The nitrogen topdressing mode of indica-japonica and indica hybrid rice are different after side-deep fertilization with machine transplanting

Determination of the optimal fertilization method is crucial to maximize nitrogen use efficiency and yield of different rice cultivars. Side-deep fertilization with controlled-release nitrogen, in conjunction with machine transplanting and subsequent topdressing, was applied to Indica–japonica hybrid rice ‘Yongyou1540’ (YY1540) and indica hybrid rice ‘Tianyouhuazhan’ (TYHZ). Four nitrogen treatments were applied in 2018 and 2019: traditional nitrogen application with quick-release nitrogen (T1), single-dose deep fertilization at transplanting with 100% controlled-release nitrogen (T2), and deep fertilization of 70% controlled-release nitrogen and topdressing of 30% quick nitrogen at tillering (T3), or at panicle initiation (T4). Side-deep fertilization reduced the fertilizer application frequency without causing yield loss, T4 enhanced the yield of YY1540 by increasing the number of productive tillers and number of spikelets per panicle compared with T1, T2 and T3. The yield of TYHZ showed no significant difference among treatments. The T4 treatment decreased the number of tillers at the tilling peak stage and increased the percentage productive tillers and number of differentiated spikelets. Compared with the other treatments, T4 increased dry matter accumulation and leaf area index during panicle initiation and grain ripening, and contributed to enhanced nitrogen uptake and nitrogen utilization in YY1540. On average, nitrogen uptake and utilization in YY1540 were highest in T4, but no significant differences among treatments were observed in TYHZ. Dry matter accumulation and nitrogen uptake from panicle initiation to heading of YY1540 were correlated with number of spikelets per panicle, but no significant correlations were observed for TYHZ. Supplementary topdressing with quick-release nitrogen at the panicle initiation stage was required to increase yield of indica–japonica hybrid rice, whereas single-dose deep fertilization with controlled-release nitrogen is satisfactory for the indica hybrid cultivar.

controlled-release nitrogen application (Table 1). In contrast to the traditional fertilization mode (T 1 ), singledose fertilization (T 2 ) decreased the yield of YY1540 by 9.5% (p > 0.05) in 2018 and slightly increased the yield in 2019, whereas T 2 slightly increased the yield of TYHZ in 2018 and 2019. The T 4 treatment increased YY1540 yield by 10.5%, 22.1%, and 17.2% compared with the T 1 , T 2 , and T 3 treatments in 2018, and by 8.0%, 8.0%, and 18.4% in 2019, respectively, which reflected an increase in spikelet number. The T 3 treatment caused a reduction in YY1540 yield in 2019. Regarding TYHZ, a slight decrease (p > 0.05) in yield was observed under the T 3 treatment compared with T 1 , T 2 and T 4 , whereas no significant difference in yield was observed between T 1 , T 2 , and T 4 .
Among the different treatments, yield changes depended on the number of productive tillers and number of spikelets per panicle., The highest number of tillers at the tillering peak stage was observed in the T 3 treatment for YY1540 and T 2 treatment for TYHZ both in 2018 and 2019 (Fig. 1). However, the highest percentage productive tillers was achieved in the T 4 treatment in YY1540 and TYHZ except for the no-nitrogen application control (T 0 ) in 2019 (Fig. 2). The T 3 treatment resulted in the lowest productive tillers percentage for YY1540 in 2018 and 2019, whereas no significant difference in productive tillers percentage was observed between T 1 , T 2 , and T 3 for TYHZ.
With regard to spikelet formation, the main stem was sampled to determine spikelet differentiation and degeneration. The number of spikelets that survived depended primarily on the number of differentiated spikelets. The T 3 treatment resulted in the minimum number of differentiated spikelets for YY1540 both in 2018 and 2019. The number of differentiated spikelets for YY1540 was highest in the T 4 treatment, which was 8.6% and 9.8% higher than those of T 1 in 2018 and 2019, respectively, 18.6% and 20.0% higher than those of T 2 , and 24.3% and 18.7% higher than those of T 3 in 2018 and 2019, respectively. The T 4 treatment significantly promoted spikelet degeneration in 2018, but the highest number of spikelets was observed in T 4 for YY1540. The highest number Table 1. Effect of nitrogen fertilization mode on grain yield and yield components of indica-japonica hybrid rice ('Yongyou1540' , YY1540) and indica hybrid rice ('Tianyouhuazhan' , TYHZ). Values followed by different lower-case letters within the same column are significantly different among treatments. *p < 0.05; **p < 0.01.

Year
Cultivar Treatment The number of productive tillers (10 5 ha −1 ) The number of spikelet per panicle Filled grain rate (%) Grain weight (mg) Grain yield (t·ha −1 ) www.nature.com/scientificreports/  Bars with different lower-case letters are significantly different at the 0.05 probability level among treatments. T 0 no-nitrogen application control, T 1 traditional nitrogen application, T 2 single-dose controlled-release nitrogen application, T 3 70% controlled-release nitrogen as side-deep fertilization with machine transplanting and 30% as topdressing applied at the tillering stage, T 4 70% controlled-release nitrogen as side-deep fertilization with machine transplanting and 30% as topdressing applied at the panicle initiation stage. Nitrogen accumulation and transport. Nitrogen accumulation showed a significant difference between the T 1 , T 2 , T 3 , and T 4 treatments (Fig. 5). Except in T 0 , the nitrogen uptake at the panicle initiation of T 3 was higher than that of the T 1 , T 2 , and T 4 treatments for YY1540, the maximum nitrogen uptake was observed in T 1 in 2018 and 2019 at the heading stage, and the total nitrogen uptake at maturing stage were observed in T 4 treatment. And in TYHZ, the nitrogen uptake at different developmental stages in 2018 showed no significant differences, whereas significantly lower nitrogen uptake at the maturity stage was observed in T 3 compared with that in the T 1 , T 2 , and T 4 treatments. Except for the T 0 treatment in 2019,, the highest proportion of nitrogen accumulation in panicles at the heading stage for YY1540 was higher in the T 4 treatment in 2018 and 2019 (Fig. 6A,B). An identical trend was observed at the maturity stage. In contrast, the proportion of nitrogen accumulation in panicles for TYHZ not differ significantly among the treatments at the heading stage in 2018 (Fig. 6C), whereas in 2019, the highest proportion was observed in the T 2 treatments at the heading and maturity stages disregarding the T 0 control (Fig. 6D). These results indicated that the T 4 treatment promoted nitrogen accumulation in YY1540 panicles, whereas the T 2 treatment fulfilled the panicle nitrogen need demand of TYHZ. With regard to nitrogen transport TNT, NTE and NCR were significantly higher in TYHZ than in YY1540. The different treatments more strongly influenced TNT, NTE, and NCR in YY1540 than those of TYHZ. The highest TNT was observed in the T 1 treatment in 2018 and 2019 for YY1540 and in comparison, TNT in the T 2 , T 3 and T 4 treatments was 29.3%, 29.7% and 34.7% lower respectively, than that of the T 1 treatment on average in 2018 and 2019. The highest TNT for TYHZ was observed in the T 4 treatment, but there were no significant difference observed in among the T 1 , T 2 , T 3 and T 4 treatments. The NTE was significantly higher in the T 3 treatment than that in T 1 , T 2 , and T 4 in 2018, whereas no significant differences were observed in 2019 for YY1540, and for TYHZ in 2018 and 2019. The T 1 treatment showed the highest NCR for YY1540 in 2018 and 2019, which was higher than that of T 2 , T 3 , and T 4 by 22.3%, 14.9%, by 73.0% in 2018, by 42.5%, 10.1%, and 62.0% in 2019, respectively (Table 4). No significant difference was observed among the T 1 , T 2 , T 3 , and T 4 treatments for TYHZ. Table 2. Effect of nitrogen fertilization mode on spikelet differentiation and degeneration of primary stems of indica-japonica hybrid rice ('Yongyou1540' , YY1540) and indica hybrid rice ('Tianyouhuazhan' , TYHZ). Values followed by different lower-case letters within the same column are significantly different among treatments. *p < 0.05; **p < 0.01.

Year
Cultivar Treatment The number of spikelets www.nature.com/scientificreports/   (Table 5). In YY1540, NMP was highest in T 0 and in comparison, the NMP was decreased by 31.4%, 25.4%, 26.0%, and 24.4.0% in T 1 , T 2 , T 3 , and T 4 respectively. The NU t E of YY1540 was highest in the T 2 treatment, but did not differ significantly from T 0 , and that of the T 1 , T 3 and T 4 treatments decreased by 9.7%, 4.2%, and 9.5% compared with T 0 . The highest NRF of YY1540 was observed in the T 4 treatment, which was 11.6%, 30.3% and 40.2% higher than those of the T 1 , T 2 and T 3 treatments, respectively. The NAE of the T 4 treatment was 12.7%, 4.5%, and 32.2% higher than those of T 1 , T 2 , and T 3 treatments, respectively, for YY1540. In contrast, no significant difference among the T 1 , T 2 , T 3 , and T 4 treatments were observed for NMP, NU t E, NRF, and NAE of TYHZ.

Correlation analysis.
To assess the relationship between dry matter accumulation and nitrogen uptake, correlation analysis of the relevant variables at different developmental stages and number of productive tillers, number of spikelets per panicle, or yield was performed on data from 2018 and 2019 combined ( Table 6). Under the different nitrogen application treatments, dry matter accumulation from panicle initiation to heading was significantly correlated to the number of spikelets per panicle and yield in YY1540, whereas nitrogen uptake from panicle initiation to heading was significantly correlated to number of spikelets per panicle. However, for TYHZ, no significant correlations were observed between dry matter accumulation and number of productive tillers and number of spikelets per panicle, in the period from sowing to panicle initiation and from panicle initiation to heading, however a significant correlation was observed between the nitrogen uptake from panicle initiation to heading and number of productive tillers and yield. These results suggested the dry matter accumulation and nitrogen uptake from panicle initiation to heading were more important for panicle formation to increase yield in YY1540 than in TYHZ, and that the different fertilization treatments had little effect on TYHZ.

Discussion
Effect of nitrogen application treatments on yield formation. Compared with the traditional fertilization treatment (T 0 ), nitrogen topdressing at the panicle initiation stage (T 4 ) resulted in increases in yield and nitrogen utilization efficiency over those attained with controlled-release nitrogen application, The present results verified that application of controlled-release nitrogen can improve the nitrogen utilization efficiency of rice, and that side deep fertilization with machine transplanting significantly improves fertilization efficiency 21,22 . However, the yield under single-dose controlled-release nitrogen application (T 2 ) and quick-release nitrogen top dressing at the tillering stage (T 3 ) showed a tendency to reduce yield in YY1540 compared with the T 1 treatment, whereas no significant difference in yield of TYHZ was observed between the T 1 , T 2 , T 3 , and T 4 treatments, These results suggested that controlled-release nitrogen application did not necessarily promote an increase in yield, and that the advantage of controlled-release nitrogen fertilizer application depended on cultivar characteristics and fertilization methods. The yield of indica-japonica hybrid rice is considerably higher than that of indica hybrid rice ( Table 1). The principal advantage of the former is the production large panicles under high biomass 23 , which was correlated Bars with different lower-case are significantly different at the 0.05 probability level among treatments. T 0 no-nitrogen application control, T 1 traditional nitrogen application, T 2 single-dose controlled-release nitrogen application, T 3 70% controlled-release nitrogen as side-deep fertilization with machine transplanting and 30% as topdressing applied at the tillering stage, T 4 70% controlled-release nitrogen as side-deep fertilization with machine transplanting and 30% as topdressing applied at the panicle initiation stage. www.nature.com/scientificreports/ with the grain yield among the four treatments applied to YY1540 24 . In the present study, the release cycle of the controlled-release fertilizer was about 120 days, Correlation analysis showed that dry matter accumulation and nitrogen uptake from the panicle initiation stage to the heading stage was correlated with spikelet number per panicle in YY1540 (Table 6), which suggested that the slow release of nitrogen by the controlled-release fertilizer did not meet the nitrogen demand of Indica-Japonica hybrid rice The productivity advantage of indica-japonica hybrid rice is dependent on the development of large panicles through enhanced by the spikelet differentiation, which may be associated with the cytokine in synthesis 25,26 . However, in the present study, the nitrogen application rate was 195 kg·ha −1 , which was slightly lower than the traditional nitrogen application rate used for indica-japonica hybrid rice 27 . Whether an increased rate of nitrogen application without topdressing can meet the nitrogen needs for panicle development requires further study, However increased nitrogen application at the tillering stage, as applied in the T 3 treatment, resulted in a significant reduction in productive tiller percentage and reduced the number of spikelets per panicle, whereas the productive tiller percentage in the T 4 treatment was significantly increased. Therefore, substantially increased nitrogen application at the tillering stage was not suitable to enhance yield of YY1540. For TYHZ, a slight reduction in yield was also observed in the T 3 treatment, which was consistent with previous results that postponement of nitrogen topdressing may increase rice yield 28 .
In addition, the release of controlled-release nitrogen at two stages is required to meet the nitrogen requirements at different growth stages of indica-japonica hybrid rice. For TYHZ, the T 4 treatment slightly increased (P > 0.05) www.nature.com/scientificreports/ the number of spikelets per panicle, but yield showed no significant difference between the T 1 and T 4 treatments owing to the higher tiller number (P > 0.05). Thus, a single-dose controlled-release nitrogen application could be adopted for indica hybrid rice cultivars such as TYHZ, which are more dependent on the number of productive tillers to attain a high yield than the number of spikelets per panicle 26 .

Effect of nitrogen application treatments on dry matter accumulation. Dry matter accumulation
is significantly influenced by nitrogen application, which plays an important role in maintaining photosynthesis 29 . In the T 3 treatment, a higher number of tillers were observed at the tillering peak stage compared with that observed in the T 1 , T 2 , and T 4 treatments in YY1540, which was induced by the excessive nitrogen. Dry matter accumulation from sowing to panicle initiation was not significantly influenced by the nitrogen treatment, but the dry matter accumulation of a single tiller was decreased, which lead to the low percentage productive tillers (Fig. 2), lower leaf area index (Fig. 3), and lower number of spikelets per panicle (Table1) in YY1540 These responses may have been induced by carbohydrate competition between tillers 26,30 . Dry matter accumulation from sowing to panicle initiation was higher in TYHZ than that in YY1540 (Table3), which may explain why the tiller number at the tillering peak stage was higher in TYHZ, This finding may also account for the stronger tillering ability of TYHZ, whereas indica-japonica hybrid rice exhibits intersubspecific heterosis for panicle initiation under similar dry matter accumulation from panicle initiation to heading 31,32 . Bars with different lower-case letters are significantly different at the 0.05 probability level among treatments. T 0 no-nitrogen application control, T 1 traditional nitrogen application, T 2 single-dose controlled-release nitrogen application, T 3 70% controlled-release nitrogen as side-deep fertilization with machine transplanting and 30% as topdressing applied at the tillering stage, T 4 70% controlled-release nitrogen as side-deep fertilization with machine transplanting and 30% as topdressing applied at the panicle initiation stage. www.nature.com/scientificreports/  www.nature.com/scientificreports/ Nitrogen supplementation is important to maintain carbohydrate supply for grain ripening of indica-japonica hybrid rice 19 . In the present study, a larger reduction in grain weight was observed in YY1540 compared with that of TYHZ under the T 3 treatment. The indica-japonica hybrid rice YY1540 has a longer grain-filling stage than that of TYHZ. The significant decrease in dry matter accumulation from heading to maturity may be caused by leaf senescence 4 . The lower SPAD value is consistent with these results (Fig. 4). The SPAD value and dry matter accumulation in the heading-maturity stage were higher in YY1540 than in TYHZ, which indicated that the leaf photosynthetic capacity during the grain-filling stage was higher in YY1540 than in TYHZ. Thus, a large amount of supplementary nitrogen is needed to maintain the chlorophyll content and photosynthesis capacity of YY1540 33 , which was the reason that nitrogen topdressing at panicle initiation stage was needed in YY1540.

Effect of nitrogen application treatments on nitrogen utilization. Nitrogen accumulation in
YY1540 was higher than that in TYHZ (Fig. 5), which is consistent with the stronger nitrogen absorption capacity in indica-japonica hybrid rice than that of indica hybrid rice 34 . Although the total nitrogen uptake at the heading stage was slightly lower in YY1540 than that in TYHZ, the nitrogen accumulation in panicles was higher in YY1540 than in panicles of TYHZ (Fig. 4), which showed that the contribution of nitrogen to panicle development in YY1540 is greater than that in TYHZ. In TYHZ nitrogen was used mainly for leaf growth, which is consistent with the higher NMP of YY1540 compared with that of TYHZ ( Table 5).
The NCR was lower in YY1540 compared with that of TYHZ, and the TNT of YY1540 was highest in the T 1 treatment (Table 4). These findings are is consistent with a previous study that showed nitrogen accumulation in indica-japonica hybrid rice primarily depends on absorption rather than transport 35 , In addition, an adequate late growth nitrogen supply is extremely important for indica-japonica hybrid rice 27 . The lower NRF on average of YY1540 compared with that of TYHZ also supported this interpretation. However, NRF a lower in the T 2 and T 3 treatments than that in the T 1 and T 4 treatments (Table 5). Furthermore the NRF of YY1540 in the T 2 and T 3 treatments was lower than that in TYHZ, which may be associated with the ability of indica-japonica hybrid rice to contribute to nitrogen priming in soil with high microbial activity. The T 2 treatment resulted in the highest NU t E in YY1540 and TYHZ (Table 5). In addition, a reduction in NU t E was observed at the quick-release nitrogen application treatments, which showed that controlled-release nitrogen is beneficial for improvement of nitrogen utilization efficiency and reduction of nitrogen losses 36,37 .

Conclusions
Controlled-release nitrogen side-deep application with machine transplanting can reduce the required fertilization frequency. Nitrogen topdressing at the panicle initiation stage is necessary for high yield formation in indica-japonica hybrid rice with large panicles. This treatment leads to increase in the number of spikelets per panicle by ensuring sufficient nitrogen fertilizer supply and increase in dry matter accumulation, and results in maximum nitrogen use efficiency in indica/japonica hybrid rice. Single-dose controlled-release side-deep fertilization can satisfy the nitrogen absorption demand for high-yield formation in indica hybrid rice, which produce medium-type panicles and show enhanced tillering capacity.

Materials and methods
Experimental site and meteorological conditions. The field experiments were conducted from May to October in 2018 and 2019 at the China National Rice Research Institute, Hangzhou, Zhejiang Province (119° 55ʹ 48ʺ E, 30° 2ʹ 24ʺ N), China. The experimental field contained common paddy soil with pH 5.45, organic matter 32.13 g·kg −1 , total nitrogen 1.69 g·kg −1 , available phosphorus 85.8 mg·kg −1 and available potassium 95 mg·kg −1 . Soil analyses were performed on samples collected from the uppermost 20 cm following the methodology of Ke et al. 38 . Table 6. Correlation analysis between dry matter accumulation and nitrogen uptake in indica-japonica hybrid rice ('Yongyou1540' , YY1540) and indica hybrid rice ('Tianyouhuazhan' , TYHZ). *p < 0.05; **p < 0.01.

YY1540 TYHZ
The number of productive tillers www.nature.com/scientificreports/ During the 2-year experiment, rainfall, relative humidity, temperature, and solar radiation were measured in the field using a HOBO weather station (MAH-H21, Onset Computer Corporation, Bourne, MA, USA). The average rainfall of 2018 was higher than that of 2019 and the humidity showed a similar trend ( Table 7). The monthly average temperature in 2018 ranged from 17.1 °C to 29.4 °C, which was slightly lower than that in 2019 (19.8-32.2 °C). The cumulative solar radiation was 391.7 MJ·m −2 in 2019, which was higher than that in 2018 (272.8 MJ·m −2 ). Plant material. We selected two super-rice cultivars grown commercially in China. The indica-japonica hybrid rice 'Yongyou1540' (YY1540) and indica hybrid rice 'Tianyouhuazhan' (TYHZ), have been widely cultivated in the middle and lower regions of the Yangtze River for single-season rice production for more than 10 years, and are representative of the indica-japonica hybrid rice and indica rice cultivars in cultivation 39,40 .
The growth period of YY1540 from sowing to maturity in 2018 and 2019 was 161 days and 157 days, respectively. The average spikelet number of YY1540 was 338 per panicle. The growth period of TYHZ from sowing to maturity in 2018 and 2019 was 129 days and 131 days, respectively. The average spikelet number of TYHZ was 217 per panicle. Details of the rice growth period in both years of the study are shown in Table 8.
Experimental design. The experiment was conducted using a split-plot design with cultivars as primary plots and nitrogen treatments as subplots. A total of 195 kg N ha −1 was applied during the rice growth period. The treatments are summarized in Table 9.   www.nature.com/scientificreports/ For all treatments, quick-release nitrogen was applied as common urea (nitrogen content: 46%) and controlled-release nitrogen was applied as a slow release fertilizer (N content: 41.6%, Kingenta International Co., Ltd., Shandong, China). Phosphorus fertilizer was applied as calcium superphosphate at the rate of 510 kg·ha −1 as a basal dressing. Potassium fertilizer was split-applied with 50% as a basal level and 50% at the panicle initiation stage at a rate of 280 kg·ha −1 potassium. The deep fertilization machine used was developed by the Jinhe Agricultural Science and Technology Co, Ltd. (Zhejiang, China) following the methodology of Zhu 20 . The depth of fertilization was 5 cm. The irrigation method followed local practices established to achieve high-yield crops. The treatments were repeated three times. The subplot size was 216 m 2 .
Determination methods. Yield and yield components. Yield and yield components were determined as described by Yoshida 41 . The grain yield was determined from a harvest area of 6 m 2 in each subplot at the rice brown stage and adjusted to 13.5% grain moisture. The yield components (panicle number, spikelets number per panicle, percentage of filled grains and grain weight) were determined from the plants within a 1 m 2 area randomly chosen in each subplot (excluding border plants) in accordance with the method of Kamiji 42 . The filled-grain percentage was calculated as follows: number of filled grains per panicle/(number of filled grains per panicle + number of unfilled grains per panicle) × 100.
Tiller number dynamics. After transplantation, the number of tillers was determined at 7-days intervals.
Thirty plants per plot were selected for investigation. The criterion for tiller recognition was presence of three leaves on each tiller. The proportion of effective panicles was calculated as the number of effective panicles/tiller number at the peak stage.

Spikelet differentiation and degeneration.
To maintain a consistent panicle development period for each treatment, ten primary tillers (tagged during rice growth) from each subplot were collected when the panicle was 50% elongated and used to quantify spikelet differentiation and degeneration. The number of degenerated spikelets was calculated by counting the vestiges present on the panicles in accordance with the method of Yao 43 . The number of differentiated spikelets per panicle is the sum of surviving and degenerated spikelets per panicle. The proportion of degenerated spikelets was calculated as follows: number of degenerated spikelets per panicle/number of total differentiated spikelets per panicle × 100.
Dry matter accumulation, leaf area index and SPAD value. Ten plant hills were sampled from each plot based on the average tiller number at the panicle initiation, heading and maturity stages. To determine level of dry matter accumulation, the sampled plants were dried at 105 °C for 30 min and then dried to a constant weight at 80 °C following the method of Zhu 20 . Panicles, leaves, and stems with leaf sheaths were separated from the plants after the heading stage. An LI-3100C Area Meter (LI-COR, Inc., Lincoln, NE, USA) was used to measure the leaf area of each green leaf. The leaf area per square meter was calculated as the leaf area index following the method of Wu 19 . The SPAD value was determined at the maturity stage using a SPAD chlorophyII meter (SPAD-502PLUS, Spectrum Technologies, Aurora, IL, USA).
Plant nitrogen content, nitrogen uptake and nitrogen use efficiency. After measurement of the sample dry matter at the panicle initiation, heading and maturity stages, the total nitrogen content of the samples was determined using the micro-Kjeldahl digestion method in accordance to the procedure described by Bremner 44 . To determine the nitrogen content, ground samples (0.20 g) of the panicle, leaf, stem, and sheath were digested in H 2 SO 4 -H 2 O 2 solution at 420 °C for 2 h and distilled with 10 mol·L −1 sodium hydroxide solution. The evaporated NH 3 was absorbed with Na 2 B 4 O 7 and titrated with 0.01 mol·L −1 sulfuric acid. According to the consumption of the titrant, the percentage of nitrogen was calculated, which was analyzed by the micro-Kjeldahl method (Kjeltec TM 8400, FOSS, Helleröd, Denmark). Nitrogen uptake was calculated using the following formula: dry matter accumulation × nitrogen concentration.
Measures of nitrogen transport and nitrogen use efficiency, namely the amount of nitrogen amount transport from stem, sheaths and leaves to panicles (TNT), apparent nitrogen translocation efficiency of stem, sheaths, and leaves (NTE), rate of contribution of transferred nitrogen into grains (NCR), nitrogen dry matter production efficiency (NMP), nitrogen utilization efficiency (NU t E), apparent nitrogen recovery fraction (NRF), and nitrogen agronomic efficiency (NAE) were calculated using the formulas described by López-Bellido et al.
TNT = Nitrogen amount of leaf and stem sheath at the heading stage − Nitrogen amount of leaf and stem sheath at the maturity stage; NTE = TNT/Nitrogen amount of leaf and stem sheath at the heading stage × 100; NCR = TNT/Nitrogen amount of grains at the maturity stage × 100; NMP = Total dry matter accumulation uptake at the maturity stage/Total nitrogen accumulation uptake at the maturity stage; NU t E = Grain yield/Total nitrogen accumulation uptake at the maturity stage; NRF (%) = Nitrogen accumulation above ground with nitrogen treatment − Nitrogen accumulation above ground with no-nitrogen/Total nitrogen application rate × 100; NAE = Grain yield with nitrogen treatment − Grain yield with no nitrogen treatment/Total nitrogen application rate.
Statistical analyses. The experimental data for YY1540 and TYHZ were analyzed statistically using oneway analysis of variance as implemented in SAS  www.nature.com/scientificreports/ of the different treatments (three replicates) were analyzed with Duncan's multiple range test (p < 0.05) using SAS 9.1 (SAS Corp.). Graphs were generated using Origin 9.1 (Origin Lab, Northampton, MA, USA).

Data availability
The data used or analyzed during the current study are available from the corresponding author on reasonable request.