Overexpression of rice gene OsATG8b confers tolerance to nitrogen starvation and increases yield and nitrogen use efficiency (NUE) in Arabidopsis

Nitrogen (N) is an important element required for plant growth and development, which also affects plant yield and quality. Autophagy, a conserved pathway in eukaryotes, degrades and recycles cellular components, thus playing an important role in N remobilization. However, only a few autophagy genes related to N remobilization in rice (Oryza sativa) have been reported. Here, we identified a core autophagy gene in rice, OsATG8b, with increased expression levels under N starvation conditions. It was investigated the function of OsATG8b by generating three independent homozygous 35S-OsATG8b transgenic Arabidopsis thaliana lines. The overexpression of OsATG8b significantly enhanced autophagic flux in the transgenic Arabidopsis plants. It was also showed that over-expressing OsATG8b promoted growth and development of Arabidopsis, in which the rosette leaves were larger than those of the wild type (WT), and the yield increased significantly by 25.25%. In addition, the transgenic lines accumulated more N in seeds than in the rosette leaves. Further examination revealed that overexpression of OsATG8b could effectively alleviate the growth inhibition of transgenic Arabidopsis under nitrogen (N) stress. N partitioning studies revealed that nitrogen-harvest index (NHI) and nitrogen use efficiency (NUE) were significantly increased in the transgenic Arabidopsis, as well as the 15N-tracer experiments revealing that the remobilization of N to seeds in the OsATG8b-overexpressing transgenic Arabidopsis was high and more than WT. Based on our findings, we consider OsATG8b to be a great candidate gene to increase NUE and yield, especially under suboptimal field conditions.


Introduction
Nitrogen (N) is one of the most important nutrient elements for plants and is required for growth and development. N is also a vital component of amino acids, protein, nucleic acid, chlorophyll, and plant hormones [1]. In plants, especially cereal crops, the yield and quality of the grain depends not only on the absorbed N prior to flowering, but also on the mobilization of reused N from the mature leaves during senescence [2,3]. Plants are sedentary and cannot move to acquire nutrients and minerals as needed. Survival depends on their ability to consume the mineral nutrients available in the rhizosphere, metabolize, recycle, and conserve them efficiently during their lifespan [4,5]. The chloroplasts are the main source of C and N recycling in plants, which contribute to the 75-80% nitrogen of the leaf total protein in C3 plants [6]. Rubisco, a photosynthetic CO 2 fixing enzyme, accounts for more than 50% of the soluble protein content in the leaves, and is the main source of nitrogen mobilization and reuse [7]. In the later stages of development, leaf protein (especially the chloroplast protein) is rapidly degraded as the output of nitrogen, as the released free amino acids and other compounds are transferred to the developing reproductive and storage organs, such as new leaves or seeds [2,8]. This mobilized nitrogen is an important source of nitrogen for seeds [9]. The circulation process for organic nitrogen transfer from aging tissues to seeds in plants is an important determinant of productivity and yield, especially under nitrogen deficiency stress [7]. Although the pathways and steps in protein turnover in aging leaves are not clear, plastid resident proteases, senescence-associated vacuoles (SAVs), and macroautophagy (hereafter referred to as autophagy) are considered to be three important routes [4,10]. Autophagy, which is a conserved vacuolar degradation pathway by which cells recycle components, including unwanted macromolecular substances or damaged organelles, plays a vital role in nutrient remobilization. Autophagy is continuously maintained at a basal level to maintain cell homeostasis under normal conditions. In response to stress or nutrient starvation, autophagy is enhanced to facilitate the degradation of increasing levels of toxic and damaged components, and nutrients mobilized from this recycled cell material is then used for the maintenance of cellular processes and adaptation to stress [11,12]. A pioneering study by Professor Yoshinori Ohsumi led to the discovery of the AuTophaGy-related genes (ATG genes) in yeast [13]. A total of 16 ATG genes were identified in the first yeast screening. Subsequently, many ATG genes have been characterized in different species, including mammals and plants [14,15]. In Arabidopsis (Arabidopsis thaliana), most ATG genes are transcriptionally up-regulated during leaf senescence and nutrient starvation [16,17]. Early senescence symptoms and hypersensitivity to C and N starvation were displayed in autophagy mutants [18], and a previous study revealed that the overaccumulation of salicylic acid (SA) is involved in the early senescence in atg mutants [19]. A study of Arabidopsis atg mutants provide the first evidence that N remobilization from the leaves to the seeds is controlled by autophagy [20]. In Arabidopsis, the wild type (WT) and atg mutants (atg18a RNAi, atg5 and atg9) were fed with 15 NO 3 during the vegetative growth stage, 15 N remobilization sharply decreased in all the atg mutants compared to the WT under N-limited conditions, and the NUE decreased significantly to about 50% in the seeds. In the rice Osatg7-1 mutant, biomass and NUE during the vegetative growth stage were significantly reduced, and the mutant was unable to mobilize N in aging leaves [21]. In maize (Zea mays), 15 N pulse-chase analysis revealed that the amount of recycled N was reduced by two-fold in the atg12 mutant compared with that of the WT [22]. Among the many plant ATG proteins, the abundance of ATG8 proteins at the cellular level can regulate the size of autophagosome. ATG8 is a ubiquitin-like protein conjugated to phosphatidylethanolamine (PE) on the autophagic membrane. ATG8 plays a central role in autophagy, which is involved in autophagic vesicle expansion [23,24] and is located on the autophagosome membrane [25]. The ATG8 is often used as a reliable marker of autophagic activity in plants and animals [12,26]. While ATG8 protein is unique in yeast, many isoforms exist in plants and are encoded by different genes. There are 7 identified ATG8 isoforms (OsATGa to OsATGg) in rice genome. OsATG8a was the first cloned rice ATG8 gene, interacted with OsATG4 [27]. OsATG8a, OsATG8b and OsATG8c share high levels of amino acid sequence similarity, while OsATG8d has high similarity to AtATG8i [28,29]. In addition, OsATG8a and OsATG8c had more closely phylogenetic relationships with ZmATG8a and ZmATG8c, respectively [30]. However, OsATG8e lacks supporting expressed sequence tag (EST) data, OsATG8f has no corresponding full-length cDNA, and OsATG8g has not been mapped to the rice genome [29]. Several studies indicate that overexpression of ATG8 promotes plant growth and increases tolerance to N starvation in plants. For example, transgenic Arabidopsis plants overexpressing AtAtg8f were slightly larger in size than the control plants, and flowered earlier than the control plants in a N-deficient medium [31]. Constitutive expression of apple MdATG8i and soybean GmATG8c in apple and soybean calli, respectively, not only led to better tolerance to N starvation, but also accelerated the growth of calli under optimal growth conditions. Overexpression of soybean and apple ATG8 genes, GmATG8c and MdATG8i, respectively, promoted growth and improved the yield in the transgenic Arabidopsis [30,32]. Similarly, Li et al observed that heterologous expression of foxtail millet SiATG8a in Arabidopsis and rice improved their performance under normal growth conditions and conferred tolerance to N starvation [33,34]. Recently, research of Masclaux-Daubresse's group showed that overexpression of AtATG8s (AtATG8a, AtATG8e, AtATG8f, AtATG8g, respectively) in Arabidopsis increased autophagosome number and stimulated autophagic activity, increased N remobilization efficiency (NRE) only under full N conditions but did not affect yield and biomass [35].
In order to screen for potential yield and N remobilization genes in rice, the rice ATG8 gene OsATG8b (LOC_Os04g53240) was identified. OsATG8b has a 360-bp coding sequence (CDS) and encodes 120 amino acids. We then confirmed its function using 35S-OsATG8b in transgenic Arabidopsis plants. Compared to the WT, three independent transgenic lines were found to have improved vegetative growth, developed to the reproductive stage slightly earlier, were taller, and produced more siliques and seeds. Overexpression of OsATG8b therefore, not only led to better performance under normal growth conditions but also maintained growth under N starvation. The autophagic flux was increased in the transgenic lines. At the same time, the number of autophagosomes were significantly enhanced under N deficiency (N free with 0 mM nitrogen) with or without the addition of the autophagy inhibitor. Growth inhibition caused by nitrogen stress was effectively alleviated. N partitioning studies have shown that overexpression of OsATG8b increased N allocation to seeds and have enhanced NHI and NUE indicated by NHI/ HI ratio [20], while the 15 N allocation to seeds was significantly higher than that in the WT using 15 N isotope tracer experiment. Therefore, our results indicate that OsATG8b may be an important candidate gene for rice with synergistically-enhanced NUE and better yield potential.

OsATG8b displays N stress inducible expression in rice and the expression patterns of OsATG8b in transgenic Arabidopsis
In order to analyze whether the expression of OsATG8b in rice is responsive to N limitation, we first examined the transcript levels of OsATG8b by real-time RT-PCR. When rice seedlings were placed under N stress for 1 day, the expression level of OsATG8b in leaves was not different in low N (NL) conditions when compared with that of the control, but was highly induced under N deficient (ND) conditions ( Fig 1A). In addition, the levels of OsATG8b in the root increased with increasing degree of N stress (Fig 1B). After 3 days of N stress, the expression of OsATG8b in leaves increased with increasing degree of N stress but was highly induced only under ND conditions in the root (Fig 1). These results suggest that expression of OsATG8b in rice is induced by N stress.
To examine the temporal and spatial expression patterns of OsATG8b, we performed histochemical analysis of the promoter-GUS fusion report systems. We transformed Arabidopsis with the Pro OsATG8b -GUS construct, and histochemical staining indicated that GUS activity could be detected in the germinating seed, cotyledon, hypocotyl, the rosette leaf, root, as well as the mature inflorescence and silique in the transgenic seedlings (S1 Fig).

Overexpression of OsATG8b in transgenic Arabidopsis promotes growth and increases yield
To fully understand the functional roles of OsATG8b, we transformed Arabidopsis plants with the 35S-OsATG8b binary expression vector. We obtained over 10 independent transgenic lines. In order to eliminate the influence of AtATGs, real-time RT-PCR was used to monitor the transcript levels of AtATGs. We found that the transcript levels of the endogenous AtATGs were unchanged and not affected in the transgenic lines (S2 Fig). The transcript and protein levels of OsATG8b were confirmed in three homozygous transgenic lines using PCR, real-time RT-PCR and western blotting (Figs 2A, 2B and 3). Lines 13, 14 and 21 (L-13, L-14, L-21) refer to the over-expressed lines. Phenotypic effects of OsATG8b were observed throughout the plant developmental stages. Under normal growth conditions, the rosette of the transgenic lines was significantly larger and grew rapidly than the WT (Fig 2C and 2E). Overexpression of OsATG8b accelerated plant development, bolting as flowering occurred 6 days earlier and also produced more total leaves in the transgenic lines when compared to that in the WT (Table 1), while the leaf number at bolting was not significant different with WT. On an average, plant height in the transgenic lines was significantly increased by up to 10.42% when compared to the WT plant height (Fig 2D and 2F). The chlorophyll content of the rosette leaves in the transgenic lines was consistently higher than those seen in the WT as the seedlings aged ( Fig 2G). There was minimal or almost no difference in the soluble protein content in the rosettes of 8-day-old transgenic seedlings and the WT seedlings; however, from 8 days until plant maturity, the soluble protein content in the transgenic lines was consistently higher than those seen in the WT plants ( Fig 2H). We further investigated the effect of OsATG8b on yield and found that the transgenic lines produced significantly higher number of siliques. As a result, the transgenic lines had a significantly higher thousand grain weight and total seed weight per plant than those in the WT (Table 2). These results show that constitutive expression of OsATG8b both promotes growth and increases yield in transgenic Arabidopsis.

Overexpression of OsATG8b in transgenic Arabidopsis enhances tolerance to N deficiency
To investigate the effect of OsATG8b overexpression on the plant responses to N stress, sevenday old transgenic seedlings and WT were transferred to ½ MS medium with sufficient or free N and incubated for 9 days. Changes to leaves and root were closely monitored. The transgenic lines overexpressing OsATG8b performed better than the WT, both under N-sufficient and Nfree conditions. The transgenic lines had significantly larger rosettes (Fig 3A), as also the fresh weight of leaves in the transgenic lines was significantly higher both under N-sufficient and Ndeficient conditions when compared to the WT ( Fig 3B). The chlorophyll and soluble protein contents were also significantly higher than that of the WT (Fig 3C and 3D), and the enzymatic activity of nitrate reductase (NR) in transgenic Arabidopsis was obviously increased (Fig 3E), whether the supply of N was sufficient or deficient. These results indicate that the transgenic lines were able to tolerate N deficiency better and improve the N assimilation efficiency.

Overexpression of OsATG8b in transgenic Arabidopsis enhances the autophagic activity and flux
To investigate whether the autophagy activity is enhanced in the transgenic Arabidopsis overexpressing OsATG8b, seven-day-old seedlings of transgenic lines and WT were treated with  Values are means ± SD of three biological replicates (n = 24) �� P < 0.01 (t-test), significant difference from the WT, the time to flowering and bolting were scored as days after sowing (DAS). The total leaf number was counted after the rosettes growth completed.
ND for 12 h, and MDC staining was used to determine autophagic activity in the leaves. Concanamycin A (ConA) enables distinct observation of the autophagosome by increasing the internal pH in the vacuolar and resident hydrolase activity in an optimal environment [36,37]. The autophagy receptor NBR1 (Neighbor of BRCA1) protein of Arabidopsis is an autophagy substrate that degrades in the vacuole, and the degradation of NBR1 can be used as a measure of selective autophagic flux in plants [38,39]. As shown in Fig 4, the consumption of NBR1 protein was higher in transgenic lines, under NS condition, the expression level of NBR1 protein in transgenic lines were 15%-28% less than that of WT. While the degradation of NBR1 protein was 34% in WT under ND compared to NS condition, but it was as much as 41%-64% in transgenic lines (Fig 4, DMSO-ND). However, when applied with ConA to block the degradation of autophagic bodies and NBR1 protein, the accumulation of NBR1 protein was greatly increased (Fig 4, +ConA), especially under the ND treatment. The previous reports showed that the lipidation of ATG8 can be used to indicate the autophagic activity [40,41]. As the autophagy core protein ATG8a usually acts as the autophagic flux marker protein, and previous studies have used Arabidopsis anti-ATG8a antibody to exam the autophagic flux in various species (such as Arabidopsis, maize and tomato) [22,[40][41][42]. The autophagy flux in our transgenic plants was monitored with both anti-AtATG8a and anti-OsATG8b antibodies. These results showed that overexpression of OsATG8b significantly increased the accumulation of both AtATG8a-PE and OsATG8b-PE in transgenic lines, thus the lipidation of OsATG8b was enhanced, especially under N deficiency condition (Fig 4, DMSO), and increased to a greater extent after ConA treatment (Fig 4, +ConA). The detection signal of lipidation used by anti-OsATG8b antibody was better than that of anti-AtATG8a antibody. Monodansylcadaverine (MDC), an acidophilic dye widely used in mammals and plants can be used as a probe to detect autophagosome [42]. To facilitate the observation of autophagic bodies, the H + -ATPase inhibitor ConA was added. The leaves of transgenic Arabidopsis and WT were excised and immediately observed by confocal microscopy. We observed significantly more MDC-positive autophagic structures in the mesophyll cells of the transgenic lines, and the fluorescence signal intensity was higher than that seen in the WT (Fig 5A). The autophagy inhibitor 3-methyladenine (3-MA), specifically inhibits PI3K activity [43,44]. The addition of 5 mM of 3-MA to the N-deficient treatment clearly suppressed the fluorescence signal of MDC-stained autophagosomes, but some fluorescent signal could still be observed in the transgenic lines (Fig 5B). The fluorescence ratio (positive signal/aera) in the WT was 30.87 under ND condition, and decreased to 13.07 after 3-MA treated, while the ratios in transgenic lines were 121.85 and 62.53. It was showed that the fluorescence ratios in the overexpressing lines were up to 3 or 4 times more than that in WT with or without 3-MA. All these results reveal that overexpression of OsATG8b could significantly increase the autophagosome number and autophagic flux in transgenic Arabidopsis. This can effectively improve the tolerance of transgenic Arabidopsis to N deficiency stress.

Overexpression of OsATG8b in transgenic Arabidopsis modifies N use efficiency and N remobilization efficiency
To determine whether the overexpression of OsATG8b modifies the N status at the whole plant level, the N concentration in rosettes, stem, and seeds was measured after harvesting. The N concentration (N%) in rosettes of the transgenic lines was much lower than that of the WT, but there was no difference in the stem samples. In contrast, the N% in seeds was significantly higher in the transgenic lines ( Fig 6A). However, there was no significant difference in the C concentration (C%) between the transgenic lines and the WT (Fig 6B). Given the increased yield in transgenic lines, we further analyzed the biomass of plants. Relative to WT, the biomass of the transgenic lines increased by 25.95% to 26.89% (Fig 6C).
We studied the effect of OsATG8b on plant yield, seed N filling, harvest index (HI) and N harvest index (NHI). While NHI is an important indicator of production, it also highlights the N allocation efficiency to seeds. As shown in Fig 6D and 6E, the HI and NHI were significantly higher in the transgenic lines compared to that in the WT plants. In order to determine the NUE, we used the NHI/HI ratio as an indicator of NUE variation [20]. The NHI/HI ratio in the transgenic lines was significantly higher than that of the WT (Fig 6F), suggesting that OsATG8b might be essential for both seed yield and NUE. These results clarify that overexpression of OsATG8b effectively redeploys N allocation from the leaves to seeds and increases the NUE of Arabidopsis plants. Enhanced autophagy with increased yield and nitrogen use efficiency To monitor N movements accurately, 15 N-pulse chase experiments were performed at the bolting stage similarly as described in previous studies [20,22]. Samples of OsATG8b-overexpressing lines and WT were dissected into seed and remains (rosettes and stems), then traits for N utilization and remobilization were calculated. To determine the N remobilization efficiency (NRE), the partitioning of 15 N in WT and transgenic lines were tracked. It was showed that overexpression of OsATG8b obviously reduced the partitioning of 15 N into remains ( Fig  6G). In contrast, the OsATG8b-overexpressing lines accumulated significantly more 15 N in seeds relative to WT, as indicated by the higher 15 NHI in transgenic lines (Fig 6H). The ratio of 15 NHI to HI was next used to compare the NRE variation as described previously [20,22], and the 15 NHI/HI were substantially increased in OsATG8b-overexpressing lines than in WT (Fig 6I), suggesting that through overexpression of OsATG8b could significantly improve the N remobilization into seeds, which resulted in favorably increased yield and NUE.

Discussion
Autophagy is involved in removal of damaged proteins and organelles, and response to biological and abiotic stresses [11]. Previous studies have shown that nutrition-limitation induces autophagy in plants [45,46]. Under C and N stress conditions, transcription of many plant ATG genes was significantly up-regulated similar to the high autophagy activity [28,29,33,47,48]. The transcriptional level of OsATG8b increased with N deficiency in rice seedlings (Fig 1), which is evidence that OsATG8b response to N stress. On the other hand, it would provide evidence that overexpression of OsATG8b could effectively increase tolerance to N stress (Fig 3A). The GUS activity of OsATG8b promoter was detected in the inflorescence, seedpod, and in particular was higher in the aging leaves, which indicated that OsATG8b gene may participate in   15 NHI was measured as the partitioning total 15 N in seeds. (I) 15 NHI/HI ration was calculated as a percentage of WT and used to estimate NRE between the transgenic lines and WT. Values are means ± SD of three biological replicates (n = 12), � P < 0.05, �� P < 0.01 (t-test). https://doi.org/10.1371/journal.pone.0223011.g006 Enhanced autophagy with increased yield and nitrogen use efficiency seed formation in plants and also be involved in autophagic degradation during leaf senescence (S1A and S1E Fig).
With defective autophagy, the redundant N in the cells cannot be effectively recovered or utilized, leading to nutrient accumulation, wastage, cell death, and lower amino acid content, the atg deletion mutants showed accelerated senescence, decreased survival rate and other unfavorable phenotypes under nutrient stress conditions [11,[20][21][22]49,50]. Consistent with this view, the transgenic Arabidopsis plants containing 35S-OsATG8b not only performed stronger and better than the WT under normal growth conditions, but also effectively overcome growth inhibition under N deficiency conditions (Fig 3A). In addition, we also found that the transgenic lines had increased chlorophyll content and soluble protein content in the leaves (Fig 2), so did the NR activity (Fig 3D). These results indicate that overexpression of OsATG8b leads to increased accumulation of available N in the 'source' of transgenic Arabidopsis. The overexpression of OsATG8b accelerated the switch from vegetative to reproductive growth in the transgenic Arabidopsis (Table 1), which was probably because of a healthy vegetative phase to have an extended reproductive phase resulting in increased number of siliques, and consequently, higher thousand grain weight and better yield ( Table 2).
In the OsATG8b-overexpressing lines, the enhanced degradation of NBR1 and lipidation of ATG8, as well as the ConA-mediated increased accumulation of NBR1 and AtATG8a-PE/ OsATG8b-PE, all these results revealing that the autophagic flux and activity significantly increased under ND condition (Fig 4-ND). Additionally, the MDC staining also supported this conclusion (Fig 5). Even under N sufficient condition, the nutrients still could be mobilized and reused from leaves to reproductive organs (seeds) to the maximum extent, as a result of enhanced autophagy activity (Fig 4-NS). The growth, development, and yield of transgenic Arabidopsis could be enhanced.
Autophagy is also as an important role in the remobilization and transfer of N to the grain [11,24]. Both NHI and seeds N concentration (N% Seeds ) are major indicators of NUE and seed nutritional quality [7]. Overexpression of OsATG8b in transgenic Arabidopsis resulted in higher NHI, HI, NHI/HI, N% Seeds , and at the same time reduced N waste in their dry remains (Fig 6), indicating that OsATG8b plays a vital role in NUE. This could explain OsATG8b increases both grain production and the allocation of biomass to grain. On the other hand, 15 NHI/HI ratio is used to measure the variations as Nitrogen Remobilization Efficiency (NRE). A significantly higher 15 NHI/HI ratio observed in the transgenic lines compared to the WT (Fig 6I) implied that OsATG8b could also increase the allocation of N to seeds. The increased seed yield was at least partially caused by effective remobilization of assimilated N from the vegetative tissue to the developing seeds. Corresponding to the increased autophagy activity described above, since autophagy likely contributes to the nitrogen pool for seed production, we speculated that the transgenic Arabidopsis could increase N transported and accumulated during the vegetative growth stage, which in-turn could provide abundant nutrients for reproductive growth and higher yield even under N full condition.
A recent paper of Masclaux-Daubresse's group has just reported that overexpression of four AtATG8 genes (AtATG8a, AtATG8e, AtATG8f, AtATG8g), respectively, in Arabidopsis could significantly increase the number of autophagosomes and the transcripts of other AtATG genes [35]. Under full N conditions, although N remobilization efficiency significantly increased relative to the control line by 15 N isotope tracer test, the ATG8 overexpressors did not affect yield and biomass [35]. We two did similar but different studies during the same period. Although these similar conclusions were got in our research, there were also clear differences between us. It seemed that our OsATG8b gene was powerful in the development of plant. The overexpressing OsATG8b transgenic Arabidopsis not only unaffected the expression of other AtATGs (S2 Fig), but also significantly increased seed yield and thousand grains weight under normal growth conditions. Another point worth mentioned was the nitrogen absent (ND) condition of in our experiment was 0 mM N, it was a very very harsh stress condition. Our transgenic plants were still doing well and made a better state of growth, even restored the growth inhibition caused by N stress (Figs 2 and 3). Although rice and Arabidopsis are two type plants, there may be differences in gene function. In addition, the conclusions of we two groups could support or complement each other.
Taken together, our results suggest that the rice autophagy gene OsATG8b that can enhance NUE and improve crop yield, which has a positive role in the maintenance of plant fitness and adaptation to nutrient limitations. These results have also linked autophagy to plant nitrogen distribution/homeostasis, as well as increased yield.

Plant materials and growth conditions
Rice (Japonica cv Shennong9816) seeds were sterilized and germinated at 28˚C for 2 d, then grown hydroponically in a growth chamber (28˚C/25˚C and 10 h light/14 h dark). Arabidopsis seeds (Columbia-0) were surface-sterilized and stratified at 4˚C for 2 days in the dark, then grow on ½ Murashige and Skoog (½ MS) medium (with 0.75% agar and 1% sucrose (w:v), pH = 5.7) in a growth chamber (22 ± 1˚C and 16 h light/8 h dark). 8-day-old Arabidopsis seedlings were transferred to soil with sufficient nutrient and cultivated under the same conditions.

Total RNA extraction and gene expression analysis
To analysis OsATG8b expression responsive to N limitation in rice, rice seedlings cultured with half Hoagland's solution (NS) as previous described [30] for 14 days, then transferred to the same solution as control, low N (NL) solution (with 0.6 mM KNO 3 and 0.1 mM (NH 4 ) 2 SO 4 , pH = 5.7) and the N-free (ND) solution (pH = 5.7), the lack of K + was replaced with KCl. After 1 day and 3 days N limitation treatments, the leaves and roots were harvested separately for OsATG8b gene expression analysis. 10-day-old seedlings of transgenic lines and WT grown on NS were sampled for expression analysis of the endogenous AtATGs by realtime RT-PCR. Total RNA extraction was conducted using the Eastep Super Total RNA Extraction Kit (Promega) and first strand cDNA was synthesized with the PrimeScript RT Master Mix (TaKaRa). Real-time RT-PCR was performed as described previously [51] by using SYBR Premix Ex TaqII (TaKaRa) on an Applied Biosystems 7500 Real Time PCR System. The following standard thermal profile was used for all PCRs: 95˚C for 30 s, 40 cycles of 95˚C for 5 s and 60˚C for 34 s. All reactions were done at least in triplicates. TIP41 was used as an internal control. The primers used for RT-PCR analysis were listed in S1 Table.

Binary vector construction and Arabidopsis genetic transformation
The complete coding region of OsATG8b was amplified by PCR using a pair of primers with full-length cDNA of rice as the template. The PCR product was then ligated into binary vector pCAMBIA1301, resulting in the 35S-OsATG8b fusion gene. For the Pro OsATG8b -GUS construct, a 1988-bp promoter fragment of the OsATG8b was amplified from rice genomic DNA by PCR, and then ligated into binary vector pCAMBIA1301. All the construction was confirmed by DNA sequencing. The plant expression vector was transformed into Arabidopsis by Agrobacterium-mediated floral dipping method [52]. All the primers used were listed in S1 Table. T 0 transgenic Arabidopsis seeds were screened on ½ MS medium containing 30 mg�L -1 hygromycin and seedlings with green true leaves were identified as transformants, 3 independents homozygous T 3 lines were analyzed in this study.

Protein isolation and immunoblot analysis
Total proteins extraction was conducted using the Minute™ Total Protein Extraction Kit (Invent Biotechnologies, Inc). Total membrane protein extraction was conducted using the Minute™ Membrane Protein Isolation Kit for plants (Invent Biotechnologies, Inc). 14-day-old seedlings of transgenic lines and WT grown on NS were sampled to exam the protein level of OsATG8b, the full length OsATG8b was used to raise a polyclonal antibody (GenScript, Nanjing China). Protein concentration was determined with the BCA Protein Assay Kit (TaKaRa). For autophagic flux analysis, 9-day-old seedlings of transgenic lines and WT were transferred to NS or ND medium with 0.5 μM ConA or solvent control DMSO for 24h. All SDS-PAGE gels were prepared with 6 M urea, antibodies against Arabidopsis anti-ATG8a and anti-NBR1 (1:1000), anti-α-tubulin (1:1000) (Agirsera, Sweden) and anti-OsATG8b (1:1000) were used as primary antibodies; goat anti-rabbit (IgG) (Sigma-Aldrich) was used as the secondary antibody. Immunoblotting were done as described [28,30]. The protein band intensities of immunoblots was analyzed by Image J software. Each experiment was repeated for three times, and one representative result was shown.

Phenotypic analysis of OsATG8b transgenic Arabidopsis under optimum condition and N starvation
For the basal phenotypic analysis of transgenic lines and the WT under sufficient nutrient condition, 8-day-old (days after germination) seedlings were transferred to soil. The maximum rosette radius and the plant height were recorded every 3 days, the bolting and flowering times were recorded every day from Stage5.1 and Stage6.0 [53], respectively. The leaf number at bolting was counted when the inflorescence emergence at the Stage5.1, and the total leaf number was counted after the rosettes growth completed at about Stage6.0. Total number of siliques was counted when the plants growth to Stage6.9, the total seeds produced by a single-plant were harvested at the growth Stage9.7, and the yield per plant was expressed as the weight of total seeds per plant, and the thousand-grain weight were weighed. For N starvation treatments, 7-day-old seedlings of transgenic lines and WT were transferred to NS or ND for 9 days. The phenotypic changes were recorded by scanner (EPSON Perfection v33), the fresh weight of leaves was measured.

GUS histochemical staining
Seeds of Pro OsATG8b -GUS transgenic Arabidopsis were screened on ½ MS medium containing 30 mg�L -1 hygromycin for temporal and spatial expression patterns analysis, then GUS histochemical staining was conducted as described [54].

MDC staining for detecting autophagosome activity
7-day-old seedlings of transgenic lines and WT were transferred to ND liquid medium containing 1 μmol�L -1 ConA with or without 5 mM 3-MA for 12h, then the leaves were excised and immediately vacuum infiltrated with 50 mM MDC (Sigma-Aldrich) for 10 min, MDC staining was conducted as described [48], and confocal images were acquired using an inverted Zeiss LSM 780 laser scanning microscope. Quantification of the MDC-stained autophagosomes in leaves was presented as the fluorescence ratio = positive signal/area. The number of positive signals was calculated from areas of 400 μm 2 , and more than 10 independent areas were used for the quantification.

Measurement of chlorophyll, soluble proteins, and nitrate reductase (NR) enzyme activity
Chlorophyll content was measured with Arnon method as describe [55]. Soluble protein content was determined with the Bradford method [56]. The NR enzyme activity was assayed as described previously [57].

Biomass, total N, C concentrations and indicator used to monitor NUE calculations
Plants were sampled by separation into rosettes (including cauline leaves), stems, and seeds at harvest, then all samples were kept in a dry oven at 80˚C for 7 days, and dry mass were weighted as DW Dr (dry weight of rosettes and stems) and DW Seeds (dry weight of total seeds). Then the dry mass of all these tissue organs were ground to powder, total N and C concentrations (N% and C%) were measured using an Elemental Analyzer (Elementa, Vario MAX C, Germany), which presented as N% (C%) = mg N (C)/100 mg DW. Biomass was calculated as DW Dr +DW Seeds . Indicator used to monitor NUE were calculated as described [20]. The harvest index (HI) was calculated as DW Seeds /(DW Seeds +DW Dr ) ratio and was used as an important indicator for yield. The N harvest index (NHI) was used to evaluate grain N filling and was calculated as (N% Seeds ×DW Seeds )/(N% Seeds ×DW Seeds +N% Dr ×DW Dr ). The ratio of NHI/HI was estimated as the percentage of WT to compare the NUE performance among genotypes.

N-Labeling and NRE calculation
The 15 N-Labeling experiment was conducted as Guiboileau et al (2012) [20] described on WT and 35S-OsATG8b transgenic lines at Stage5.1 (long time before flowering), the unlabeled watering solutions was replaced with one containing the same nutrient composition except that plants were supplied with 15 NH 4 and 15 NO 3 (10.6 atom% excess) for 3 d. Afterwards, the pots were rinsed thoroughly with distilled water to remove the remaining 15 N, and unlabeled watering solutions were then applied until seed harvest. When the plants were grown to maturity, the plants were separated into remains [rosettes (including cauline leaves) and stems] and seeds. All samples were kept in a dry oven at 80˚C for 7 days, then dry mass was weighted and ground to fine powder to assess N remobilization using an elemental analyzer/isotope ratio mass spectrometry system (Thermo Scientific, Flash2000/Delta V Advantage), and unlabeled plants were also collected to determine the natural 15 N abundance. The 15 N abundance of each sample was expressed as a percent of total N and calculated as atom% sample (A%sample), A%sample = 100×( 15 N)/( 15 N+ 14 N). The natural abundance of 15 N (Acontrol%) obtained from unlabeled samples were considered using the average value 0.365% to calculate 15 N enrichment of labeled samples (E%), E% = A%sample-A%control. NRE in this experiment were estimated as described by Guiboileau et al (2012) [20], using the ratio of 15 NHI/HI, and they were estimated as the percentage of WT to compare the NRE performance among genotypes. 15

Statistical analysis
The experiments were repeated three times, and data from one representative experiment were shown here. The statistically significant differences of all data were calculated based on Student's t-test at significance levels of � P < 0.05 and �� P < 0.01.