Target of rapamycin (TOR) regulates the response to low nitrogen stress via autophagy and hormone pathways in Malus hupehensis

Abstract Target of rapamycin (TOR) is a highly conserved master regulator in eukaryotes; it regulates cell proliferation and growth by integrating different signals. However, little is known about the function of TOR in perennial woody plants. Different concentrations of AZD8055 (an inhibitor of TOR) were used in this study to investigate the role of TOR in the response to low nitrogen (N) stress in the wild apple species Malus hupehensis. Low N stress inhibited the growth of M. hupehensis plants, and 1 μM AZD alleviated this effect. Plants supplied with 1 μM AZD had higher photosynthetic capacity, which promoted the accumulation of biomass, as well as higher contents of N and anthocyanins and lower content of starch. Exogenous application of 1 μM AZD also promoted the development of the root system. Plants supplied with at least 5 μM AZD displayed early leaf senescence. RNA-seq analysis indicated that TOR altered the expression of genes related to the low N stress response, such as genes involved in photosystem, starch metabolism, autophagy, and hormone metabolism. Further analysis revealed altered autophagy in plants supplied with AZD under low N stress; the metabolism of plant hormones also changed following AZD supplementation. In sum, our findings revealed that appropriate inhibition of TOR activated autophagy and jasmonic acid signaling in M. hupehensis, which allowed plants to cope with low N stress. Severe TOR inhibition resulted in the excessive accumulation of salicylic acid, which probably led to programmed cell death in M. hupehensis.


Introduction
Apple is one of the most economically important fruits in the world [1]; it is also one of the main fruit tree species in northern China, such as the Bohai Bay area and the Loess Plateau. Apple orchards mainly occur in barren areas, and nutrients in the soil in such areas are often depleted. Nitrogen (N) is an essential nutrient that plays a key role in plant growth, development, and production; it is an integral element of the basic components of life, such as proteins and nucleic acids [2]. However, >98% of N in soil is organic N, which plants are not able to use efficiently [3]. The use of chemical N fertilizers in orchards has increased to promote the development of apple trees and the production of apple fruits. But the application rate in most orchards exceeds demand and results in negative consequences, such as low N utilization efficiency, which exacerbates shortages of N in apple trees. The part of N not absorbed would eventually leach into rivers or evaporate into the air, accelerating environmental degradation [4,5]. Achieving high yield under N-limited conditions and reducing N fertilizer input are major challenges in apple production and require urgent attention.
N is an essential element for the synthesis of nucleic acids, proteins, lipids, and hormones. N deficiency affects cell division and differentiation, as well as many other physiological and biochemical processes, and alters plant morphology [6]. In the absence of N, chlorophyll synthesis is impaired; the subsequent decline in the rate of photosynthesis results in lower biomass production [7]. Plants have evolved several mechanisms to cope with N starvation stress. They can modify the architecture of their roots, especially the deep roots, and promote the growth of lateral roots to enhance the efficiency of N absorption from the soil [8]. The transport of N from roots to shoots and N utilization also change in response to low N stress [9]. These processes involve a complex signaling network consisting of different regulators, like plant hormones. For example, auxins have been shown to be involved in the response of plants to low N stress. They enhance N absorption from soil by promoting root elongation and lateral root formation [10]. Cytokinins are signals of N saturation in plants [11]. Abscisic acid (ABA) inhibits the expression of ABI2, which is an enhancer that promotes N uptake by regulating NPF6.3 (also known as NRT1.1) [12]. Ethylene and salicylic acid (SA) also function in the response of plants to low N stress [13,14]. The jasmonic acid (JA) signal could mediate large-scale systemic changes in N assimilation [15].
Target of rapamycin (TOR) is a serine/threonine protein kinase that is highly conserved in eukaryotes; it integrates nutrient, light, hormone, stress, and energy signals to regulate various biological processes, thereby promoting cell proliferation and growth [16]. Nutrient uptake and assimilation are central to the life activities of plants; TOR is a central node involved in the control of cell growth and can regulate signaling networks by sensing nutrient signals such as carbon (C), N, sulfur (S), and phosphorus [16,17]. It coordinates a variety of complex processes to regulate the C/N balance in aboveground and belowground organs. N can regulate the growth and development of apical and shoot tips by activating the ROP2-TOR pathway, and N depletion strongly inhibits TOR activity [18]. TOR inhibition induces autophagy in plants [19], which is a conserved degradation and recycling mechanism in eukaryotes, and is involved in plant development [20], nutrient recycling [21], and responses to stresses [22,23]. When plants experience N starvation, autophagy degrades unwanted or damaged organelles and proteins to recycle N, and promotes the absorption and assimilation of N to increase N use efficiency [24,25].
TOR is a key regulator of various biological processes in organisms in response to environmental stimuli. Rapamycin is often used in humans and yeast to characterize the biological function of TOR and pathways relevant to it. However, most terrestrial plants are insensitive to rapamycin and show embryonic lethality when TOR is mutated. Consequently, several ATPcompetitive TOR-specific inhibitors have been developed and used in various studies, including AZD8055, Torin1, Torin2, and KU0063794, which can effectively inhibit TOR kinase activity and alter plant growth [26,27]. Here, we studied the role of TOR in Malus hupehensis plants under low N stress. We used a hydroponic system to impose low N stress on M. hupehensis plants and used AZD8055 (hereafter referred to as AZD) to inhibit TOR kinase activity. The effect of TOR on the growth of M. hupehensis under low N conditions is discussed, as well as the underlying mechanism of the TOR signaling pathway. Our results provide new insights into the role of TOR in the response of plants to low N stress, especially in perennial woody plants.

Effects of TOR inhibitor on the growth of M. hupehensis under low N stress
To evaluate the effects of TOR inhibition on the growth of M. hupehensis under low N stress, plants were cultured in a hydroponic system with normal (6 mM) and low N (0.2 mM) levels for 35 days and supplied with different concentrations of AZD. Compared with plants under control growth conditions (NC group), plants under low N stress had a stronger root system but chlorotic leaves. Among the six groups under low N stress (supplemented with 0, 0.5, 1.0, 2.5, 5, and 10 μM AZD and referred to as groups 0A, 0.5A, 1A, 2.5A, 5A, and 10A, respectively), the roots were longest in group 1A, and the leaves in this group were as green as those in the NC group. The roots were the second longest in the 0.5A and 2.5A groups; plants from the 5A and 10A groups were smaller and weaker, and leaf margin necrosis could be observed in them ( Fig. 1A and B). After treatment, values for fresh and dry weight were much lower in the 0A group than those in the NC group, and were highest in the 1A group among the six low N groups, followed by the 0.5A group, in which values for fresh and dry weights were much higher than those in the other three groups (2.5A, 5A, and 10A) ( Fig. 1C and D). In addition, the N content of the leaves and roots in the low N groups was much lower than that in the NC group at the end of the treatment period. However, the exogenous application of AZD increased the N content when it was applied at a suitable concentration (Fig. 1E).
Low N promoted the growth of roots (0A versus NC), and values for root length, root surface area, average root diameter, root volume, and numbers of tips and forks were higher (Fig. 1A, Table 1). The root architecture was also altered under TOR inhibition when different concentrations of AZD were applied in the hydroponic system. Values for all the root parameters increased with low concentrations of AZD and decreased with high concentrations of AZD (Table 1). These results indicated that low N conditions suppressed the growth of M. hupehensis plants, whereas the application of 1 μM AZD effectively alleviated the adverse effect of low N stress on M. hupehensis.

Effects of TOR inhibitor on the photosynthetic capacity of M. hupehensis under low N stress
Inhibition of TOR activity changed the biomass accumulation of plants under low N conditions; thus, we evaluated the photosynthetic capacity of plants. We made continuous measurements of the gas exchange parameters for all groups and found no difference in the net photosynthetic rate (Pn), stomatal conductance (Gs), or intercellular CO 2 concentration (Ci) among them at the beginning of the treatments. Pn decreased in response to low N stress, and the differences among the seven groups became more pronounced over time. Under low N stress, plants supplied with a relatively low concentration of AZD had higher Pn values (e.g. 0.5A and 1A versus 0A); by contrast, plants supplied with a relatively high concentration of AZD had lower Pn values (e.g. 5A and 10A versus 0A) over time. At the end of the low N treatments, the lowest Pn value was found for the 10A group and the highest Pn value was found for the 1A group ( Fig. 2A). Data are mean ± standard error of 10 replicate samples. Different lowercase letters indicate statistically significant differences between treatments according to Tukey's multiple range test (P < .05). NC, plants without any treatment; 0A, low N stress treatment (0.2 mM N); 0.5A-10A, low N stress with 0.5 μM (0.5A), 1 μM (1A), 2.5 μM (2.5A), 5 μM (5A), and 10 μM (10A) AZD8055 treatments. The number of tips and forks is a measure of the branchiness of the fine root system; a higher number of tips and forks means easier uptake of water and nutrients through the fine roots. All data are mean ± standard error of 10 plants. Percentages in parentheses after the standard error are for data in 0.5A, 1A, 2.5A, 5A, and 10A groups versus the 0A group. Different lowercase letters indicate statistically significant differences between treatments according to Tukey's multiple range test (P < .05). NC, plants without any treatment; 0A, low N stress treatment (0.2 mM N); 0.5A-10A, low N stress with 0.5 μM (0.5A), 1 μM (1A), 2.5 μM (2.5A), 5 μM (5A), and 10 μM (10A) AZD8055 treatments.
Gs was much higher in the 1A group than in the other six groups, and it decreased over time in most groups (Fig. 2B). Ci fluctuated almost in the same pattern over time in all seven groups (Fig. 2C).
In addition, we compared the maximum photochemical efficiency of photosystem II (PSII) (Fv/Fm) among the seven groups. The Fv/Fm value was ∼0.8 in the NC group. After treatment with low N for 35 days, the value of Fv/Fm gradually decreased to ∼0.4 in the 0A group, and the exogenous application of AZD reversed this effect. The Fv/Fm value first increased and then decreased as the AZD concentration increased, and the highest value was observed in the 1A group, in which it was nearly the same as in the NC group ( Fig. 2D and E). These data indicated that the reduction in the photosynthetic capacity of plants under low N stress was mitigated by exogenous AZD supplementation, especially by 1 μM AZD.

Differential effects of TOR inhibition on accumulation of starch and anthocyanins in M. hupehensis under low N stress
We examined the effects of two extreme concentrations of AZD (1 μM for the positive role and 10 μM for the negative role) on the response of plants to N deficiency. C metabolism is tightly coupled to N metabolism, and starch is one of the two forms of carbohydrate stored in plant cells after C fixation [28]. In this study, we found that M. hupehensis plants accumulated large amounts of starch after low N treatment, as their leaves were shallow-or deep-stained blue. Among the four groups, the darkest blue was observed in the 0A group, followed by the 10A group. Leaves in the 1A group were stained light blue, and leaves in the NC group were almost white with a few blue flecks on the edge (Fig. 3A). This pattern was also consistent with the starch content observed in the four groups (Fig. 3B). These results demonstrated that the inhibition of TOR activity promoted the degradation of starch in N-deficient plants.
N deficiency modifies the secondary metabolism of plants and promotes the production of plant pigments such as anthocyanins, which could function as antioxidants in response to N deficiency [29]. Here, we found that the upper stems of plants turned red under low N conditions. The petioles of the upper leaves also turned red in response to low N stress. Among the four groups, deeper red was observed in the 1A group (Fig. 3C). We then measured the content of anthocyanins, which are responsible for red color of plants. Plants in the 1A group accumulated 12.86 nmol/g FW anthocyanins, which was approximately two and a half times higher compared with plants in the 0A group, and nearly three times higher compared with plants in the NC group (Fig. 3D), indicating that the beneficial role of appropriate TOR inhibition was partly linked to the accumulation of anthocyanins under low N stress.

Differential effects of TOR inhibition on transcription level in M. hupehensis under low N stress
Gene expression and its regulation were evaluated in response to TOR inhibition under low N stress. A total of 3923 differentially expressed genes (DEGs) were identified, which were clustered into eight groups according to their expression profiles (P < .05, log2 fold change >1) (Supplementary Data Fig. S1). Comparisons between groups 0A and NC, 1A and 0A, and 10A and 0A revealed 199, 122, and 2500 upregulated DEGs, respectively, and 1079, 4, and 1151 downregulated DEGs, respectively (Fig. 4A). Upregulated DEGs in the 1A versus 0A group were involved in hormone metabolism, protein degradation, and protein post-translation modification pathways, especially the former two. In the 10A versus 0A group, 2500 upregulated DEGs could be found in almost all identified clusters; moreover, downregulated DEGs were involved in protein synthesis, light signaling, major carbohydrate metabolism, and cell wall (Fig. 4B).
The transcriptional profiles of DEGs involved in several biological processes were analyzed in different groups, including photosynthesis, N, starch, and anthocyanin metabolism, protein degradation, and plant hormone metabolism (Supplementary Data Table S2). Several DEGs are shown in Fig. 5. Genes involved in anthocyanin anabolism and starch degradation were highly expressed in the 1A group, and genes involved in anthocyanin reductase and starch synthesis were highly expressed in the 10A group, consistent with the content of anthocyanins and starch we detected in the four groups. Genes involved in protein degradation were grouped into two clusters; some of them were highly expressed in the 10A and NC groups, and the others oppositely were highly expressed in the 0A and 1A groups (Supplementary Data Table S2). However, the expression of all genes involved in protein degradation through autophagy peaked in the 1A group (Fig. 5). Moreover, genes involved in plant hormones were divided into four categories according to their functions, comprising synthesis, degradation, signal transduction, and inducedregulated-responsive-activated. Most of them were suppressed by low N stress (0A versus NC); among groups under low N stress, the expression of most genes was highest in the 10A group (Supplementary Data Table S2). For example, most genes involved in JA and SA synthesis and their related signal transduction pathways were highly expressed in the 10A group compared with the others (Fig. 5).

Differential effects of TOR inhibition on autophagic activity in M. hupehensis under low N stress
TOR is the major negative regulator of autophagy [30]. We found that all genes involved in autophagy were upregulated in the 1A group in the RNA-seq data; thus, we performed several experiments to investigate how TOR affects autophagy in M. hupehensis under low N stress. We first analyzed the transcription levels of several autophagy-related genes (ATGs) [31,32], and found that most of them were induced by low N stress (0A versus NC). The expression of MdATG3b, -8c, and -12 was also increased, but not as much as other genes. In addition, application of 1 μM AZD increased the expression of all 12 candidate MdATGs. For example, the expression of MdATG8c, -10, and -12 in the 1A group was nearly twice that in the NC group (Fig. 6A).
Transmission electron microscopy was used to observe autophagosome formation in leaves under low N stress ( Fig. 6B and C). At the end of the experiments, plants in the NC group contained few autophagosome structures, whereas plants in the 0A group had a greater number of these structures. Three times as many autophagosomes and autophagic bodies could be observed in the 1A group (1A versus 0A). Overall, these data suggested that autophagic activity was significantly enhanced by appropriate inhibition of TOR activity in M. hupehensis under low N stress.

Differential effects of TOR inhibition on hormone metabolism in M. hupehensis under low N stress
Several upregulated genes in the 1A versus 0A group were involved in hormone metabolism; they might be positively correlated with TOR inhibition under low N stress. Thus, we analyzed hormone metabolism in M. hupehensis plants. The accumulation of plant hormones changed dramatically in response to AZD, particularly in the 10A group; the total level of plant hormones was twice as high in the 10A group than in the NC group ( Supplementary Data Fig. S2). We finally detected 39 different metabolites, which were assigned to cytokinin (15), auxin (10) Table S3, Supplementary Data Fig. S3). The accumulation of most metabolites in the cytokinin, JA, ABA, ethylene, and strigolactone groups was inhibited, whereas the accumulation of most metabolites in the auxin and SA groups was promoted by low N stress (0A versus NC). Differential inhibition of TOR activity further changed plant hormone metabolism under low N stress (Fig. 7).
The accumulation of metabolites in the cytokinin, auxin, and gibberellin groups changed irregularly upon TOR inhibition. In the auxin group, the content of six metabolites was highest in the 10A group, and the content of the other four was highest in the 1A group. The accumulation of the active forms indole-3-acetic acid (IAA) and l-tryptophan increased as the concentration of AZD increased. The accumulation of degradation products such as IAA-Glu, IAA-Asp, and OxIAA changed irregularly upon AZD application. The accumulation of ABA, ABA-glucosyl ester (ABA-GE), and 1-Aminocyclopropanecarboxylic acid (ACC) changed little following AZD application.
The accumulation of metabolites in the JA (5) and strigolactone groups (5-Deoxystrigol (5DS) only) under low N stress was promoted by 1 μM AZD and inhibited by 10 μM AZD, respectively. The excess accumulation of SA resulted in hypersensitivity reactions. The accumulation of SA and Salicylic acid 2-O-β-Glucoside (SAG), the most abundant metabolites we detected (Supplementary Data Fig. S2), peaked in the 10A group, which was consistent with the observed necrosis in the leaves.
In summary, these findings indicated that the positive effect of appropriate TOR inhibition on M. hupehensis under low N stress might be related to JA. IAA metabolism was antagonistic to TOR activity in leaves, and SA might cooperate with TOR in inducing necrosis in plants.

Inhibition of TOR activity alters the response to low N stress in M. hupehensis
N is one of the main nutrients limiting plant growth and production, as it is the major element of most compounds in cells. N starvation prevents cells from synthesizing new proteins, nucleic acids, and other essential substances for plant growth, leading to growth retardation. N starvation also inhibits the synthesis of chlorophyll, reduces the photosynthetic capacity of plants, and results in yield decreases in plants [4,7]. In this study we did not observe dwarf plants and the roots were stronger under low N stress; growth retardation was exhibited by the obvious reduction in the biomass of plants under low N stress (Fig. 1, Table 1).
Plants have evolved various mechanisms to ensure their growth and development under various environmental conditions, and the TOR signaling pathway is a conserved and central regulatory hub in these processes in all eukaryotes [33]. It senses various signals that affect plants, including nutrients, hormones, and stress signals . More than 10 cells were used to quantify structures; data are mean ± standard error of 10 replicate samples. Different lowercase letters indicate statistically significant differences between treatments according to Tukey's multiple range test (P < .05). [34,35]. Inorganic N and amino acids can activate the TOR signaling pathway to promote development [36]. We also found that the TOR signaling pathway responded to low N stress, and there was a dosage effect between TOR inhibition and the low N stress response in M. hupehensis. When we applied AZD to M. hupehensis plants to inhibit TOR activity under low N stress, plants supplied with a low concentration of AZD (1 μM) accumulated more biomass and had stronger roots and higher photosynthetic efficiency; the opposite pattern was observed for plants supplied with a high concentration of AZD (10 μM) (Figs 1 and 2, Table 1). AZD at 1 μM alleviated the suppressive effect of low N stress on plants, whereas AZD intensified the suppressive effect beyond 5 μM, which confirmed that TOR plays a core role in regulating various biological processes in plants; the severe inhibition of its activity or its mutants results in the severe suppression of plant growth and can even induce lethal phenotypes.

Appropriate inhibition of TOR activity promotes starch catabolism and anthocyanin anabolism in M. hupehensis to alleviate low N stress
C-containing compounds are the basic and abundant organic compounds in plant cells; most of them cannot be synthesized without N. N deficiency leads to excessive C in cells, resulting in surplus carbohydrates stored in chloroplasts as starch [9,37]. This interplay between low N and starch accumulation also exists in M. hupehensis; we observed excessive accumulation of starch in leaves under low N stress (Fig. 3A). Generally, starch in chloroplasts is transitory; it is degraded into glucose when it enters C metabolism in plants. The excessive accumulation of starch in chloroplasts disrupts the distribution of C in plant organs, which leads to energy metabolism imbalance and eventually growth suppression [38]. TOR is a master regulator that regulates central energy metabolism, C partitioning, and growth in plants; inhibition of its activity results in the massive accumulation of storage lipids and starch [39,40]. This probably explains the high amount of starch observed in plants in the 10A group, in which TOR activity was severely inhibited. However, when the inhibition of TOR activity was weak, as it was in the 1A group, the accumulation of starch decreased (Fig. 3A and B). Furthermore, we found that several genes involved in starch synthesis, such as ISA1, SSI1, and GBSS1, were highly expressed in the 10A and 0A groups, whereas genes involved in starch degradation, such as BAM3 and PHS2, were highly expressed in the 1A group (Fig. 5). TOR can directly interact with these genes in the protein-protein interaction (PPI) network; it probably alters their expression to regulate the accumulation of starch under low N stress (Fig. 8A).
In addition, when sufficient N for synthesizing amino acids is not available, the excess of C-containing compounds might result in the synthesis of anthocyanins, which are the most abundant secondary metabolites in plants; they can then contribute substantially to low N stress tolerance in plant species [24,41,42]. We found that 1 μM AZD promoted the accumulation of anthocyanins in M. hupehensis under low N stress; the upper stems appeared redder than those in other groups ( Fig. 3C and D). Several genes involved in the synthesis of anthocyanins, such as LDOX, UGT78D2, and UGT88A1, were highly expressed in the 1A group (Fig. 5); these genes also directly interacted with TOR in the PPI network (Fig. 8A). However, the interaction between TOR and these genes, as well as the mechanism by which TOR regulates them, requires further experimental confirmation.

Appropriate inhibition of TOR activity promotes autophagy in M. hupehensis to alleviate low N stress
Precise regulation of the spatiotemporal expression of TOR is critical for plant growth and development. Appropriate TOR inhibition is beneficial for improving N use efficiency under low N stress, but excessive reduction of TOR activity leads to accelerated aging and abnormal apical dominance [18,43,44]. Autophagy is an important and highly conserved process in plants adapting to nutrient conditions. Changes in autophagy affect the remobilization, absorption, and transport of nutrient elements, which alters the adaptability of plants to N starvation conditions [45]. For example, enhanced autophagy promotes N uptake and assimilation in apples under low N stress [24]. It also controls S metabolism in leaves and facilitates S remobilization to seeds in Arabidopsis [46]. Autophagy regulates multiple genes in cells, and TOR is a negative upstream regulator of autophagy in both yeast and animals; this relationship has also been demonstrated in plants sensing nutrients [19,30]. We found that the expression of several ATG genes was induced by low N stress, along with the increased number of autophagosomes in cells, and appropriate inhibition of TOR activity (treated with 1 μM AZD) further promoted autophagy, as revealed by the higher expression of ATG genes and the greater number of autophagosomes (Fig. 6). TOR has been shown to inhibit autophagy at the The protein interaction relationships of targeted DEGs were determined based on the genome-wide PPI network of apple (GDDH13). The predicted protein associations were determined from the STRING database, and the PPI was built based on experimental, text mining, database, and co-expression evidence. The size of the node is the parameter of the connectivity of the network. Multiple copies of identical genes are shown as a single node. For example, the large green circle at the bottom of the figure represents five sequences predicted as the TOR gene in the RNA-seq data, including MD05G1107800, MD00G1031100, MD10G1086500, MD10G1112000, and MD15G1045400. They coincided completely in the PPI network. The top small green circle represents novel.4829 in the RNA-seq data. (B) Interaction analysis between predicted TOR and proteins in the JA pathway. The 618 amino acids of the MD05G1107800 N-terminal and 400 amino acids of the MD10G1086500 N-terminal were used to construct BD vectors. JMT, JAZ10, and NPR3 were separately cloned into AD vectors. The recombined BD and AD vectors were then co-transformed into yeast and grew on different media. T, L, H, and A are short for Trp, Leu, His, and Ade. The line connecting TOR and NPR3 was colored red after the interaction was confirmed in yeast. The blue line connecting TOR and JMT and the blue line connecting TOR and JAZ10 were added to the PPI network according to the results in yeast. post-translational level by over-phosphorylating ATG13, which prevents its association with ATG1; TOR inhibition under nutrient starvation fails to dampen autophagy and leads to low N tolerance in plants [47]. With the exception of ATG13, TOR likely alters the expression of other ATG genes to regulate autophagy in cells, as we observed a direct interaction between TOR and nearly all the ATGs detected in the RNA-seq data in the PPI network (Fig. 8A). Appropriate inhibition of TOR induced autophagy, which facilitated the adaptation of M. hupehensis to low N stress; however, severe inhibition of TOR (treatment with 10 μM AZD) failed to activate autophagy. Plants were sensitive to low N stress, and their growth was inhibited (Figs 1 and 6). SnRK1 is another central metabolic regulator, which antagonizes TOR in regulating autophagy under energy-deficient conditions [48]. SnRK1 activates autophagy via the TOR signaling pathway; it also regulates autophagy in a TOR-independent manner [49].
It might have been responsible for the suppression of autophagy in the 10A group; however, we did not detect changes in the expression of genes involved in the SnRK1 signaling pathway in the RNA-seq data. We found that the expression of TAP46, TIP41, and PP2A, which function downstream of TOR to positively regulate autophagy [33], was suppressed by 10 μM AZD; this might explain why autophagy decreased in the 10A group. Other regulators and signal pathways related to TOR and autophagy might also be responsible for this pattern; more work is needed to clarify these findings.

TOR regulates the metabolism of jasmonic acid and salicylic acid in M. hupehensis under low N stress
Plant hormone signaling activates responses to environmental stresses. Their integration with the TOR signaling pathway is essential and occurs frequently in plant cells [50][51][52][53]. The accumulation of JAs was higher in plants supplied with 1 μM AZD under low N stress. Under strong inhibition of TOR, plants accumulated SA instead (Fig. 7). JA and SA are important defense hormones and regulate plant growth and development. Exogenous methyl jasmonate improves growth characteristics and increases the lateral root numbers of M. hupehensis [54]. It also affects photosynthesis in a dose-dependent manner in Citrus plants [55]. These positive roles of JA likely contribute to the better growth of M. hupehensis when supplied with 1 μM AZD under low N stress. Although the crosstalk between TOR and JA remains unclear, TOR negatively affects the expression of genes related to JA biosynthesis and signal transduction in Arabidopsis and cotton [56,57]. It suppresses JA signaling under abiotic stress [50]. In this study, we demonstrated the interaction between TOR (N-terminal) and NPR3 (Fig. 8B), which is a negative regulator in the JA pathway and a positive regulator in the SA pathway [58], and TOR also interacted with Jasmonic acid carboxyl methyltransferase (JMT) and Jasmonate ZIM-domain protein 10 (JAZ10) (Fig. 8B). JAZ10 is also a negative regulator in JA signaling transduction, and JMT is involved in the synthesis of JA. In addition, we found that TOR regulates JA signaling via some bridges, such as MYBs and MPKs (Supplementary Data Fig. S4). The present results suggested that TOR regulates the JA signaling pathway both directly and indirectly. Additional studies are still needed to reveal the crosstalk between TOR and JA in M. hupehensis responding to low N stress.
SA acts as a developmental regulator at low concentrations; it stimulates systemic acquired resistance to regulate plant defense responses at high concentrations [59]. Excessive accumulation of SA leads to hypersensitive reactions, which promotes leaf senescence in plants, also a form of programmed cell death [60]. This process is characterized by the high expression of a specific set of senescence-associated genes (SAGs). In plants supplied with 10 μM AZD under low N stress, the expression of several SAGs, such as SAG101, SAG20, and SAG21 (Supplementary Data Table S2), was high, and leaf margin necrosis was observed (Fig. 1); SA content was also higher than in the other groups (Fig. 7), indicating the occurrence of programmed cell death. In sum, stronger TOR inhibition in the 10A group led to the excessive accumulation of SA, which resulted in early senescence in M. hupehensis under low N stress. Like JA, TOR antagonizes the action of SA in the plant immune system [61]. In addition, TOR might directly regulate the SA pathway; it interacted with UGT74E2 and UGT74F2, which are related to SA metabolism, in the PPI (Fig. 8A), and with NPR3, which is a positive regulator of SA signaling, in yeast (Fig. 8B).

Conclusions
In this study, we showed that low N repressed the growth of M. hupehensis. Application of AZD, an inhibitor of TOR, altered plant physiology profiles to mediate the response to stress (Fig. 9). In plants supplied with 1 μM AZD, biomass accumulation was greater, the root system was improved, the photosynthetic rate, N and anthocyanin contents were higher, and the starch content was lower. Further analysis revealed that autophagy was increased with the appropriate inhibition of TOR activity, which made plants recycle more nutrients under energy starvation. However, autophagy decreased when TOR was severely inhibited. TOR directly interacted with several ATGs in the PPI and altered their expression to regulate autophagy; it also indirectly regulated autophagy via TAP46. Finally, TOR altered plant hormone metabolism under low N stress. We suggest that TOR negatively regulates JA signaling in plants responding to low N stress; TOR inhibition also cooperates with SA to cause programmed cell death in M. hupehensis.

Plant materials and growth
M. hupehensis was used because of its high degree of apomixis, which results in superior consistency during its development. The seeds were collected in Pingyi County (35 • 07 N, 117 • 25 E), Shandong Province, China. We cultivated the seedlings as described by Gong et al. [62]. Briefly, after stratification for 50 days at 4 • C, germinated seeds were planted and cultured in a greenhouse. When six true leaves were fully expanded, seedlings were transferred to a hydroponic system with 1/2 Hoagland's nutrient solution [63]. All the experiments were carried out at Northwest A & F University, Yang Ling (34 • 20 N, 108 • 24 E), China.

TOR inhibitor application and low N treatment
To determine whether TOR plays a role in M. hupehensis under low N stress, different concentrations of the TOR inhibitor AZD were applied in the hydroponic system. All the seedlings were pre-cultured for 2 weeks, and uniform seedlings were selected and divided into a control group and a low N treatment group, with ∼6 mM N and 0.2 mM N in 1/2 Hoagland nutrient solution, respectively. AZD was applied to the low N group at six different concentrations (0, 0.5, 1.0, 2.5, 5, and 10 μM), which were refreshed every 5 days along with the nutrient solution.
Overall, there were two groups and seven different treatments (48 plants each) as follows: (i) normal conditions with 1/2 Hoagland nutrient solution (hereafter referred to as NC; the N level was ∼6 mM); (ii) low N conditions with 0.2 mM N in 1/2 Hoagland nutrient solution, with 0, 0.5, 1.0, 2.5, 5, and 10 μM AZD supplementation separately; these six latter treatments were subsequently referred to as 0A, 0.5A, 1A, 2.5A, 5A, and 10A, respectively. The stress period ended after 35 days, and different parts of the selected plants were collected for measurements.

Growth measurements
At the end of the experiment, 10 seedlings were randomly selected from each group to measure the growth parameters, including fresh and dry weights and root architecture according to Gong et al. [62]. Root architecture was first captured using an Epson V700 scanner (Epson, Japan), and root morphological indexes, including root length, surface area, average diameter, root volume, numbers of root tips and forks, were analyzed using the WinRHIZO Pro root analysis system (Regent Instruments, Quebec, Canada).

Evaluation of photosynthetic efficiency and related parameters
The photosynthetic capacity of the seedlings was determined every 5 days at 9.00 a.m. during the study period using a CIRAS-3 portable photosynthesis system (PP Systems, USA), according to Jia et al. [32]. Ten individually mature and fully expanded leaves were measured for each treatment. The chlorophyll fluorescence was detected before and after the treatments according to Jia et al. [31], using an Open FluorCam FC 800-O imaging fluorometer (Photon Systems Instruments), and analyzed with Fluorcam7 software (PSI, Brno, Czech Republic).

Determination of N, starch, and anthocyanin contents
After determining the dry weight, the samples (leaves and roots) were ground and sieved through a 60-mesh sieve. Samples (0.1 g) were then used to detect the N content in plants in different treatments, according to Sun et al. [24].
To dye leaf starch, the middle mature leaves of the plants were boiled in 95% absolute ethanol until the leaves turned white. They were then washed twice in ultrapure water and dyed in 5% Lugol's solution (5% w/v I 2 and 10% w/v KI) for 10 minutes, and then destained in water until a clear background was obtained. After that, leaves were photographed on a white background plate. Quantitative determination of starch and anthocyanin contents was carried out after extraction as described in a previous study [24].

Detection of autophagy
Leaf autophagosomes were observed by transmission electron microscopy following the methods of Jia et al. [32]. At the end of each treatment, leaves were cut into small pieces, fixed in 4% glutaraldehyde solution in 0.1 M phosphate-buffered saline (pH 6.8), and then stored at 4 • C for 12 hours. The autophagosomes were observed using an HT7700 transmission electron microscope (Hitachi, Japan). Quantitative PCR was conducted to analyze the transcript levels of ATG genes in different groups (Supplementary Data Table S1), according to Gong et al. [62]. There were four replicates for each sample.

Measurement of plant hormone levels
Leaf samples were collected at 10 days for quantification of plant hormone levels in the different groups. After being ground into powder in liquid N, samples (50 mg) were mixed with internal standard and extracted in 1 mL of methanol/water/formic acid (15:4:1, v/v/v). The extracted supernatant was evaporated until dry with an N gas stream, redissolved in 100 μL of 80% methanol, and filtered through a 0.22-μm filter for liquid chromatography-tandem mass spectrometry (MS/MS) analysis. The data acquisition instrument system included ultraperformance liquid chromatography (UPLC) (ExionLC™ AD, USA, https://sciex.com.cn/) and MS/MS (TRAP ® 6500+, USA, https://sciex.com.cn/). There were three replicates for each sample.

RNA-seq analysis
Leaf samples were collected at 10 days for transcriptome analysis following the RNA-seq methods of Hrdlickova et al. [64]. Total RNA was extracted, and the concentration and integrity of the RNA were assessed using a Qubit 2.0 fluorometer (Life Technologies, USA) and Agilent 2100 bioanalyzer (Agilent Technologies, USA), respectively. After removing rRNA, the rest of the sample was used to construct a library using the NEBNext ® Ultra™ RNA Library Prep Kit (NEB, USA) and diluted to 1.5 ng/μL. The insert size of the library was determined with the Agilent 2100 Bioanalyzer. The libraries were then sequenced on a Novaseq 6000 system. Raw data were filtered using fastp v0.19.3 (https://github.com/OpenGene/fastp); clean reads were retained for analysis, which were then mapped to the reference apple genome GDDH13 [65]. To estimate the expression of genes, the FPKM (fragments per kilobase of transcript per million reads) value of each gene was calculated based on the length of the gene and the number of reads mapped to each gene. Differential expression analysis between the two groups was performed using the DESeq2 package (v1.22.1). Pvalues were adjusted using the Benjamini-Hochberg procedure [66]. In addition, DEGs were annotated based on Gene Ontology (GO), Kyoto Encyclopedia of Genes Genomes (KEGG), and MapMan pathways [67]. The PPI network of DEGs was constructed using the STRING database [68].

Yeast two-hybrid assay
The full-length sequences of three genes involved in JA (NPR3, JMT, and JAZ10) were cloned into the pGADT7 vectors, and the two predicted sequences of TOR (MD05G1107800 and MD10G1086500) in the RNA-seq data were divided into several parts according to their predicted domains and then cloned into pGBKT7 vectors. The yeast two-hybrid assays were then performed as described in the Clontech protocol. All the primers are listed in Supplementary Data Table S1.

Statistical analysis
Statistical analysis was performed using SPSS 22.0 software (IBM, Chicago, USA). Differences among means were determined by one-way ANOVA and Tukey's multiple range test (P < .05).