Over-Expression of a Tobacco Nitrate Reductase Gene in Wheat (Triticum aestivum L.) Increases Seed Protein Content and Weight without Augmenting Nitrogen Supplying

Heavy nitrogen (N) application to gain higher yield of wheat (Triticum aestivum L.) resulted in increased production cost and environment pollution. How to diminish the N supply without losing yield and/or quality remains a challenge. To meet the challenge, we integrated and expressed a tobacco nitrate reductase gene (NR) in transgenic wheat. The 35S-NR gene was transferred into two winter cultivars, “Nongda146” and “Jimai6358”, by Agrobacterium-mediation. Over-expression of the transgene remarkably enhanced T1 foliar NR activity and significantly augmented T2 seed protein content and 1000-grain weight in 63.8% and 68.1% of T1 offspring (total 67 individuals analyzed), respectively. Our results suggest that constitutive expression of foreign nitrate reductase gene(s) in wheat might improve nitrogen use efficiency and thus make it possible to increase seed protein content and weight without augmenting N supplying.


Introduction
Wheat (Triticum aestivum L.) is one of the most widely cultivated and most important food crops in the world, and its higher yield depends on heavy field-supply of nitrogen (N) fertilizer [1][2][3]. However, the N use efficiency of crops was low (approximately 33%) [4,5] and over 50% of the N applied was lost from the plantsoil system [6], leading to environmental damage and negative impacts on human health [7][8][9][10]. That was particularly pronounced in the areas along the Yellow River, Huai Rover and Hai River (called ''Huanghuaihai Area'') in central China [11] where is one of the major areas of wheat production but with saline and alkaline sandy soils and relatively lower yield. Nitrate (NO 3 2 ) is the main N source for crops under normal field conditions [9,12,13] and its availability strongly affects crop productivity and food quality [14], especially in wheat [15][16][17]. The nitrate up-taken in plant is well known to be first reduced to nitrite and then to ammonium via the Glutamate synthesis cycle (GOGAT cycle) in two successive steps catalyzed by nitrate reductase (EC 1.6.6.1, NR) and nitrite reductase (EC 1.7.7.1, NiR) in cytosol and chloroplast, respectively [18]. Thus, the NR is considered a key enzyme in the overall process of nitrate assimilation [19], and how to increase NR content and/or activity, therefore, becomes one of the major challenges for increasing N use efficiency in crops including wheat. Using biotechnology to introduce and over-express exogenous tobacco NR gene was tested for lowering nitrate content in the leaf and edible organs of dicotyledonous crops [20][21][22][23][24][25][26][27], but no information about the effect on seed protein content and grain weight was released. To the best of our knowledge, integration and overexpression of foreign NR gene have not been tested in wheat although its foliar NR activity was demonstrated significantly correlated with yield [17,28], flour quality [16] and grain protein content [15,17]. The purpose of the present work was to test whether or not introduction and expression of a foreign NR gene in wheat could increase N use efficiency and hence improve quality and/or yield without augmenting N supply, or could maintain quality and/or yield with a diminished use of N fertilizer. Our results demonstrated that over-expression of a CaMV 35Sdriven NR gene in two cultivated winter wheat cultivars remarkably enhanced foliar NR activity and significantly increased seed protein content and grain weight under normal soil N conditions.

Explants and Agrobacterium tumefaciens-mediated transformation
Two winter wheat (Triticum aestivum L) cultivars, ''Nongda146'' (ND146) and ''Jimai6358'' (JM 6358) which are widely cultivated in the ''Huanghuaihai Area'', China, were used throughout this study. Their immature embryos were isolated from the young caryopses 12-14 days after anthesis, and induced to produce embryogenic callus as previously described [29]. The calli were pretreated for 8-12 h on an osmotic medium with 0.4 M mannitol before Agrobacterium tumefaciens (strain LBA4404) inoculation. The LBA4404 harbored a binary vector pBCSL16 [21] which was kindly provided by Drs. Cabouche and Meyer (INRA, France). The vector carried a kanamycin-resistant gene (Npt II) and tobacco nitrate reductase cDNA (nia) which was functionally fused to CaMV 35S promoter and terminator. The inoculation and coculture of the pretreated calli with Agrobacterium were performed as previously reported [29].

Selection and regeneration of G418-resistant wheat plants
After co-culture, the calli were subcultured, G418-resistance selected and shoot-regenerated, and the regenerated green shoots rooted as previously described [29] except that G418 (Geneticin, an aminoglycoside antibiotic similar in structure to gentamicin B1; 25 mg/L) instead of PPT was used as the selective agent. The plantlets were vernalized for 2 weeks at 4uC, and then transplanted in pots in greenhouse and self-fertilized to produce T 1 seeds. During greenhouse stage, one young leaf from each independent T 0 transformant and WT was sampled for PCR verification.

Screening and cultivation of kanamycin-resistant T 1 plants
Screening of kanamycin-resistant (Kan-R) T 1 plants was conducted according to Xi and co-workers [30] and Zhang et al. [31] with slight modification. Briefly, the Kan tolerant threshold of WT (ND146 and JM6358) was first determined. The seeds were germinated in a set of Kan concentration (0, 40, 60, 80, 100, 120, 160 or 200 mg/L) at room temperature, and the seedlings were transferred into vermiculite-containing Petri dishes, irrigated with corresponding concentration of Kan and vernalized for 2 weeks at 4uC. After vernalization, the cultures were irrigated with water and placed under the conditions of 2561uC and 16/8 h (light/dark) photoperiod of ca. 3000 lux. About one week later, green and white seedlings were accounted for each Kan concentration, and the lowest Kan concentration that resulted in more than 90% white seedlings was chosen as the threshold. In order to select T 1 transformant, the T 1 seeds were germinated and seedlings were selected as WT except with Kan at the threshold concentration. The green seedling was considered Kan-R.
The Kan-R T 1 plants were further verified by PCR, and then transplanted in flowerpots (14616.5 cm) together with untransformed control (WT) in greenhouse, one plant per pot. All pots contained equal quantity of the nutrient soil (1 vermiculite: 3 garden nutrient soil) and were randomly placed in an experimental plot with normal field managements.

Southern blot analysis
Southern blot analysis of PCR products was used to verify PCRpositive T 0 transformants, and both PCR and Southern blot to identify T 1 progeny.
For Southern blotting of PCR products, the PCR was run as described above with the genomic DNA from PCR-positive T 0 transformants as template. PCR products were separated by electrophoresis in 0.8% (w/v) agarose gel. For Southern identification of T 1 progeny, about 30 mg of genomic DNA from T 1 individuals or the control (WT) were digested at 37uC for 12 h with Nde I that has no recognized site in the T-DNA region of pBCSL16. The digested DNA was fractionated in 0.8% (w/v) agarose gel by electrophoresis run at 22 V for approximately 8 h. The PCR DNA and fractionated DNA were then transferred onto positively charged Hybond TM -N+ nylon membrane (Amersham Pharmacia Biotech) by capillarity and fixed by UV cross-linking. The membranes were hybridized using the probe of npt II+nos fragments that were labeled with digoxigenin using the random primer labeling kit (DIG DNA Labeling and Detection Kit). Prehybridization, hybridization and detection of the probe were carried out using a non-radioactive, DIG Luminescent Detection Kit for Nucleic Acids (Roche Diagnostics) according to the manufacturer's instructions.

Determination of nitrate reductase activity
The nitrate reductase activity (NRA) was measured in vivo according to Freschi et al. [34] with slight modification. Real and potential NRAs were those measured without and with KNO 3 induction, respectively. The fresh flag leaf at grain filling period was collected between 9:00 and 10:00 a.m. from greenhousegrown wheat, and cut into equal two parts along the main vein. One part was used for measurement of real NRA and another part, for potential NRA. To measure the potential NRA, the sample was first induced in 50 mM KNO 3 for 12 h at 25uC under light of 3000 lux, and then vacuum-infiltrated. For vacuuminfiltration, leaf samples (0.2 g fresh weight) with or without KNO 3 induction were cut into pieces (0.5-1 cm 2 ), immersed in an incubation buffer (5 ml phosphate buffer (pH 7.5) + 5 ml 0.2 M KNO 3 solution), vacuum-infiltrated 3-4 times, each for 20 min, and then incubated in darkness for 30 min at 30uC. After infiltration, the nitrate reduction was carried out at room temperature for 30 min in a reaction mixture containing 1 ml of sample infiltrate, 1 ml of 1% (w/v) sulfanilamide in 36% HCl and 1 ml of 0.2% (w/v) 1-naphthylamine. The nitrite (NO 2 2 ) formed was detected spectrophotometrically at 540 nm, and the NRA was expressed inmg of nitrite (NIR) produced per hour and per gram of fresh leaf. The experiment was triplicated.

Measurement of nitrate contents
The foliar nitrate content was determined according to Cataldo et al. [35] slightly modified. Leaf segments were dried at 85uC until constant weight. The dried material (25 mg) was grounded to powder and then incubated in 10 ml of distilled water for 2.5 h. Aliquots of 0.1 ml were mixed thoroughly with 0.4 ml of 5% (w/v) salicylic acid in concentrated H 2 SO 4 . After 20 min incubation at room temperature, 9.5 ml of 2 M NaOH were added. The samples were cooled to room temperature and nitrate concentration determined spectrophotometrically by measuring the absorbance at 410 nm.

Protein and 1000-grain weight analysis of T 2 seeds
At harvest, the seeds from 57 T 1 -individuals with good seedsetting rate were chosen for determination of 1000-grain weight and protein content. To determine 1000-grain weight, 15 seeds per individual plant were picked up randomly and weighted. For detecting protein content, the seeds were first dried at 40uC to constant weight, and then milled and sieved (100 mesh). The protein content of the flour was blindly measured by a commercial company using Kjeldahl method with a continuous flow analyzer (Auto Analyzer 3 Bran + Luebbe, Germany) on three replicates, and calculated by using a conversion factor of 5.7.

Statistical analysis
Data of NRA, nitrate content, protein content and grain weight were analyzed by analysis of variance (ANOVA) followed by Duncan's multiple test and T-test with SPSS 17.0 software (SPSS Inc, Chicago, IL, USA).

PCR and Southern blot identification of T 0 transformants
Among independent G418-resistant T 0 transformants, 8 and 53 individuals from JM6358 and ND146 had one expected band of about 740 bp in the PCR product (Parts shown in Figs. 2a & 2b). This gave a transformation efficiency of 0.4% (8/2000) and 1.6% (53/3300), respectively, based on the number of PCR-positive plants/number of the embryogenic calli trans-infected. When Southern blotted, all PCR-positive products and the vector plasmid had a clear hybridized band, whereas no such a band appeared from untransformed control plant (Fig. 2c & 2d).
The T 1 green seedlings were further verified by PCR. Overall 71.6% and 70.6% T 1 green seedlings of NR-ND146 (225 plants) and NR-JM6358 (85 plants) were PCR-positive (PCR+), respectively (Table 1), but none of the albino seedlings from two families were PCR+ (Data not shown). As presented in Table 1, in NR-ND146 family 7 out of 9 lines had a ratio of 1 : 1 of the Kan-R : PCR+ individual, whereas in the family NR-JM6358, this ratio was only noted in 1 of 7 lines.

Southern blot analysis of T 1 offspring
The presence of the transgene in PCR+ T 1 progeny was further verified by Southern blot analysis. In 8 PCR+ individuals randomly picked (4 from NR-ND146 and 4 from NR-JM6358), the hybridizing band was clearly present, and the band number varied from 1 to 5, with fewer bands in the individuals from NR-ND146 family (lanes 1-4) than in those from NR-JM6358 (lanes 5-8) (Fig. 4). Real and potential NR activities of T 1 progeny The foliar NRA was significantly enhanced by 50 mM KNO 3 induction, and this increment took place both in WT and T 1 progeny (Fig. 5). Compared with WT, the T 1 offspring of NR-ND146 had a significant higher NRA in 5 of 7 individuals tested (146-50-5, 146-90-87, 146-90-110, 146-90-189 and 146-95-4), no matter with or without KNO 3 inducement (Fig. 5a). However, in NR-JM6358 descendants, all tested T 1 individuals from 5 T 0 lines displayed remarkably stronger NRA than WT when induced with

Nitrate content of T 1 plants
In NR-ND146 family, the foliar nitrate content of 15 T 1 individuals varied from 2.67 to 44.67 mg/g FW. The lowest, detected in plant 146-95-4, was 14.2% of the WT (18.9 mg/g FW) and the highest, in plant 146-90-110, approximately 2.4-fold of the WT (Fig. 6a). Among 15 T 1 individuals, 8 plants displayed significantly lower nitrate content than WT, but 5 plants, higher than WT (Fig. 6a).

1000-grain weight and protein content of T 2 seeds
Mature T 2 seeds were collected from 27 individuals of 9 T 1 NR-ND146 lines and 30 individuals of 7 T 1 NR-JM6358 lines. Among them, 3 individuals from NR-ND146 family and 4 from NR-JM6358 family had the flag leaf sampled for NRA and NO 3 2 content determination. In order to exclude the influence of leafsampling on seed protein content and grain-weight, we analyzed the T 2 seeds of the plants whose leaves sampled in one group and the those with intact leaf, in another one.
1000-grain weight. In all leaf-sampled T 1 plants of NR-ND146 family, the T 2 seeds had a very significant higher 1000grain weight than WT, whereas in 4 leaf-sampled T 1 individuals of NR-JM6358 family, only two displayed such a significant increment, and the rest, much modest ( Table 2).
Among 24 leaf-intact T 1 plants of NR-ND146 family, 70.8% increased their seed protein content in comparison with WT (protein content: 19.08%), with an increment range of more than 30% in 33.3% individuals and 20%-30% in 25% individuals. The highest seeds protein content reached at 31.49% (plant 146-93-3) which is 1.65 times of WT.
In NR-JM6358 family, 4 out of 26 leaf-intact T 1 plants had the protein content higher than the WT (21.7%) by over 5%, 8 T 1 individuals by 2%-5%, but about one half of individuals even declined their seed protein content, more or less (Table 2).

Discussion
Transformation and regeneration of cultivated winter wheat and rapid screening of T 1 transformants Although the first report on successful transformation and regeneration of wheat mediated by Agrobacterium tumefaciens was reported in 1997 by Cheng et al. [36], most reported transformation events were still limited to some ''model'' spring-type cultivars such as ''Bobwhite'' and ''Chinese Spring'' [37,38]. We successfully transferred a tobacco nitrate reductase gene (Nia2) into two commercially cultivated winter wheat cultivars, ''ND146'' and ''JM6358'' with Agrobacterium-mediation and obtained numbers of fertile transgenic plants (Figs. 1 & 2) following our protocol established [39] and improved [29,40,41]. We realized a transformation efficiency of 1.68% in ''ND146'' and 0.40% in ''JM6358'' based on the number of PCR-positive plants/number of calli inoculated.
After successful transformation and regeneration, we turned our attention to how to select transformants rapidly, efficiently and cost-effectively. In wheat as in other cereals, using hygromycin resistance gene was considered an effective selection system that allowed few escape plants to survive [42]. However, taking consideration of the existing biosafety/regulatory rules about genetically modified crops (GMC) and possible commercial cultivation of the transgenic wheat, we used Kan-R gene (npt II) in place of hygromycin-R one as the selection gene. We used G418 in the place of Kan as the selective agent at different in vitro stages of the transformation due to wheat's native resistance to Kan. Our results demonstrated that G418 at 25 mg/L was efficient for selecting npt II-transgenic calli, shoots and plantlets at corresponding stages of the transformation (Fig. 1c to 1h). Even so, we were aware that the G418 was much expensive than Kan, and its amount requested for ''field'' selection of T 1 and then-after offspring would be much more than the selection in vitro of T 0 transformants. In order to reduce the selection cost, we adopted the method of Xi and colleagues [30] and Zhang et al. [31] by germinating T 1 seeds in Kan solution in dishes and then choosing green seedlings. Our results showed that more than 70% of the green seedlings in both genotypes (161/225 in NR-ND146 and 60/85 in NR-JM6358) developed in 80 mg/L of Kan solution were PCR-positive (Table 1), and all albino seedlings was PCRnegative (data not shown). The similar results were reported by Zhang et al. [31] and Ren et al. [43] in other genotypes of transgenic wheat. This indicated that npt II could be used as selectable marker gene in wheat transformation and Kan was efficient and cost-effective to screen primarily the T 1 offspring and then-after generations. Nitrate reductase activity and nitrate content in the flag leaf of T 1 progeny In untransformed wheat, the nitrate reductase activities (NRA) of the leaf tissues [15,16], basipetal part of the youngest ligule emergent leaf [28], third leaf [44], flag leaf [17] and even shoots [45] were found to be correlated more or less with yield and/or grain (flour) quality. We used the flag leaf for determining NRA and nitrate content of T 1 progeny, because its NRA was significantly correlated with both yield and grain protein content in winter wheat [17].
Over-expression of 35S-NR gene remarkably enhanced foliar NRA in more than 70% of T 1 descendants analyzed in NR-ND146 and NR-JM6358 families, and a maximum increment level reached at 3.46 times and 4.08 times of the WT, respectively (Fig. 5). This kind of NRA-increment was also reported in 35S- NR-transgenic dicotyledonous crops, such as in tobacco [20], Arabidopsis [46], lettuce [21], Chinese cabbage (Brassica. campestris L. ssp. pekinensis) and pakchoi (B. campestris L. ssp. chinensis) [22]. Our data showed that without NO 3 2 inducement, the NRA of T 1 progeny was T 0 parent line-dependent, and the different T 1 individuals from one single T 0 line had also remarkably varied NRA (Fig. 5). Under NO 3 2 inducement, both WT and T 1 plants enhanced their leaf NRA, but the increment was much more pronounced in T 1 plants than in WT (Fig. 5), with a maximal 3.9fold and 6.2-fold increment in the T 1 offspring of the family NR-ND146 (Fig. 5a) and NR-JM6538 (Fig. 5b), respectively. This implied that both endogenous and transgenic NR genes were nitrate-inducible, at least, in wheat, although the transgene NR was driven by constitutive promoter 35S.
Over-expression of 35S-NR gene significantly declined leaf nitrate content in 53.3% (8/15) to 76.5% (13/17) of T 1 individuals of NR-ND146 and NR-JM6358 families, respectively, with a maximal decrement of 78.6% (plant 6358-14-4) to 85.9% (plant 146-95-4) (Fig. 6). Such decrement of foliar nitrate content was also observed in numbers of NR-transgenic dicotyledonous crops: such as in tobacco [20,[47][48][49], lettuce [21], potato [23][24][25], Chinese cabbage and pakchoi [22,27]. It was well known that the NR, as a rate-limiting enzyme, catalyzed reduction of NO 3 2 into NO 2 2 , and thus logically over-expression of NR could decrease nitrate content in plant. Hu et al. [50] even speculated that the higher NRA was the more nitrate would be reduced. However, in our NR-transgenic wheat under the greenhouse growth conditions, the increment of foliar NRA was sometimes correlated with the nitrate decrement in the leaves of T 1 offspring of both NR-ND146 and NR-JM6358 families (Fig. S1). Sun et al. [51] reported that Arabidopsis plants transformed with a Chinese cabbage NR gene exhibited an enhanced level of both NO 3 2 and NRA in leaves under NO 3 2 inducement. Hoff et al. [52] also observed that Arabidopsis mutants affecting Nia2 and barley Nar1 mutants expressing only 10% of the WT NRA did not alter nitrate content and biomass under the greenhouse growth conditions. What is the reason remained to be investigated. In our transgenic wheat, the accumulated nitrate in the leaf would be used later for grain development.

Seed weight, protein content and their relationship with foliar nitrate reductase activity
Our data demonstrated that 70.8% (17/24) NR-ND146 and 50% (13/26) NR-JM6358 T 1 descendants had significant higher protein content than WT, and a more than 30% increment was detected in 33% of T 1 offspring in NR-ND146 family ( Table 2). For a limited number of leaf-sampled T 1 plants, the seed protein content looked like have a tendency of positive correlation with foliar NRA in both families (Fig. S2). In non-transformed spring wheat [15] and winter one [17], the foliar NRA was observed positively correlated with seed protein content. Kumari [17] thought that grain protein accumulation depended on the accumulation and partitioning of the reduced N accumulated during the vegetative stage and on the relative contributions of nitrate assimilation and N redistribution during grain development. In N-deficient wheat plants, lower shoot NRA resulted in decrement of reduced N accumulation daily in the shoots [45]. The plants grown in nitrate-rich conditions not only enhanced the activities of NR, ribulose bisphosphate carboxylase-oxygenase (RuBPCO) and glutamine synthetase etc. in growing and full expensed leaves, but also slowed the decrease of those activities in older leaves and delayed leaf senescence [44]. An increase in the supply of glutamine could enhance the rate of protein deposition in the wheat grain [53]. In 35S-NR-transgenic tobacco plants with higher foliar extractable NRA, Ferrario-Méry et al. [54] observed the increased glutamine level in the leaves. We suggested that the increased foliar NRA in 35S-NR-transgenic wheat might speed up nitrate assimilation and facilitate the N-flux to and/or N redistribution in seeds during grain development, and hence increased grain protein content.
We noted that in T 1 plant with intact leaf, 70.8% and 65.4% of them remarkably augmented their grain weight in NR-ND146 and NR-JM6358 families, respectively (Table 2). In order to know whether the increase in grain weight has some relationship with foliar NRA, we analyzed T 1 plants whose flag leaves were sampled for NRA and NO 3 2 determination. Our data showed that the correlation between grain weight and flag leaf NRA was varied with transgenic wheat families: R 2 = 1 in NR-JM6538, and R 2 = 0.4569 in NR-ND146 (Fig. S3). In untransformed wheat, the NRA in the basipetal part of the youngest ligule emergent leaf [28], flag leaf [17] and leaf tissues at boot stage of maturity [15] correlated well with yield. Kumari [17] observed that on induction of NRA by nitrate supply at post anthesis stage, the flag leaf retained the ability to synthesize RuBPCO. We suggested that constant expression of the NR gene in 35S-NR-transgenic wheat might help consistent synthesis of RuBPCO, and thus confer to the leaves longer and higher capacity of photosynthesis, and hence increase grain weight.
Our data indicated also that in 35S-NR-transgenic wheat, there was not obvious relationship between grain weight and seed protein content (Fig. 7), and in some T 1 individuals, the increment of grain weight was accompanied by the increase of seed protein content (Table 2). Jenner et al. [55] showed that the duration and rate of both starch and protein deposition in the endosperm of wheat were all independent events, controlled by separate mechanisms. Under N-rich growing conditions, both the duration and rate of starch deposition during grain filling were determined primarily by factors that worked close to or within the grain itself, whereas those of protein deposition were decided predominantly by factors of supply outside the grain. In studying the relationships between carbon and nitrogen metabolism in the leaves of NRtransgenic tobacco that expressed either a 5-fold increase or a 20fold decrease in NRA, Foyer and colleagues [56] concluded that large decreases in NRA had profound repercussions for photosynthesis and carbon partitioning within the leaf, but the increases in NRA had negligible effects. In Arabidopsis, over-expression of NR led to 200% increase of seedlings protein content without any gain in the fresh and dry weights [46]. We are aware that more works are needed to address the mechanism of the cell-and organspecific expression and metabolic regulation of NR gene and other genes involved in the nitrogen assimilatory pathway and to investigate the role of the enzymes in regulating flux through the nitrogen assimilation pathways, as indicated by Cullimore and Bennett [57].
We noted also, there was obvious variability in both seed protein content and grain weight among independent transformants and their progeny ( Table 2). Random insertion of the transgene in the genome of T 0 transformants and random recombination of the transgene in producing the progeny might be one of the explanations, because the insertion might change the expression of adjacent genes [58].
In conclusion, over-expression of 35S-NR gene in winter wheat significantly increased grain weight and seed protein content. This might be realized by an increased foliar NRA. The enhanced NRA might speed up nitrate assimilation and facilitate N-flux to and/or N redistribution in seeds during grain development in one hand, and make the leaf to have longer and higher capacity of photosynthesis, in other hands. Our results would provide an alternative way to breeding new wheat cultivars of higher protein content and higher nitrogen use efficiency, which makes it possible to reduce the need for excessive input of N fertilizers and improve or stabilize quality.