The Arabidopsis AMOT1/EIN3 gene plays an important role in the amelioration of ammonium toxicity

We identified a novel Arabidopsis ammonium tolerance 1 (amot1) mutant and reveal that blocking of ethylene signaling reduces tissue H2O2 accumulation and enhances seedling tolerance to ammonium stress.


Introduction
Ammonium (NH 4 + ), an important source of nitrogen for many species (Kronzucker et al., 1997;Balkos et al., 2010), is frequently present in soils and in the atmosphere in significant quantities (Britto and Kronzucker, 2002;Dupre et al., 2009). However, NH 4 + is toxic at moderate levels, frequently achieved in soils, to most plants, in particular those used in temperate agriculture, with stunted root and leaf growth as major symptoms of toxicity (Britto and Kronzucker, 2002;Coskun et al., 2013). Several important physiological processes have been linked to excessive NH 4 + exposure, such as ionic imbalance, relationships with carbon biochemistry, energy consumption, and modifications of hormonal balance (Britto and Kronzucker, 2002;Coskun et al., 2013). Ethylene production has been shown to increase linearly with tissue NH 4 + accumulation (Barker, 1999a), concurrent with the development of toxicity symptoms Barker, 2002, 2005). In addition, the application of ethylene synthesis and ethylene action inhibitors can ameliorate NH 4 + toxicity symptoms (Barker and Corey, 1991;Feng and Barker, 1992a, b;You and Barker, 2005; G. . Ethylene is synthesized from S-adenosyl-l-methionine (SAM) via 1-aminocyclopropane-1-carboxylic acid (ACC). The ratelimiting step in ethylene biosynthesis lies in ACC production via ACC synthase (ACS), followed by ACC conversion to ethylene by ACC oxidase (ACO) (Adams and Yang, 1979). ACS and ACO are encoded by multigene families and are regulated by both developmental and environmental factors. Evolution of ethylene in response to biotic and abiotic stress resulting from ACS and ACO up-regulation is a common phenomenon (Yamagami et al., 2003;Schellingen et al., 2014). In Arabidopsis, the downstream component of the ethylene signaling pathway includes ethylene receptor ETRs (e.g. ETR1), CTR1 (constitutive triple response 1), EIN2 (ethylene-insensitive 2), EIN3/ EIL (ethylene-insensitive 3/EIN3-like1), and ERF1 (ethylene-response factor 1) transcription factors (Lei et al., 2011). Ethylene signaling is transduced into the nucleus to cause the accumulation of two master transcriptional activators EIN3 and EIL1, which initiate transcriptional re-programming in various ethylene responses (Yoo et al., 2009;Tao et al., 2015). However, the detailed mechanisms of ethylene biosynthesis and signaling in NH 4 + stress responses remain unclear. Reactive oxygen species (ROS) are induced under a wide range of environmental stresses; they can cause oxidative damage leading to injury and death, depending on cellular concentrations. NH 4 + was found to induce higher levels of H 2 O 2 and oxidative stress response reactions in leaves of the aquatic plant Vallisneria natans , Arabidopsis (Podgórska et al., 2013), and other species . However, data regarding the pathways involved in NH 4 + regulation of H 2 O 2 production are still rare. ROS can be generated in the apoplast via the activity of NADPH oxidases under stress (Mittler et al., 2004). A group of NADPH oxidases and respiratory burst oxidase homologs (RBOHs) have been identified in Arabidopsis (Sagi and Fluhr, 2006). ROS are tightly regulated via a production/scavenging equilibrium. The expression of genes involved in the ROS regulatory network, including APX1 (ascorbate peroxidase 1) and CAT1 (catalase 1) in Arabidopsis (Davletova et al., 2005;Xing et al., 2008), also affects ROS levels. Peroxidases (PODs) have been proposed as alternative producers of ROS (Apel and Hirt, 2004;Bindschedler et al., 2006). PODs catalyze the oxidoreduction of various substrates using H 2 O 2 . PODs, rather than NADPH oxidase, have been proposed as the major ROS producer in French bean (Phaseolus vulgaris) treated with a cell wall elicitor of Colletotrichum lindemuthianum, the fungus that causes anthracnose (Bolwell et al., 1998). Kim et al. (2010) found that a POD contributes to ROS production during the Arabidopsis root response to potassium deficiency, showing the POD to be a component of the low potassium signal transduction pathway. Recently, Balzergue et al. (2017) showed that -Pi induces root tip ROS accumulation, indicating that PODs play a role. Further, the POD inhibitor salicylhydroxamic acid (SHAM) restored root growth and reduced ROS accumulation under -Pi conditions (Balzergue et al., 2017). Although ethylene and ROS have been reported to be involved in NH 4 + sensitivity, there has been no study to evaluate the role of ethylene in high-NH 4 + -induced ROS production in leaves, and further research is necessary to clarify the circumstances under which NH 4 + causes oxidative stress in plants. One approach to elucidating mechanisms of NH 4 + toxicity in plants is to use mutant lines. Qin et al. (2008) isolated the first NH 4 + -sensitive root elongation mutant, vtc1 (vitamin C defective 1), disrupted in GDP-mannose pyrophosphorylase (GMPase). Recently, several genetic regulators controlling root sensitivity to NH 4 + have been identified in Arabidopsis, such as aux1 (auxin resistant 1) (Cao et al., 1993;Li et al., 2011), trh1 (tiny root hair 1) (Zou et al., 2012), dpms1 (dolichol phosphate mannose synthase1) (Jadid et al., 2011), and gsa1 (gravitropism sensitive to ammonium1) (Zou et al., 2013). Elucidation of the function of these genetic regulators in determining the root sensitivity to NH 4 + offered insight into the molecular basis of historically described physiological responses to NH 4 + toxicity. In contrast, the underlying mechanisms of impaired leaf growth under NH 4 + toxicity are still largely unknown. Reduced shoot biomass and leaf chlorosis are important symptoms (Britto and Kronzucker, 2002). Based on their chlorotic phenotypes, ammonium-overly-sensitive 1 (amos1)  and amos2 (G.  mutants were recently identified: the AMOS1 locus is identical to EGY1 (ethylene-dependent gravitropism-deficient and yellow-green-like protein1), which encodes a membrane-bound, ATP-independent metalloprotease localized to plastids, required for chloroplast biogenesis (Chen et al., 2005). However, the genetic locus responsible for the amos2 mutation has not been identified. These studies, in combination, provide a significantly improved understanding of the process of NH 4 + toxicity in plants. Here, we report a novel Arabidopsis thaliana mutant, amot1 (ammonium tolerance 1), which displays enhanced shoot growth in response to NH 4 + stress. Gene cloning shows amot1 to be allelic to EIN3. Our results demonstrate that the disruption of AMOT1/EIN3 reduces high-NH 4 + -induced ROS accumulation in leaves, leading to reduced oxidative stress in the shoot. Moreover, AMOT1/EIN3 up-regulates shoot expression of the genes coding for PODs, previously shown to correlate positively with NH 4 + -induced changes in ROS content and cell growth inhibition.

Screening conditions
Transfer DNA (T-DNA) lines were constructed in the laboratories of D. Weigel and C. Somerville using the pSKI15 vector. Approximately 7500 independent lines (stock no. N21995) were provided by the ABRC. After surface sterilization, seeds were sown and grown on vertically oriented growth medium plates. After 5 d, seedlings were transferred to growth medium plates supplemented with 20 mM (NH 4 ) 2 SO 4 . Potential NH 4 + tolerance mutants were selected after 6 d and rescued, transferred to soil, and allowed to self-fertilize. The homozygous M 4 amot1 mutant was backcrossed to the WT Col-0, and the resulting F 1 generation was crossed with WT Col-0 twice to remove unlinked mutations caused by the mutagenesis.

Thermal asymmetric interlaced PCR
DNA for PCR amplification was extracted according to Weigel and Glazebrook (2002). Plant T-DNA-flanking sequences were amplified by PCR according to the protocols of Rodrigues et al. (2009).

Growth assays
For high-NH 4 + stress experiments, 5-day-old seedlings were transferred onto growth medium containing various concentrations of (NH 4 ) 2 SO 4 . Following 6 d of treatment, photographs were taken, and relative rosette size and shoot biomass were measured. To study the effect of precursors or inhibitors, the medium was supplemented with NH 4 + plus the indicated concentrations of ACC (Sigma), AgNO 3 (Shanghai yuanye biotechnology Co. Ltd, Shanghai, China), H 2 O 2 (Shanghai yuanye biotechnology Co. Ltd), or SHAM (Shanghai yuanye biotechnology Co. Ltd). The ratio of average rosette size on NH 4 + -stressed plates to the average rosette size on control plates was calculated as relative rosette size, according to Lei et al. (2011). The fresh weight of each individual shoot was measured immediately after harvest using a high-precision balance (0.000001) (XP105, Mettler Toledo).

NH 4 + , H 2 O 2 , MDA, and ACC content, and peroxidase and glutamine synthetase activity assay
Shoots (30-50 mg FW) of each sample were washed with 10 mM CaSO 4 , frozen in liquid nitrogen, and then extracted with 1 ml of 10 mM formic acid for the NH 4 + content assay by HPLC, following derivatization with o-phthaldialdehyde (Sigma) as described previously (G. . H 2 O 2 content was determined by the POD-coupled assay according to Veljovic-Jovanovic et al. (2002). Arabidopsis shoots were ground in liquid nitrogen, and the powder was extracted in 2 ml of 1 M HClO 4 in the presence of insoluble polyvinylpyrrolidone (5%). The homogenate was centrifuged at 12 000 g for 10 min, and the supernatant was neutralized with 5 M K 2 CO 3 to pH 5.6 in the presence of 100 ml of 0.3 M phosphate buffer (pH 5.6). The solution was centrifuged at 12 000 g for 1 min, and the sample was incubated for 10 min with 1 U of ascorbate oxidase (Shanghai yuanye biotechnology Co. Ltd) to oxidize ascorbate prior to use in the assay. The reaction mixture consisted of 0.1 M phosphate buffer (pH 6.5), 3.3 mM DMAB (Shanghai yuanye biotechnology Co. Ltd), 0.07 mM MBTH (Shanghai yuanye biotechnology Co. Ltd), and 0.3 U of POD (Shanghai yuanye biotechnology Co. Ltd). The reaction was initiated by the addition of the sample. The absorbance change at 590 nm was monitored at 25 °C. The malondialdehyde (MDA) level was measured using a thiobarbituric acid-reactive substance (TBARS) assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) (Ren et al., 2015). Shoot ACC content was detected by negative ion chemical ionization (NICI) GC-MS (Schellingen et al., 2014). Data are expressed in nmol g -1 FW. POD activity was detected by a micro-POD assay kit (BC0095, Solarbio, Beijing, China) according to the manufacturer's recommendations. POD activity is expressed as U mg -1 based on protein content. Glutamine synthetase (GS) activity was detected by a GS kit (BC0910, Solarbio) (Zhao et al., 2017). The specific enzyme activity (U g -1 FW) was defined as the amount of enzyme units catalyzing the transformation of 1 μM substrate per minute by the amount of fresh weight in grams.

Real-time quantitative PCR analysis
Total RNA was extracted from Arabidopsis shoots. Gene sequences were provided by the National Center of Biotechnology Information (NCBI), and gene-specific primers for real-time quantitative PCR (qRT-PCR) were designed using Primer-5 software (see Supplementary Table  S1 at JXB online). CBP20 (nuclear-encoded cap-binding protein) and ACTIN2 were used as the internal reference genes, and relative RNA abundance was normalized to the CBP20 or ACTIN2 internal control

Histochemical staining and image analysis
Histochemical staining of H 2 O 2 was performed as previously described (Dong et al., 2009) with minor modifications. Shoots were vacuuminfiltrated with 0.1 mg ml -1 3,3′-diaminobenzidine (DAB) in 50 mM Tris-acetate buffer, at pH 5.0. Samples were incubated for 24 h at room temperature in the dark prior to transfer to 80% ethanol. Histochemical analysis of β-glucuronidase (GUS) reporter enzyme activity was performed as described by Weigel and Glazebrook (2002). GUS or H 2 O 2 staining in the shoot was carried out using an Olympus SZX10 stereo microscope. Intensities of the GUS-and DAB-stained zone were quantified using Image-J software. All staining and image analysis procedures were repeated at least twice.

Ethylene measurements
After seedling exposure to 40 mM NH 4 + for varying durations, as indicated, shoots from the control and treatments were weighed and put into 5 ml gas-tight vials containing 1 ml of agar medium (0.7% agar). Headspace samples (1 ml) were withdrawn and analyzed using a GC-6850 gas chromatograph (Agilent Technologies Japan, Ltd), which was equipped with an FID detector.

Yeast one-hybrid (Y1H) analysis
Promoter fragments from At1g49570 (2217 bp) and At5g19890 (1692 bp) were cloned into the pAbAi vector to produce the bait constructs pAbAi-At1g49570 and pAbAi-At5g19890, respectively. The coding seuqence (CDS) of AMOT1/EIN3 was fused to the pGADT7 vector to generate a prey construct, AD-EIN3. The bait construct and the empty vector (AD) served as the negative control; p53-AbAi/pGAD-p53 were used as a positive control and transformed separately into yeast cells. Transformed yeast cells were diluted with a 10× dilution series and dotted onto SD plates lacking Ura and Leu (with or without antibiotic).

Statistical and graphical analyses
For all experiments, data were statistically analyzed using the SPSS 13.0 program (SPSS Chicago, IL, USA). Details are shown in the figure legends. Graphs were produced using Origin 8.0. All graphs and images were arranged using Adobe Photoshop 7.0.

Results
Enhanced tolerance of the amot1 mutant to ammonium toxicity Under our growth conditions, NH 4 + , at concentrations of 20-30 mM, caused slight reductions in shoot size and biomass ( Supplementary Fig. S1). A concentration of 40 mM significantly inhibited shoot size and biomass ( Supplementary  Fig. S1). To explore the mechanisms of NH 4 + -induced shoot growth inhibition, we performed a forward genetic screen for seedlings that show a shoot phenotype that was more resistant than WT plants when grown on medium containing 40 mM NH 4 + . Seedlings that appeared similar to the WT without NH 4 + but displayed significantly higher resistance of shoot growth to NH 4 + were amot (ammonium tolerance) mutants. We present the characterization of the amot1 mutant.
In agar plates without NH 4 + , the amot1 shoot growth phenotype was indistinguishable from that of the WT (Fig. 1A). However, when grown in the presence of high NH 4 + , amot1 shoot growth displayed greater resistance to NH 4 + than the WT (Fig. 1A).
NH 4 + -treated WT and amot1 plants showed a dose-dependent inhibitory effect of NH 4 + on the growth of aerial parts in response to a range of NH 4 + concentrations, but WT shoot growth was inhibited more than in amot1 at the concentrations used ( Fig. 1B, C). Furthermore, we analyzed the shoot phenotypes of amot1 and WT seedlings in response to 40 mM NH 4 + over time. The shoot growth between WT and amot1 seedlings remained similar 2 d after NH 4 + addition, but the difference was clearly accentuated under prolonged NH 4 + treatment, with the mutant maintaining significantly higher growth rates (Fig.  1D, E). Considering these results together, amot1 emerges as the first NH 4 + -resistant mutant, displaying superior shoot growth. The WT and amot1 seedlings were also treated on medium enriched with a variety of ions and molecules, and the results indicate that amot1 seedlings are highly resistant to both (NH 4 ) 2 SO 4 and NH 4 Cl, but responded to 15 mM and 20 mM K 2 SO 4 or 60 mM and 80 mM mannitol in a similar pattern to the WT (Fig. 2).
The amot1 mutant is a novel ein3 allele, and loss of EIN3 function enhances ammonium tolerance The WT as female parent was crossed with the homozygous mutant as the pollen donor. F 1 plants were selfed to obtain F 2 seeds. Both F 1 and F 2 seeds were assayed for growth on NH 4 + medium. All examined F 1 progeny (45 seedlings) displayed the same phenotypes as the WT. In the F 2 population, the amot1 phenotype segregated at an approximate 1:3 ratio (amot1:WT=54:142; x 2 =0.55, P>0.05), indicating that amot1 is a recessive mutation at a single nuclear locus. T-DNA-flanking sequences were isolated from the mutant by thermal asymmetric interlaced PCR, and sequence analysis revealed that the pSKI15 T-DNA was inserted into the exon of the EIN3 gene (At3g20770) in amot1, 192 bp downstream of the start codon, ATG [ Fig. 3A(a)]. EIN3 gene transcripts were greatly reduced in amot1 compared with the WT [ Fig. 3A(b) and (c)]. To ascertain further whether the NH 4 + -resistant phenotype in amot1 is due to the mutation in the EIN3 gene, we analyzed the previously reported allele ein3-1 (Chao et al., 1997) and crossed the amot1 mutant with ein3-1 plants. With exposure to 40 mM NH 4 + , shoot size and biomass in ein3-1 seedlings were indeed similar to those of amot1 seedlings (Fig. 3B). Furthermore, the amot1 mutant was crossed to ein3-1, and the F 1 progeny showed a phenotype similar to that of the parents in the presence of NH 4 + (Fig. 3B). Collectively, these results show that the amot1 mutant is a new loss-of-function allele of the EIN3 gene. AMOT1/EIN3 is a member of a protein family that includes EIN3-like (EIL) proteins (Chao et al., 1997) and initiates transcriptional re-programming in various ethylene responses (Guo and Ecker, 2004;Peng et al., 2014). We sought to determine the role of AMOT1/EIN3 in NH 4 + resistance. Because AMOT1/EIN3 and its close homolog EIL1 functionally overlap (Chao et al., 1997;An et al., 2010) 3C). Furthermore, in contrast to the ein3 single mutant's tolerant phenotype, the ein3-1eil1-1 double mutant exhibited a more tolerant phenotype than the amot1 mutant (Fig. 3C). We also examined the phenotype of the eil1 single mutant upon treatment with high NH 4 + ; however, the eil1-1 mutant was similar to the WT under high NH 4 + ( Supplementary Fig.  S2). Next, we examined the NH 4 + -responsive phenotype of a transgenic line overexpressing EIN3 under the control of the 35S promoter (35S:EIN3/Col-0 or EIN3ox), which displays an enhanced ethylene response (Chao et al., 1997;An et al., 2010;. The transcripts of AMOT1/EIN3 were significantly increased in the EIN3ox seedlings [Z. Fig. 3A(b) and (c)]. Under high NH 4 + , EIN3ox plants displayed increased sensitivity, based on shoot size and biomass, when compared with their WT counterparts (73% shoot biomass reduction in EIN3ox plants versus 52% in the WT at 40 mM NH 4 + ) (Fig. 3C). Together, these results suggest that constitutive overexpression of AMOT1/EIN3 leads to elevated shoot NH 4 + sensitivity in Arabidopsis. Consistent with a previous report (G. , the ein3eil1 lateral root number was also more resistant than that of the WT to high-NH 4 + stress ( Supplementary Fig. S3).

Enhanced shoot ethylene evolution is involved in ammonium-mediated inhibition of shoot growth
Aerial tissue NH 4 + content was determined, and the NH 4 + content increased gradually with treatment time compared with that in untreated shoots (Fig. 4A). Shoot ethylene production under NH 4 + exposure was also significantly greater than without NH 4 + and increased linearly with NH 4 + treatment time (Fig. 4B), consistent with Barker (1999b). Ethylene is synthesized from SAM via ACC, which is catalyzed by the enzyme ACS (Adams and Yang, 1979). ACC amounts in untreated and treated seedlings are presented in Fig. 4C. Consistent with previous reports, ACC amounts increased linearly with NH 4 + treatment time. As ACS and ACO are key enzymes of the ethylene biosynthetic pathway in plants, AtACS2, AtACS7, AtACS11, and AtACO2 expression was examined. Expression of the four genes was rapidly up-regulated in response to high NH 4 + (Fig. 4D). We further investigated the activity of AMOT1/EIN3 in response to NH 4 + in shoot. A transgenic reporter line harboring the GUS gene driven by five tandem repeats of the EIN3-binding site (EBS) followed by a minimal 35S promoter (5×EBS:GUS/Col-0) has been used to monitor the transcriptional activity of EIN3 (Stepanova et al., 2007;He et al., 2011). Following NH 4 + treatment, GUS staining became intensified in the cotyledons of 5×EBS:GUS/Col-0 plants (Fig. 4F), Fig. 2. Specificity of the amot1 mutant to NH 4 + . WT and amot1 seedlings were grown for 5 d on growth medium (GM) and then transferred to medium supplemented with salts and osmotica as indicated. Shoot fresh weight was measured 6 d after transfer. Growth on control nutrient (GM) was considered as 100%. Values are the means ±SD, n=10-14. Asterisks indicate statistical differences between the mutant and WT (independent samples t-test, *P<0.05).
indicative of elevated levels of AMOT1/EIN3 activity. We also observed that the expression of the ethylene-responsive gene ERF1 was up-regulated by NH 4 + in the WT (Fig. 4G). In keeping with the results on AMOT1/EIN3 activity, expression of ERF1, a direct target gene of EIN3 (Solano et al., 1998), was also lower in the NH 4 + -treated ein3eil1 mutant, but higher in EIN3ox lines, compared with the WT (Fig. 4G).
The WT plants treated with the ethylene biosynthetic precursor ACC displayed decreased tolerance to NH 4 + (Fig. 4E). Consistent with this, the ethylene overproduction mutant eto1-1 (ethylene overproducer 1) also showed reduced NH 4 + tolerance compared with the WT (Fig. 4H). In the presence of the ethylene receptor antagonist Ag + , shoot growth of the WT was significantly increased when the plants were exposed to NH 4 + stress (Fig. 4E). As ethylene is known to activate downstream signaling pathways by binding to ethylene receptors (e.g. ETR1), we examined whether ethylene regulates shoot growth sensitivity to NH 4 + in such a way. Shoot growth in the ethylene-insensitive (ethylene receptor) mutant etr1-3 and positive regulator mutants in ethylene signaling, amot1 and ein3-1, was more tolerant to NH 4 + than that of the WT; consistent with this, EIN3ox lines displayed increased shoot growth sensitivity ( Fig. 4H). These results indicate that ethylene has a negative effect on NH 4 + tolerance in Arabidopsis shoot growth.

AMOT1/EIN3 regulates ammonium-induced ROS accumulation in shoots
High NH 4 + induces an increase in ROS in plants; however, the biological mechanism of NH 4 + -induced ROS accumulation remains largely unknown. Here, we examined the levels of endogenous H 2 O 2 in the WT, ein3eil1, and EIN3ox in response to high-NH 4 + treatment. NH 4 + stress increased H 2 O 2 accumulation in the cotyledons of the WT, indicated by DAB staining (Fig. 5A, B). We further found higher levels of DAB staining in EIN3ox but lower levels in ein3eil1 than in the WT following NH 4 + stress (Fig. 5A, B), in accordance with the NH 4 + tolerance phenotypes of these genotypes (Fig. 3C). We also measured the H 2 O 2 contents in shoots. As inferred from DAB staining assays, high-NH 4 + treatment induced accumulation of H 2 O 2 in the WT. However, the level was lower in ein3eil1 and higher in EIN3ox shoots compared with Col-0 under NH 4 + stress (Fig. 5C). The H 2 O 2 content under control conditions was not significantly different among the three ecotypes ( Supplementary Fig. S4). A split-shoot experiment was devised to examine further the relationship between NH 4 + -induced AMOT1/EIN3 transcriptional activity and ROS accumulation ( Supplementary Fig. S5A). The cotyledon supplied with NH 4 + displayed significantly increased EBS:GUS expression compared with the untreated cotyledon ( Supplementary Fig. S5B). Consistent with EBS:GUS inductive sites, a higher intensity of DAB staining was also detected in the NH 4 + -treated cotyledon ( Supplementary Fig. S5C).
In this study, the concentration of MDA equivalents was increased in NH 4 + -treated leaves (Fig. 6A). However, we observed a higher MDA level in EIN3ox, and a lower level in ein3eil1 than in the WT following NH 4 + stress (Fig. 6A). The effects of NH 4 + on shoot growth were also examined in combination with external H 2 O 2 . The combined treatment with NH 4 + and H 2 O 2 in the medium markedly inhibited shoot growth compared with NH 4 + alone, and the combined treatment inhibited shoot growth more significantly in amot1 and ein3eil1 (Fig. 6B).
We further examined the expression of genes encoding antioxidant metabolic enzymes, such as APX1, APX2, CAT1, CAT2, and CAT3. NH 4 + stress did not induce APX and CAT gene expression in WT shoots, and these genes were also not much affected in ein3eil1 and EIN3ox seedlings under NH 4 + stress (Fig. 6C), suggesting that AMOT1/EIN3 regulation of NH 4 + -induced ROS accumulation might not be related to APX1-, APX2-, CAT1-, CAT2-, and CAT3-mediated antioxidant activity. Previous studies showed that drought stress increases RBOH transcript levels (Lee et al., 2012). However, the expression patterns of RBOHA, RBOHB, RBOHD, and RBOHF under control and high-NH 4 + stress were similar, and they were also not much affected in the ein3eil1 mutant and in EIN3ox lines compared with the WT (Fig. 6C). Furthermore, DAB staining and shoot growth in response to high-NH 4 + stress were also similar between the WT and the rbohd mutant ( Supplementary Fig. S6).  proposed that higher POD levels are positively correlated with NH 4 + -induced ROS generation and cell growth inhibition. We found that the expression of two of the genes encoding PODs, At5g19890 and At1g48570, was induced by NH 4 + treatment in the WT, and expression was more elevated in EIN3ox while it was reduced in ein3eil1 with or without NH 4 + (Fig. 7A, B). The transcript levels of other POD-encoding genes, such as At5g42180, At2g18980, At4g11290, and At3g49960, were not increased by NH 4 + treatment in the WT, and were also not significantly altered in stress conditions (Supplementary Fig. S7). The POD activity assay also showed that POD activity was significantly elevated in EIN3ox seedlings compared with the WT and ein3eil1 under NH 4 + stress (Fig. 7C). POD activity in EIN3ox seedlings under control conditions was also slightly elevated compared with the WT and with ein3eil1 ( Supplementary Fig. S8). We next analyzed the promoter regions of two POD genes (At5g19890 and At1g48570) and found EBSs (ATGTA) in each promoter (data not shown). To test the interaction between the AMOT1/EIN3 protein and the At5g19890 and At1g48570 promoters, a Y1H assay was performed. As shown in Fig. 7D, bait yeast cells co-transformed with the empty vector (AD) or the fusion vector (AD-EIN3) grew well on synthetic dropout medium (SD) without Ura and Leu. However, only the yeast cells co-transformed with the fusion vector AD-EIN3 survive on the selective medium supplemented with 400 ng ml -1 aureobasidin A (AbA; Fig. 7D). The data suggest that the AMOT1/EIN3 protein interacts with the At5g19890 and At1g48570 promoters in the yeast system. SHAM is a widely used POD inhibitor (Balzergue et al., 2017). The shoot biomass of the WT and EIN3ox lines was increased under NH 4 + stress after SHAM treatment, but no effects on ein3eil1 lines were observed (Fig. 7E, F). Furthermore, the POD inhibitor SHAM could decrease NH 4 + -induced DAB staining, indicating H 2 O 2 accumulation in both WT and EIN3ox leaves ( Fig.  7G; Supplementary Fig. S9).

The amot1 mutant accumulates less NH 4 + in shoot tissue under NH 4 + toxicity
There was no difference in shoot NH 4 + content of the WT, amot1, and ein3eil1 with 3 d of NH 4 + treatment (Fig. 8A), although ROS accumulation was greater in WT shoots than in those of ein3eil1 at this treatment time (Fig. 5). However, higher levels of NH 4 + in the WT, but lower levels in amot1 and ein3eil1, than in the WT were recorded following high-NH 4 + stress during a 6 d treatment (Fig. 8A). Consistently, NH 4 + accumulation was slightly higher in EIN3ox shoots than in the WT following the treatment (Fig. 8B). Activities of the enzyme GS, centrally involved in the NH 4 + assimilation process (Kronzucker et al., 1995;Hirano et al., 2008), was determined in shoots of the WT and amot1. GS activity was induced by external NH 4 + in NH 4 + -fed WT and amot1 plants, but this increase was not significantly different in the two genotypes (Fig.8C). This indicates that NH 4 + metabolism was not affected by the AMOT1 mutation.

Discussion
Stunted root system and decreased leaf bomass are among the most visible phenotypic manifestations of NH 4 + toxicity in higher plants (Britto and Kronzucker, 2002). Several genetic regulators controlling root sensitivity to NH 4 + have been identified in Arabidopsis ; however, little is known about the specific targets and pathways that lead to impaired leaf growth in the context of NH 4 + toxicity. To gain insight into the mechanisms of the effects of NH 4 + on shoot growth, we employed a molecular genetics approach, based on a mutant screen for altered response to NH 4 + . In the current work, enhanced NH 4 + tolerance of shoot growth was found in amot1. We further revealed that the nuclear AMOT1 locus is identical to EIN3, which encodes a transcriptional activator required for initiating transcriptional re-programming in various ethylene responses (Guo and Ecker, 2004;Yoo et al., 2009). It was found that amot1 and ein3-1, a reported allele, showed enhanced shoot growth tolerance compared with the WT, but the transgenic line overexpressing EIN3 (EIN3ox) was more sensitive. The activity of AMOT1/EIN3, indicated by using EBS:GUS in shoots, was markedly enhanced on NH 4 + (Fig. 4F). These results suggest that AMOT1/EIN3 plays an important role in the NH 4 + -induced impairment of shoot growth. It was demonstrated furthermore that this inhibitory effect is related to enhanced shoot ACC and ethylene accumulation. More importantly, it was found that AMOT1/EIN3 positively regulates shoot ROS accumulation, which leads to oxidative stress in Arabidopsis shoots under NH 4 + stress, and up-regulates the shoot expression of the genes coding for PODs previously shown to correlate positively with NH 4 + -induced increases in ROS content and cell growth inhibition .
A role for ethylene evolution has long been suggested under NH 4 + excess (Barker and Corey, 1991;Barker, 1999a, b), but its involvement remains incompletely understood. The present data indicate that ethylene evolution increases linearly with NH 4 + treatment time (Fig. 4B), consistent with previous reports (Barker 1999b). Foliar ethylene evolution increased sharply in tomato when foliar NH 4 + accumulation passed a critical value (Barker, 1999a). Our present and previous data also link the stimulation of EBS:GUS activity to NH 4 + exposure (G.  Supplementary Fig. S5). The present data show that shoot NH 4 + content increases linearly with Five-day-old seedlings were transferred to medium supplemented with NH 4 + alone or in combination with 2 mM H 2 O 2 for 6 d. n=7-10. (C) Effect of NH 4 + on gene expression of WT, EIN3ox, and ein3eil1 shoot tissue by qRT-PCR for 6 h. Values are means ±SD of three replicates. ACTIN2 was used as the internal reference gene, and the WT control was considered as 1. Different letters indicate statistical differences at P<0.05 (one-way ANOVA with Duncan post-hoc test).
increased treatment time (Fig. 4A). Furthermore, a previous observation showed high ethylene evolution to correlate with high tissue NH 4 + but to be independent of nitrogen form and pH regime (Feng and Barker, 1992c). Hence, together with previous reports, our study suggests that greatly increased shoot NH 4 + content may be the intrinsic trigger leading to enhanced ethylene evolution under NH 4 + stress. The rate-limiting step in ethylene biosynthesis lies in the production of ACC by ACS (Schellingen et al., 2014). Shoot ACC amounts also increased linearly with NH 4 + treatment time (Fig. 4C). Furthermore, we show here that AtACS2, AtACS7, AtACS11, and AtACO2, which encode ACS and ACO, the two key enzymes responsible for ethylene synthesis, are transcriptionally up-regulated by NH 4 + treatment (Fig. 4D). Therefore, it is conceivable that the increased ACC biosynthesis resulting from up-regulation of ACS and ACO gene expression is involved in NH 4 + -induced Fig. 8. Effects of NH 4 + treatment on shoot NH 4 + content and GS activity. (A) NH 4 + contents in the shoot tissues of WT, ein3eil1, and amot1 seedlings. Five-day-old WT, ein3eil1, and amot1 seedlings were grown on growth medium and transferred to fresh medium with control or NH 4 + for 3 d and 6 d of growth, and then NH 4 + tissue content was determined. Values are means ±SD of three replicates. (B) NH 4 + contents in the shoot tissues of WT and EIN3ox seedlings. Five-day-old WT and EIN3ox seedlings were grown on growth medium and transferred to fresh medium with control or NH 4 + for 3 d and 6 d of growth, and then NH 4 + tissue content was determined. Values are means ±SD of three replicates. (C) GS activities in the shoots of WT and amot1 seedlings. Five-day-old WT and amot1 seedlings were grown on growth medium and transferred to fresh medium with control or NH 4 + for 6 d, and then GS activity was determined. Values are means ±SD of three replicates. Different letters indicate statistical differences at P<0.05 (one-way ANOVA with Duncan post-hoc test).  A and B) qRT-PCR analysis of the expression of two POD genes (At5g19890 and At1g48570) in WT, EIN3ox, and ein3eil1 shoot tissue after NH 4 + treatment for 6 h. Values are means ±SD of three replicates. ACTIN2 was used as the internal reference gene, and the WT control was considered as 1. (C) Measurement of POD activity of WT, EIN3ox, and ein3eil1 shoot tissue. Seedlings were exposed to 40 mM NH 4 + for 5 d. Values are means ±SD of three replicates. (D) Y1H assay showing the EIN3 binding to the promoter of the two POD genes. The yeast expression plasmid pGADT7-EIN3 was reintroduced into the yeast strain Y1H Gold carrying the pAbAi-At1g49570 or pAbAi-At5g19890 vectors. The transformants (with or without dilutions) were screened for their growth on yeast synthetic defined medium (SD/-Ura -Leu) in the presence of 400 ng ml -1 AbA (antibiotic) for stringent selection. The empty vectors pGADT7 and p53-AbAi/pGAD-p53 were used as a negative and positive control, respectively. (E and F) Effect of salicylhydroxamic acid (SHAM) on the shoot biomass of the WT, EIN3ox, and ein3eil1. Five-day-old seedlings were transferred to medium alone or in combination with 10 μM SHAM for 6 d. A photograph of representative seedlings is shown in (E). Scale bars=1 cm. The shoot biomass is shown in (F). Values are the means ±SD, n=12. Different letters indicate statistical differences at P<0.05 of control and NH 4 + treatment, respectively (one-way ANOVA with Duncan post-hoc test). (G) Effect of SHAM on the NH 4 + -induced H 2 O 2 accumulation in shoots of WT and EIN3ox. Seedlings at 5 d were exposed to 40 mM NH 4 + with or without 10 μM SHAM for 3 d, and then DAB staining of shoots was performed. Scale bars=200 μm. ethylene evolution. Further study on the detailed mechanisms of NH 4 + -regulated ethylene evolution is warranted. In our study, we provide several lines of evidence supporting the notion that ethylene biosynthesis and signaling play a negative role in the adaptation of Arabidopsis shoot growth to NH 4 + stress. ACC content, ethylene production, and AMOT1/ EIN3 activity, and the expression of genes encoding key enzymes responsible for ethylene synthesis in the Arabidopsis shoot showed dramatic increases after NH 4 + treatment. The dual evidence that shoot biomass was inhibited by NH 4 + to a greater extent in the ethylene overproduction mutant eto1-1 and the EIN3ox line, compared with the WT, and that mutations in ethylene receptors (e.g. etr1-3) and key positive regulators in ethylene signaling (e.g. amot1 and ein3-1) showed increased shoot growth compared with the WT under NH 4 + stress (Figs 3, 4H), supports this notion. The observations that the externally supplied ethylene inhibitor Ag + alleviated the phenotypic manifestation of toxicity, but that ACC, a precursor of ethylene, aggravated NH 4 + -suppressed shoot growth in the WT (Fig. 4E), further demonstrate the important role of shoot ethylene signaling. Our findings are in excellent agreement with a previous study showing that Ag + improved plant growth on NH 4 + (Barker and Corey, 1991). Therefore, the data suggest that ethylene biosynthesis and signaling negatively regulate NH 4 + tolerance of shoot growth in Arabidopsis. The underlying mechanisms determining ROS accumulation by NH 4 + in leaves are only partially understood (Bittsánszky et al., 2015). Our results show that AMOT1/ EIN3 is involved in H 2 O 2 metabolism in leaves under NH 4 + stress. First, a higher level of H 2 O 2 cytochemical staining in EIN3ox was found, while a lower level of H 2 O 2 staining was seen in ein3eil1 than in the WT following NH 4 + stress (Fig.  5A, B), in accordance with the NH 4 + tolerance phenotypes and lipid peroxidation profiles of these genotypes (Figs 3, 6A). In agreement with the above results, NH 4 + stress increased leaf H 2 O 2 concentrations in the WT, while these were lower in ein3eil1 and higher in EIN3ox under identical treatment conditions (Fig. 5C). Moreover, our split-shoot experiment showed that a higher DAB staining intensity was detected in the components of EBS:GUS cotyledons exposed to NH 4 + ( Supplementary Fig. S5). These results suggest that AMOT1/ EIN3 positively regulates NH 4 + -induced leaf H 2 O 2 accumulation. Investigations on whether oxidative stress is involved in NH 4 + phytotoxicity have led to conflicting conclusions. The results of Domínguez-Valdivia et al. (2008) in spinach and pea suggest that stress originating from applying NH 4 + as the only nitrogen source is not an oxidative stress. However, evidence is accumulating that NH 4 + can induce oxidative stress in leaves, including in Arabidopsis , the aquatic plant Hydrilla verticillata (Wang et al., 2010), and duckweed (Huang et al., 2013). Our present results show that externally supplied H 2 O 2 increases shoot growth sensitivity of amot1 and ein3eil1 to NH 4 + stress (Fig. 6B). The extent of lipid peroxidation, estimated by monitoring the decomposition product MDA, has also been reported as elevated in NH 4 + grown seedlings (Hachiya et al., 2010;Podgórska et al., 2013;Fig. 6A). Together, these findings indicate that increased H 2 O 2 accumulation and oxidative stress in leaves under NH 4 + stress at least partially result from elevated shoot AMOT1/EIN3 activity, providing a molecular basis for NH 4 + -induced accumulation of H 2 O 2 in leaves. However, the present study shows that, although lower, there was still increased H 2 O 2 accumulation and oxidative stress in ein3eil1 mutant leaves under NH 4 + stress (Figs 5, 6A), suggesting that there are pathways through which AMOT1/EIN3 functions independently to regulate shoot H 2 O 2 accumulation and oxidative stress in response to NH 4 + stress. Drought can promote ROS biosynthesis by inducing expression of several Atrboh genes, such as AtrbohA, AtrbohD, and AtrbohE (Lee et al., 2012). However,  found that expression of the RBOHD gene is not induced in leaves under NH 4 + stress. In this study, we show that there is no increase in RBOHA, RBOHB, RBOHD, or RBOHF expression in NH 4 + -treated WT, EIN3ox, and ein3eil1 leaves. These results suggest that AMOT1/EIN3 is not involved through modulating expression of Atrboh genes under high NH 4 + , such as RBOHA, RBOHB, RBOHD, and RBOHF. However, we cannot exclude NADPH oxidase as a potential ROS source in NH 4 + -treated plants, as other RBOH genes might be up-regulated. NH 4 + -mediated changes in apoplastic pH (Husted and Schjoerring, 1995) may induce ROS generation, possibly through the modulation of the activities of cell wall PODs (Lager et al., 2010). Furthermore, under low-Pi conditions, increased POD activities also inhibited Arabidopsis root growth by regulating ROS levels and cell wall stiffening, and the POD inhibitor SHAM could restore root growth and reduce ROS accumulation under -Pi (Balzergue et al., 2017).  showed that higher POD levels positively correlate with NH 4 + -induced ROS content and cell growth inhibition. Consistent with this, by examining POD gene expression and POD activity under NH 4 + conditions, we show that NH 4 + stress increases expression of some POD genes (At5g19890 and At1g48570) and POD activity in WT shoots ( Fig. 7A-C). Moreover, the peroxidase inhibitor SHAM could indeed alleviate NH 4 + -induced shoot ROS accumulation ( Fig. 7G; Supplementary Fig. S9) and growth inhibition (Fig. 7E, F). This result, together with previous reports, confirms that NH 4 + induces accumulation of ROS and suggests that POD expression and activity may play an important role. Our qRT-PCR analyses show that two POD genes (At5g19890 and At1g48570) were constitutively up-regulated in EIN3ox but down-regulated in the ein3eil1 double mutant, regardless of NH 4 + (Fig. 7A). Moreover, NH 4 + -induced POD activity was positively correlated with the expression of EIN3 genes, as shown for ein3eil1 and EIN3ox seedlings (Fig. 7B). Further studies revealed that the key transcription factor AMOT1/EIN3 may directly target the two POD genes (At5g19890 and At1g48570), as the AMOT1/EIN3 protein could specifically bind to the promoters of the At5g19890 and At1g48570 genes, as revealed by the Y1H assay. Further supporting our finding of increased POD activity and DAB staining, indicating H 2 O 2 accumulation in EIN3ox, the POD inhibitor SHAM was shown to enhance EIN3ox shoot growth and reduce DAB staining used to indicate H 2 O 2 accumulation under NH 4 + stress ( Fig. 7; Supplementary Fig. S9). Collectively, these data indicate that the ethylene signaling-mediated NH 4 + response of Arabidopsis shoot growth is brought about, at least partially, through POD genes (e.g. At5g19890 and At1g48570), via AMOT1/EIN3.
The accumulation of free shoot NH 4 + is widely considered to be critical to the development of NH 4 + toxicity (Gerendas et al., 1997;Szczerba et al., 2008). Because H 2 O 2 accumulation was greater in WT shoots than in ein3eil1, but higher in EIN3ox than in the WT, after a 3 d treatment with high NH 4 + , we hypothesized that loss of function or overexpression of AMOT1/EIN3 may entail reduced or enhanced NH 4 + content in the shoot at the 3 d treatment time, respectively. However, there was no difference in shoot NH 4 + content between the WT, amot1, ein3eil1, and EIN3ox (Fig. 8A). These results further highlight that AMOT1/EIN3 plays an important role in regulating NH 4 + -induced shoot ROS accumulation and rules out that reduced shoot ROS accumulation during the early phase of exposure (within 3 d in our study) by loss of function of AMOT1/EIN3 resulted from reduced NH 4 + content. However, a higher NH 4 + content in EIN3ox and the WT, and a lower content in amot1 and ein3eil1 were observed following high-NH 4 + stress for prolonged treatment times (6 d) (Fig.  8A, B). Our results indicate that GS activity was not affected by the mutation in AMOT1/EIN3 (Fig. 8C), showing that GS is not responsible for the lower NH 4 + accumulation in the amot1 mutant. Alternatively, oxidative stress itself, induced by H 2 O 2 , can increase cellular NH 4 + concentrations by inducing proteolytic activity (Sweetlove et al., 2002). It is not clear how AMOT1/EIN3 mediates shoot NH 4 + accumulation over longer periods of treatment, and more research is warranted to examine this.
In summary, we have identified and characterized a novel gene, AMOT1/EIN3, that controls shoot NH 4 + sensitivity and propose a model for ethylene-AMOT1/EIN3 functions in shoot NH 4 + sensitivity (Fig. 9). Our study shows the importance of EIN3 in shoot inhibition under high-NH 4 + stress, providing strong genetic evidence in support of the role of the ethylene biosynthesis and signaling pathway in regulating shoot NH 4 + sensitivity. Under NH 4 + stress, ethylene is perceived and transduced, affecting the transcription factors EIN3/EIL and initiating the ethylene response. Additionally, our work highlights the roles of AMOT1/EIN3 in regulating ROS accumulation in Arabidopsis shoots under NH 4 + stress. In the ein3eil1 mutant, high-NH 4 + -induced ROS accumulation is reduced, which leads to reduced oxidative stress in the shoot. However, EIN3ox shoots accumulated more ROS and displayed higher sensitivity to NH 4 + stress. Elucidation of this interaction between AMOT1/EIN3 and H 2 O 2 signaling provides novel insight into our understanding on how EIN3 regulates shoot growth under high-NH 4 + stress. Moreover, it was found that AMOT1/EIN3 can up-regulate shoot expression of the genes coding for PODs (e.g. At5g19890 and At1g48570) previously shown to correlate positively with NH 4 + -induced ROS accumulation and cell growth inhibition. Further studies using molecular approaches to investigate the transcriptional network regulated by AMOT1/ EIN3 will provide important new insights into the process of acclimation and of adaptation to NH 4 + stress in plants.

Supplementary data
Supplementary data are available at JXB online. Fig. S1. Rosette diameter and fresh shoot weight of Arabidopsis thaliana wild-type (WT, Col-0) plants following treatment with various NH 4 + concentrations. Fig. S2. Relative rosette diameter (A) and fresh shoot weight (B) of Arabidopsis thaliana WT and eil1 mutant plants following treatment with NH 4 + for 6 d. Fig. S3. Lateral root number of Arabidopsis thaliana WT and ein3eil1 mutant plants following treatment with high NH 4 + for 6 d. Fig. S4. H 2 O 2 content in WT, EIN3ox, and ein3eil1 shoot tissue under control conditions. Fig. S5. EBS:GUS expression and DAB staining in the splitshoot experiment. Fig. S6. Effect of NH 4 + treatment on shoot DAB staining and biomass of the WT and the Atrbohd mutant. Fig. S7. qRT-PCR analysis of POD gene expression in WT, EIN3ox, and ein3eil1 shoot tissue under NH 4 + treatment for 6 h. Fig. S8. Measurement of POD activity of WT, EIN3ox, and ein3eil1 shoot tissue under control conditions. sensitivity. Based on our study and previous reports (Chao et al., 1997;G. Li et al., 2013;Podgorska et al., 2013Podgorska et al., , 2015, we established a model for ethylene-EIN3 function in shoot NH 4 + sensitivity. (A) In the wild type, NH 4 + stress enhanced the expression of ACS and ACO genes, encoding ACS and ACO, the two key enzymes responsible for ethylene synthesis. Under NH 4 + stress, ethylene is perceived and transduced, affecting the transcription factor EIN3, and initiating the ethylene response. EIN3 regulates ROS accumulation, which leads to oxidative stress in Arabidopsis shoots under NH 4 + stress. The expression of EIN3-mediated POD genes (e.g. At5g19890 and At1g48570) is involved in NH 4 + stress-induced shoot ROS accumulation. (B) In the amot1/ein3 mutant, expression of the AMOT1/EIN3-dependent POD genes (e.g. At5g19890 and At1g48570) in the shoot is blocked under NH 4 + stress. Ethylene regulation of ROS accumulation and oxidative stress is lowered. Orthogons in orange represent known EIN3 functions, and orthogons in gray with dashed lines represent the inhibition of EIN3 functions due to the amot1/ ein3 mutation. Red arrows indicate increased POD gene expression, ROS accumulation, and oxidative stress, and thick and thin red arrows indicate, respectively, a high or low ROS accumulation and oxidative stress. Fig. S9. Mean relative DAB staining intensity in WT (A) and EIN3ox (B) shoots treated with NH 4 + and NH 4 + plus SHAM. Table S1. Gene-specific primers used for qRT-PCR.