the ''carbon/nitrogen balance,'' is also important for the regulation of plant growth, development and yield production [5,[8][9][10].
In higher plants, glutamine synthetase (GS; EC 6.3.1.2) is a key enzyme for the assimilation of ammonium into glutamine (Gln) [11][12][13]. The reaction of GS is coupled with glutamate synthase (GOGAT), which generates two molecules of glutamate (Glu). One molecule of glutamate is cycled back to the GS reaction as a substrate, and the other molecule of glutamate is exported or used to produce other amino acids [14]. Two major isoforms of GS, GS1 in the cytosol and GS2 in plastids, are present in higher plants [15,16]. A multigene family encodes the cytosolic isoform GS1, and a single gene encodes the plastidic isoform GS2, although several GS2 genes have been identified in soybean, alfalfa and durum wheat [17][18][19][20][21][22][23]. In rice, there are three genes that encode cytosolic GS1 (OsGS1;1, OsGS1;2 and OsGS1;3), and one gene encodes chloroplastic GS2 (OsGS2) [2]. OsGS1;1 is mainly expressed in the shoot, OsGS1;2 mRNA is abundant in the root, OsGS1;3 mRNA is present only in the spikelets, and OsGS2 is mainly expressed in the leaf [24].
Numerous studies of the GS enzyme have emphasized the importance of this enzyme in plant nitrogen metabolism. Previous reports have shown that overexpression of GS genes aimed to improve plant nitrogen assimilation presented variable results. For example, accelerated growth rate was observed in transgenic Lotus corniculatus plants overexpressing a soybean GS1 isoenzyme driven by the CaMV 35S promoter [25]. In addition, vegetative growth and photosynthetic capacity improvements were reported for cytosolic GS1-overexpressed poplar trees [26][27][28][29], tobacco [30][31][32] and L. japonicus [33]. Earlier flower and seed development were also observed in transgenic wheat lines containing Phaseolus vulgaris GS1 under the control of the Rubisco small subunit (rbcS) promoter [34]. In addition, a high level of GS activity in the root was negatively correlated with aboveground biomass in Lotus japonicus [35]. However, pea plants overexpressing a soybean GS1 under the control of a root-specific promoter showed no consistent effect on total biomass [36]. Furthermore, recent studies of mutants deficient in leaf cytosolic GS in rice, maize and Arabidopsis demonstrated the important role of cytosolic GS in nitrogen remobilization for grain filling, carbon and nitrogen metabolic coordination, and ammonium homeostasis. In rice, homozygous lines of three OsGS1;1-knockout mutants showed a decrease in cytosolic GS enzyme activity, delay in growth rate, reduced spikelet weight and number, and reduced fertility [24]. Metabolic profile analyses revealed an imbalance in the levels of sugars, amino acids and metabolites in the tricarboxylic acid cycle and an over-accumulation of secondary metabolites in OsGS1;1-knockout mutants [37]. In maize, gln1-4 mutants exhibited a reduced kernel size and the gln1-3 mutant exhibited a reduced kernel number, a higher free amino acid content and a lower carbon/nitrogen ratio with no effect on the shoot dry weight, even in double mutants [38,39]. In Arabidopsis, gln1;2 knockout mutants displayed lower glutamine synthetase activity, higher ammonium concentration, and reduced rosette biomass compared with the wildtype (WT) under conditions of ample nitrate supply [40].
Because rice growth and yield requires abundant nitrogen, large amounts of nitrogen fertilizers are used. However, crop plants use less than half of the nitrogen fertilizers applied [41]. The unused nitrogen is inevitably leached into the underground water system and lost to the atmosphere, resulting in severe environmental pollution. Recent analyses have demonstrated that soil acidification in China primarily resulted from high-N fertilizer inputs [42]. To improve nitrogen use efficiency, the rice yield and/or to ensure normal plant growth and yield under low nitrogen fertilizer supplies, we generated OsGSs-overexpressing transformants driven by the CaMV35S promoter and obtained transgenic rice plants using the Agrobacterium-mediated transformation method in the previous study [43,44]. The GS1;1-, GS1;2-overexpressing plants displayed an imperceptible growth phenotype at the seedling stage when grown hydroponically under both normal and low nitrogen conditions and showed decreases in both grain yield and total amino acids in seeds grown in fields with low nitrogen fertilizer [43]. To identify the reasons for these observations, we systematically analyzed differences in the growth, carbon-nitrogen metabolic status and gene expression profiles between the GS1;1-, GS1;2-overexpressing rice and wildtype Zhonghua 11 at different developmental stages grown under different nitrogen levels. The results in this study revealed that overexpressing GS1;1, GS1;2 genes altered plant growth and development, yield, carbon and nitrogen metabolism. The imbalance in carbon and nitrogen may be the main reason for the decreased yield in GS1;1-, GS1;2overexpressing plants.

Plant Materials and Growth Conditions
Two cDNA sequences encoding the rice cytosolic GS1;1 (AB037595) and GS1;2 (AB180688) were isolated from the Minghui 63 normalized cDNA library (http://www.redb.ncpgr. cn). GS fragments were then ligated into the pCAMBIA 1301 S vector, driven by cauliflower mosaic virus (CaMV) 35 S promoter. The chimeric gene was transformed into the japonica rice cultivar Zhonghua 11 by an Agrobacterium tumefaciens-mediated transformation method. The GS expression level and copy numbers of the transgene were analyzed by the Northern blot and Southern blot techniques in the T 0 generation which were described in the previous study [43].
Seeds of GS1;1, GS1;2-overexpressing rice in the T 3 generation and wildtype Zhonghua 11 (Oryza sativa ssp. japonica) were germinated and sown in sand. At the 2-leaf stage, the seedlings were transferred into a normal nutrient solution containing 1. 44 [45]. The culture solution was refreshed every 3 days. After a week, the plants were transferred into a culture solution without NH 4 NO 3 (06N treatment), a culture solution with 1/10 NH 4 NO 3 (0.16N treatment), a culture solution with 5-fold NH 4 NO 3 (56N treatment), and a culture solution with complete nutrients (16N treatment). The plant materials were harvested at the tillering stage and the heading stage for analysis of the growth phenotype, leaf SPAD value, photosynthesis, carbon and nitrogen content, concentration of soluble proteins and carbohydrates, metabolic profiling and gene expression. Moreover, seedlings of the GS1;1, GS1;2-overexpressing rice and wildtype Zhonghua 11 were transferred into pots fertilized with 0.15 g P 2 O 5 /kg soil and 0.2 g K/kg soil (06N treatment), pots fertilized with 0.02 g N/kg soil, 0.15 g P 2 O 5 /kg soil and 0.2 g K/kg soil (0.16N treatment), pots fertilized with 0.2 g N/kg soil, 0.15 g P 2 O 5 /kg soil and 0.2 g K/kg soil (16N treatment), and pots fertilized with 1 g N/kg soil, 0.15 g P 2 O 5 /kg soil and 0.2 g K/kg soil (56N treatment). At the mature stage, the yield production was analyzed. All the plants were grown in the pot farm in Huazhong Agricultural University, Wuhan, China.

Nitrogen Uptake Assay
Seeds of the GS1;1, GS1;2-overexpressing rice and wildtype Zhonghua 11 (Oryza sativa ssp. japonica) germinated and were sown in sand. At the 2-leaf stage, the seedlings were transferred into the normal nutrient solution as previously described [45]. At the tillering stage, the NH 4 NO 3 in the nutrient solution was replaced with 15 NH 4 15 NO 3 . Six samples of randomly mixed plant root, stem and leaf materials were harvested after 1 h, 3 h, 8 h, 1 d and 3 d. The 15 N content was analyzed using an isotope mass spectrometer (ANCA-MS, Europa Scientific, Crewe, UK) according to the manufacturer's instructions. The total carbon and nitrogen content in the root, stem and leaves of GS1;1-, GS1;2overexpressing plants and wildtype plants under the 06N, 0.16N, 16N and 56N conditions at the tillering stage was determined using a C/N analyzer (Elementar, Vario MAX CN, Germany) according to the manufacturer's instructions, with L-glutamic acid as a standard.

Determination of the Leaf SPAD Value and Photosynthetic Parameters
At both the tillering stage and heading stage, every ten plants of the GS1;1, GS1;2-overexpressing rice and wildtype Zhonghua 11 grown under the 06N, 0.16N, 16N and 56N conditions were randomly selected for the determination of the leaf SPAD value. A chlorophyll meter (SPAD-502) was used to test the SPAD value of the upper, middle and bottom portion of the flag leaf of each plant according to the manufacturer's instructions, and the average mean was used in the data analysis. At the heading stage, every ten plants of the GS1;1, GS1;2-overexpressing rice and wildtype Zhonghua 11 grown under the 06N, 0.16N, 16N and 56N conditions were randomly selected for the determination of the photosynthetic parameters. A Li-6400XT portable photosynthesis system (USA, Li-COR) was used to test the photosynthetic rate, stomatal conductance, intercellular CO 2 concentration and transpiration rate of the upper, middle and bottom portion of the flag leaf of each plant according to the manufacturer's instructions, and the average mean was used in the data analysis.

Determination of the Physiological Parameters
For the soluble protein and carbohydrate analysis, three samples of randomly mixed plant root, stem and leaf materials from three biological replications were harvested at the tillering stage and heading stage. The plant materials were homogenized by grinding the freshly harvested leaves on ice with extraction buffer [10 mM Trizma (pH 7.5), 5 mM sodium glutamate, 10 mM MgSO 4 , 1 mM dithiothreitol, 10% (v/v) glycerol, and 0.05% (v/v) Triton X-100]. The homogenates were then centrifuged at 12,000 g for 20 min at 4uC [46]. The soluble protein concentration of the extract was measured using the Bradford [47] protein assay and Coomassie Plus Protein Assay Reagent (Pierce, Rockford, IL, USA); bovine serum albumin was used as the standard protein.
The soluble carbohydrates were extracted from pre-dried plant materials with boiling water and colorimetrically measured according to the anthrone procedure [48,49]. For the metabolite profiling analysis, samples of six randomly mixed plant root and leaf materials of the GS1;1, GS1;2-overexpressing rice and wildtype Zhonghua 11 were harvested at the tillering stage and analyzed using the GC-TOF-MS method. Extracts from 200 mg FW samples were used. The data pre-treatment and normalization, alignments and metabolite identification were performed as previously described by Kusano et al. [50] and Redestig et al. [51].

Gene Expression Analysis
For gene expression analysis, both root and leaf materials of the GS1;1, GS1;2-overexpressing rice and wildtype Zhonghua 11 were harvested from three biological replications under 06N, 0.16N, 16N and 56N conditions at the tillering stage. Total RNA was extracted with TriZol reagent (Invitrogen, Germany) according to the manufacturer's instructions. For q-RT PCR analysis, firststrand cDNAs were synthesized from DNaseI-treated total RNA using Superscript II reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Q-RT PCR was performed in an optical 96-well plate with an ABI PRISM 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA). Each reaction contained 12.5 ml of 26SYBR Green Master Mix reagent (Applied Biosystems), 3.0 ml of cDNA, and 200 mM each of the gene-specific primers in a final volume of 25 ml. The thermal cycle used was as follows: 95uC for 3 min followed by 45 cycles of 95uC for 30 s, 60uC for 30 s, and 72uC for 40 s. All gene-specific primers for q-RT PCR are designed on the basis of the cDNA sequences and listed in Table S1. The specific primer for the rice actin gene (NM_197297) was used as an internal control. The primers were designed using Primer Express Software (Foster City, CA, USA) and checked using the BLAST program with the rice genomic sequence available in the database of the Institute for Genomic Research (TIGR, http://rice.plantbiology.msu.edu/) to ensure that the primers would amplify a unique and desired cDNA segment. The specificity of the reactions was checked by melting curve analysis, and three replicates of each cDNA sample were used for q-RT PCR analysis.

Growth Phenotype and Yield Analysis in GS1overexpressing Plants
In our previous study, we obtained GS1;1and GS1;2overexpressing rice plants using the Agrobacterium-mediated transformation method, and we failed to observe obvious differences in the growth phenotype between transgenic plants and wildtype plants at the seedling stage when grown hydroponically under normal and low nitrogen conditions [43]. To determine whether the accumulated mRNA transcript level of GS1;1 or GS1;2 will affect the development and productivity in transgenic plants, we analyzed the root length, plant height, root and shoot dry weight, leaf SPAD value and photosynthetic parameters of GS1;1-, GS1;2overexpressing plants and wildtype plants at the tillering stage and heading stage grown hydroponically under four different nitrogen levels (06N, 0.16N, 16N and 56N). The yields of GS1;1-, GS1;2overexpressing plants and wildtype plants at the mature stage growth in pots were also tested.
Our results showed a significant (P,0.05) decrease in the root length and plant height, root and shoot dry weight in GS1;1-, GS1;2-overexpressing plants when compared to wildtype plants at both the tillering and heading stages under the 06N, 0.16N, 16N and 56N conditions ( Fig. 1 1B).
During the leaf SPAD value determination, compared to wildtype plants, there were significant (P,0.05) decreases in GS1;1-overexpressing plants at the heading stage under the 16N (4.5% decrease) and 56N (5.8% decrease) conditions, while there were significant (P,0.05) increases in GS1;2-overexpressing plants at the tillering stage under the 06N (7.0% increase) and 0.16N (5.7% increase) conditions (Table 1). During photosynthesis analysis at the heading stage, there were significant (P,0.05) changes in GS1;1-overexpressing plants, while no significant (P, 0.05) changes in GS1;2-overexpressing plants were observed when compared to wildtype plants. For example, there was a 12.4% decrease in intercellular CO 2 concentration and 21.5% decrease in transpiration rate under the 0.16N condition, a 36.6% increase in stomatal conductance and 18.0% increase in transpiration rate under the 16N condition, and 34.0% increase in stomatal conductance and 32.6% increase in transpiration rate under the 56N condition in GS1;1-overexpressing plants compared to wildtype plants at the heading stage (Table 1). However, no significant (P,0.05) changes in photosynthetic rate were observed in GS1;2 transgenic plants (Table 1).
During the yield analysis at the mature stage, compared to wildtype plants, significant (P,0.05) decreases in both GS1;1and GS1;2-overexpressing plants were observed, except the yield in  (Table 2). In addition, the yield components (panicles/plant, filled grains/panicle, seed setting rate   Table 2). These results suggested that the accumulated mRNA transcriptional levels of GS1;1 and GS1;2 affected the plant growth at the tillering and heading stages and also seriously affected productivity at the mature stage.

Nitrogen Uptake Assay of GS1-overexpressing Plants
Because glutamine synthetase is the main enzyme involved in nitrogen assimilation, we analyzed the contents of 15 N and total N in the root, stem and leaves of GS1;1-, GS1;2-overexpressing plants and wildtype plants using the 15 N tracer assay at the tillering stage to test the differences in the nitrogen uptake ability in the root, and nitrogen transport ability from the root to the stem and from the stem to leaf between GS1;1-, GS1;2-overexpressing plants and wildtype plants. Our results showed that both GS1;1and GS1;2overexpressing plants demonstrated lower 15 N contents in the root, stem and leaf, particularly at 3 d after NH 4 NO 3 in the nutrient solution when refreshed by 15 NH 4 15 NO 3 (Fig. 2). Interestingly, similar 15 N contents in the root, while different 15 N contents in the stem and leaf between GS1;1and GS1;2overexpressing plants were observed at 3 d after NH 4 NO 3 in the nutrient solution when refreshed by 15 NH 4 15 NO 3 (Fig. 2). Furthermore, we analyzed the total carbon and nitrogen contents in the root, stem and leaves of GS1;1-, GS1;2-overexpressing plants and wildtype plants under the 06N, 0.16N, 16N and 56N conditions at this stage. Due to the two independent assays, the plant materials were different from the materials used in the 15 N tracer assay. Our results showed significant (P,0.05) decreases in the carbon/nitrogen ratio, specifically in the stem, in the GS1;1-, GS1;2-overexpressing plants when compared to wildtype plants. For the GS1;1-overexpressing plants, there were 22.3% and 9.6% decreases in the stem carbon/nitrogen ratio under the 0.16N and 16N conditions, respectively, which was caused by 22.1% and 9.4% higher stem nitrogen contents under the 0.16N and 16N conditions, respectively (Table 3). Similar results were also observed in GS1;2-overexpressing plants, in which there was a 9.9% decrease in the leaf carbon/nitrogen ratio under the 06N condition, and 13.7% and 12.3% decreases in stem carbon/ nitrogen ratio under the 16N and 56N conditions, respectively, which was caused by a 10.3% higher leaf nitrogen content under the 06N condition, and 18.0% and 12.2% higher stem nitrogen contents under the 16N and 56N conditions, respectively (Table 3).

Soluble Proteins and Carbohydrates Determination of GS1-overexpressing Plants
Because the nitrogen and carbon contents are altered by the over-expression of GS1;1, GS1;2 genes as previously mentioned, we determined the concentrations of soluble proteins and carbohy-  16N and 56N conditions to evaluate the differences in the carbon and nitrogen metabolic status between GS1;1-, GS1;2-overexpressing plants and wildtype plants. Our results showed that most of the soluble proteins were present in the leaf, while most of the soluble carbohydrates were present in the stem. Interestingly, the soluble proteins increased from a low nitrogen level to a high nitrogen level, while the opposite pattern was observed in the soluble carbohydrates, which decreased from a low nitrogen level to a high nitrogen level at both the tillering and heading stages (Fig. 3, 4) 06N, 0.16N, 16N and 56N conditions, respectively (data not shown here) were observed. Furthermore, there were many changes in the concentrations of the root, stem and leaf soluble proteins and carbohydrates in GS1;1-, GS1;2-overexpressing plants when compared to wildtype plants (Fig. 3, 4). During the analysis of soluble proteins, for GS1;1-       (Fig. 4). These results suggested that over-expression of GS1;1, GS1;2 not only changed the concentrations of soluble proteins and carbohydrates in transgenic plants but also altered their distribution in the root, stem, leaf and spikelet, which affected the carbon and nitrogen metabolic status.

Metabolic Profiling Analysis in GS1-overexpressing Plants
To further study the individual metabolite involved in the carbon and nitrogen metabolic pathway in detail, we analyzed the sugars, organic acids and free amino acids in the root and leaves of GS1;1-, GS1;2-overexpressing plants and wildtype plants at the tillering stage under the 0.16N and 16N conditions. As shown in Figure 5 and Figure S1, the fold change corresponds to the ratio of GS1;1-, GS1;2-overexpressing plants/wildtype plants calculated by the concentration of these individual metabolites. For the total sugars, total organic acids and total free amino acids, the results showed similar patterns between the GS1;1and GS1;2-overexpressing plants, particularly under the 0.16N condition. The total sugars, total organic acids and total free amino acids increased in the leaf and decreased in the roots of both GS1;1and GS1;2overexpressing plants under the 0.16N condition (data not shown here). Under the 16N condition, the total sugars, total organic acids and total free amino acids increased in the leaves of both GS1;1and GS1;2-overexpressing plants, while opposite results were found in the root between the GS1;1and GS1;2overexpressing plants, in which the total sugars and total organic As shown in Figure 5 and Figure S1, which displayed the fold change of the individual sugars, organic acids and free amino acids in GS1;1-, GS1;2-overexpressing plants compared to wildtype plants, we observed alterations in these metabolites due to the over-expression of the GS1;1, GS1;2 gene. In general, larger variations were found in the leaf compared to the root. In the leaf, larger variations of individual sugars and organic acids were found under the 0.16N condition, while larger variations in the individual free amino acids were found under the 16N condition (Fig. 5, Fig. S1). In the root, larger variations of these individual metabolites were found under the 16N condition compared to the 0.16N condition (Fig. 5, Fig. S1). In the leaves of GS1;1overexpressing plants, when compared to wildtype plants, dramatic increases in fructose (.10.8-fold), xylitol (.179.4-fold), succinate (.70.4-fold) and methionine (.166.8-fold) and dramatic decreases in benzoic acid (,0.02-fold) were observed under the 0.16N condition, In contrast, dramatic increases in glutamine (. 12.7-fold), alanine (.39.4-fold) and valine (.31.4-fold) and dramatic decreases in xylitol (,0.01-fold) and aspartate (,0.08fold) were observed under the 16N condition (Fig. 5, Fig. S1). In the roots of GS1;1-overexpressing plants, when compared to wildtype plants, dramatic decreases in ascorbic acid (,0.02-fold), leucine (,0.001-fold) and cysteine (,0.01-fold) were observed under the 0.16N condition, while dramatic increases in benzoic acid (.52.3-fold), methionine (.16.4-fold) and alanine (.89.7fold) and a dramatic decrease in succinate (,0.09-fold) were observed under the 16N condition (Fig. 5, Fig. S1). In the leaves of GS1;2-overexpressing plants, when compared to wildtype plants, dramatic increases in xylitol (.202.8-fold) and succinate (.391.1fold), and dramatic decreases in benzoic acid (,0.08-fold) and alanine (,0.01-fold) were observed under the 0.16N condition, while dramatic increases in methionine (.48.1-fold) and alanine (.11.3-fold) were observed under the 16N condition (Fig. 5, Fig.  S1). In the roots of GS1;2-overexpressing plants, when compared to wildtype plants, a dramatic increase in aminobutyric acid (. 93.8-fold) and a dramatic decrease in phenylalanine (,0.04-fold) were observed under the 0.16N condition, while dramatic increases in pyruvate (.708.3-fold), aminobutyric acid (.69.9fold), leucine (.14.4-fold) and phenylalanine (.14.8-fold) and dramatic decreases in succinate (,0.04-fold) and glutamine (, 0.03-fold) were observed under the 16N condition (Fig. 5, Fig. S1).

Gene Expression Analysis in GS1-overexpressing Plants
To analyze the effect of higher GS1;1, GS1;2 mRNA transcriptional levels on the expression patterns of key genes involved in the carbon and nitrogen metabolic pathway, the expression level of genes encoding NRT (nitrate transporter), NR (nitrate reductase), GS (glutamine synthetase), GOGAT (glutamate synthase), RU-BISCO (ribulose-1,5-bisphosphate carboxylase/oxygenase) and PEPC (phosphoenolpyruvate carboxylase) were analyzed using q-RT PCR. These genes are displayed in the carbon and nitrogen metabolic pathway in rice plants (Fig. 6A). Table S2 lists the fold change corresponding to the ratio of GS1;1-, GS1;2-overexpressing plants/wildtype plants calculated by the relative gene expression level in the root and leaf at the tillering stage under the 06N, 0.16N, 16N and 56N conditions. Our results showed the distinct expression patterns of these genes displayed between the GS1;1and GS1;2-overexpressing plants under different nitrogen levels.

Cytosolic GS1 Genes Play Important Role in Plant Growth and Grain Filling
In higher plants, GS/GOGAT cycle is the first step of incorporation of inorganic nitrogen into organic nitrogenous compounds, which is a major checkpoint for controlling nitrogen assimilation. Because of this important function of GS enzyme in plant nitrogen metabolism, particular attention has been devoted to studies on GS transformation in higher plants, which is expected to be a good molecular method used to analyze gene functions and a good strategy to improve nitrogen use efficiency. In our previous study, the full length cDNAs of cytosolic GS1;1 and GS1;2 in rice were constructed into pCAMBIA 1301S vector which driven by the CaMV 35S promoter; and the GS1-overexpressing transgenic plants were obtained. The fresh weight and dry weight of GS1;1-, GS1;2-overexpressing plants were not significantly different than those of wildtype plants at the seedling stage [43]. However, the GS1;1-, GS1;2-overexpressing plants showed significantly decreased plant height and root length, shoot dry weight and root dry weight at both the tillering stage and heading stage when compared to the wildtype plants. These results indicated a high expression level of cytosolic GS1 gene was negatively correlated with plant biomass. Similar results were also observed in GS1overexpressing Lotus japonicus and Pea, which exhibited negative effect on plant biomass [35,36]. In addition, several studies highlighted the importance of cytosolic GS1 genes in determining grain filling in cereal crops. For example, positive correlations were showed between grain number/size and cytosolic GS protein content/GS activity in rice [24,52,53], maize [38,54,65] wheat [34]. However, in our study, a negative correlation was observed between the cytosolic GS1 gene expreesion level and the grain filling in GS1;1-, GS1;2-overexpressing plants. Compared to the wildtype plants, the yields of GS1;1-, GS1;2-overexpressing plants were significantly declined. Similar results were reported in the previous study that GS1;1-, GS1;2-overexpressing plants showed decreased yield and amino acids concentrations when grown under the low nitrogen field [43].

Stable Carbon/nitrogen Ratio Plays an Important Role in Plant Development
In higher plants, carbon and nitrogen are crucial for routine and fundamental cellular activities. Plant growth and development are highly dependent on the interaction between carbon and nitrogen metabolism. The coordination and optimal functioning of the metabolic pathways for nitrogen and carbon assimilation in plants are critical in determining plant growth, biomass accumulation and, ultimately, yield [5,7,8,10]. Due to developmental changes and environmental demands on carbon and nitrogen resources, plants have evolved as a highly sophisticated and complex sensory system to regulate carbon and nitrogen assimilation, metabolism and transport [10]. Studies at the molecular level have revealed an even higher complexity, with multiple levels of regulation, including post-transcriptional control by microRNAs [55][56][57][58]. Recent reports have also shown that the adaptation of maize source leaf metabolism to stress is related to disturbances in the carbon, nitrogen and phosphorus balance [59]. In the present study, significant (P,0.05) decrease of the carbon/nitrogen ratio were shown in GS1;1-, GS1;2-overexpressing plants, particularly in the stem. Furthermore, the concentrations of soluble proteins and carbohydrates changed their distributions in various organs, such as the root, stem, leaf and spikelet, which affected the carbon and nitrogen metabolic status. In addition, variations in the leaf SPAD value, photosynthetic parameters, individual metabolites and gene expression levels indicated an imbalance of carbon-nitrogen metabolism, which may cause the poor growth and yield in the GS1;1-, GS1;2-overexpressing plants. An imbalance in levels of sugars, amino acids and metabolites in the tricarboxylic acid cycle was also reported in the rice OsGS1;1 mutant by Kusano et al., which revealed a crucial function of OsGS1;1 in coordinating the global metabolic network in rice plants [37]. Previous studies have reported similar results in which the carbon to nitrogen ratio, rather than the carbohydrate status alone, plays a predominant role in regulating various aspects of Arabidopsis seedling growth, including the cotyledon size, fresh weight, chlorophyll and anthocyanin content, storage reserve mobilization and photosynthetic gene expression [9].

GS1;1 and GS1;2 Genes have Different Functions in Rice Carbon and Nitrogen Metabolism
In higher plants, multiple homologous but distinct genes were found for cytosolic GS1 [14,[60][61][62][63][64][65], and three members have been identified in rice (Oryza sativa; OsGS1;1, OsGS1;2 and OsGS1;3;) [2,20,24]. OsGS1;1 was expressed in all organs (i.e., root, leaf blade, leaf sheath, and spikelet), with a higher expression in the leaf blade and shoot phloem during the vegetative stage and thus functions in translocating nitrogenous compounds to the developing sink tissues [24]. OsGS1;2 transcripts were also detected in all organs, with a higher expression in the root at the seedling stage, and this protein mainly functions in primary NH 4 + assimilation [2]. Although similar results of growth phenotype, yield and carbon/nitrogen ratio were found in both GS1;1and GS1;2overexpressing plants in this study, the different exhibitions of the leaf SPAD value, photosynthetic parameters, soluble proteins and carbohydrates, metabolites and gene expression levels were observed between the GS1;1and GS1;2-overexpressing plants, suggesting different roles of GS1;1 and GS1;2 genes in rice nitrogen metabolism. In addition, different profiles of individual metabolites and gene expression levels were found between GS1;1and GS1;2overexpressing plants, particularly when sufficient nitrogen was applied in the environment. Our previous study also suggested different functions between the GS1;1 and GS1;2 genes in rice nitrogen metabolism and abiotic stress responses. Basta resistance was observed in GS1;2-overexpressing rice plants at three developmental stages (seed germination, two-leaf seedling, and mature plant stages) [43]. However, Basta resistance was not observed in GS1;1-overexpressing rice plants at any developmental stage [43]. In addition, analysis of gln1-3 and gln1-4 mutants in maize also indicated distinct roles of cytosolic glutamine synthetase genes in which the gln1-4 mutant exhibited a reduced kernel size and the gln1-3 mutant exhibited a reduced kernel number [38].

Conclusion
In this study, we systematically analyzed the differences in the growth phenotype, carbon-nitrogen metabolic status and gene expression profile between GS1;1-, GS1;2-overexpressing rice plants and wildtype plants. From the results, we found that significant alterations in the carbon and nitrogen metabolism were displayed by the overexpressing of GS1;1 or GS1;2 in the transgenic plants. For example, under the 16N condition, compared to wildtype plants, most of the carbon-nitrogen metabolic genes were significantly (P,0.05) induced in GS1;1overexpressing plants (NRT1;1, NRT1;2, NRT2, NR2, GS1;1, GS1;2, Fd-GOGAT1 and NADH-GOGAT1 genes in the roots, and GS1;1, NADH-GOGAT2, RUBISCO, PEPC2, PEPC3 and PEPC4 genes in the leaves), which directly resulted the increasing of total sugars, total organic acids and total free amino acids in the leaves, the increasing of total sugars and total organic acids and the decreasing of the total free amino acids in the roots. These changes of metabolites caused the decline of soluble proteins and the increase of soluble carbohydrates in the GS1;1-overexpressing plants. While in the GS1;2-overexpressing plants, most of the genes were significantly (P,0.05) induced (including NRT1;1, NRT1;2, NR2 and GS1;2 genes in the roots, NR1, GS1;2, NADH-GOGAT2, PEPC3, PEPC4 and PEPC7 genes in the leaves), except for the GS1;1, GS2, RUBISCO and PEPC6 genes, which were significantly (P,0.05) decreased in the leaves. This type of gene expression profile directly resulted the increasing of total sugars, total organic acids and total free amino acids in the leaves, the decreasing of total sugars and total organic acids and the increasing of the total free amino acids in the roots of GS1;2-overexpressing plants. These changes of metabolites caused the increases of both soluble proteins and carbohydrates in the GS1;2-overexpressing plants. In addition, the significantly (P,0.05) decreased carbon/nitrogen ratio was observed in the stem of both GS1;1and GS1;2overexpressing plants. We concluded that the imbalance in carbon and nitrogen metabolic status may be the main reason for the decreased growth and yield in GS1;1-, GS1;2-overexpressing plants. Gal, galactose; Glu, Glutamate; Gln, Glutamine; Arg, Arginine; Pro, Proline; Orn, Ornithine; GABA, Aminobutyric; Asp, Aspartate; Asn, Asparagine; Ile, Isoleucine; Met, Methionine; Thr, Threonine; Ala, Alanine; Val, Valine; Leu, Leucine; Phe, Phenylalanine; Try, Tryptophan; Ser, Serine; Gly, Glycine; Cys, Cysteine.

(DOC)
Table S1 Primer sequences of the key genes involved in the carbon and nitrogen metabolism used in qRT-PCR. (DOC)

Table S2
Fold change corresponding to the ratio of the gene expression level in the GS1;1-, GS1;2-overexpressing plants relative to the wildtype plants for the roots and leaves at the tillering stage under the 06N, 0.16N, 16N, 56N conditions. (DOC)