Arabidopsis seeds (qko, qko+AtAMT1;1, and qko+TaAMT1;1b) (Yuan et al., 2007a) were surface sterilized and kept in a 4°C chamber for 2 days. The seeds were planted in modified 0.5× MS medium containing 1 mM KNO₃ as the sole N source. Arabidopsis were cultured in the chamber at 22°C with 12 h/12 h: light/dark cycle. Three-day-old plants were transferred to the same medium containing 0 or 10 mM methyl-ammonium (MeA) and grown for another 4 days. In order to analyze the cellular NH4+ contents, the Arabidopsis mutants (qko, qko+AtAMT1;1, and qko+TaAMT1;1b) were planted in 0.5× MS medium and grown for 7 days, after which their complete root systems from 30 plants were collected.

The stem rust-susceptible wheat (Triticum aestivum) line “Little Club” (LC) was used in the experiments to examine the effects of NH4+ on TaAMT1 gene expression. Germinated seeds were grown in deionized water in a greenhouse for 2 weeks to consume all the nutrient solution in the endosperm. Wheat plants were cultured in the chamber at 21°C with 12/12: light/dark cycle. The seedlings were then grown for another 3 days in N-free nutrient solution (–N basal salt: 7 μM Na₂HPO₄, 16 μM KCl, 7 μM CaCl₂.2H₂O, 15 μM MgCl₂.6H₂O, 36 μM FeSO₄.7H₂O, 9 μM MnSO₄.4H₂O, 45 μM H₃BO₄, 3 μM ZnSO₄.7H₂O, 0.2 μM CuSO₄.7H₂O, 0.05 μM Na₂MoO₄.2H₂O) (Abiko et al., 2005), after which they were transferred to a nutrient solution containing 0.5 mM (NH₄)₂SO₄ at pH 5.5. Whole roots and leaves were harvested at 0, 1, 3, and 6 h following the provision of 0.5 mM (NH₄)₂SO₄. Two-week-old LC plants grown in water were transferred to a nutrient solution (–N basal salt) containing 0.5 mM (NH₄)₂SO₄. After 3 days of growth, the plants were transferred to the same nutrient solution without the N source (–N basal salt). Whole roots and leaves were collected after 0, 1, 2, and 3 days of N deprivation. Seventeen-day-old LC plant roots and leaves, as well as 2-month-old plant stems and flowers, were harvested for RNA extraction.

For inoculation of the urediniospores of Pgt (race 21C3CTHTM), the urediniospores were separately inoculated to stem rust-susceptible wheat (T. aestivum) “LC” and stem rust resistant line “Mini 2761” seedlings once the primary leaves had fully expanded while the secondary leaves were being sprouted. The urediniospores of Pgt (race 21C3CTHTM) were propagated by inoculation of LC leaves in the growth chamber. The inoculated seedlings were kept in moist conditions for 14–16 h, and then cultured at 21 ± 1°C, 12 h/12 h: light/dark cycle and a light intensity of 5.8–6.0 klx. The infection types were recorded according to six classes of standards (Supplementary Figure S1; Roelfs and Martens, 1988). To analyze N fertilization dependent stem rust disease index, seeds were grown in deionized water for 2 weeks before being transferred to –N basal salt containing 0.5 mM (NH₄)₂SO₄ solution (pH 5.5) for another 2 weeks prior to inoculation of Pgt.

To search AMT amino sequences in wheat, the rice, Arabidopsis, and potato AMT sequences were used as a bait and searched in the Uniprot database¹ using BLAST with e-value cutoff of e-10 (Pearson, 2013). The retrieved sequences are listed in Supplementary Table S1. The multiple sequence alignments were performed by ClustalW and the evolutionary history was inferred using maximum likelihood based on the JTT matrix-based model. The tree with the highest log likelihood (-12,240.60) is built. Initial tree(s) for the heuristic search were obtained automatically by applying the Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model. The topology with the superior log likelihood value was then selected. The analysis involved 42 amino acid sequences. All positions containing gaps and missing data were eliminated. There was a total of 49 positions in the final dataset. Evolutionary analyses were conducted in MEGA7 (Kumar et al., 2016).

Total cellular RNA was isolated from 20 plant tissues with TRIzol (Takara, Dalian, Liaoning, China), and 2 μg of total RNA was subsequently treated with RQ1 RNase free DNase (Promega, Madison, WI, United States) to eliminate genomic DNA contamination. For cDNA synthesis, a GoScript Reverse Transcription Kit was used following the manufacturer’s instructions (Promega, Madison, WI, United States). Subsequently, qRT-PCR was performed in triplicate using the SYBR Green Mix (Bio-Rad). The three replicates of PCR in each time were analyzed for one sample and the experiments were repeated at least three times. The PCR products were quantified using the Illumina Research Quantity software Illumina Eco 3.0 (Illumina, San Diego, CA, United States). The values of each sample were first normalized against TaEF1α levels from the same samples, and next compared with indicated control group value to analyze the ratio for each gene. Changes in gene expression were calculated using the 2^-ΔΔC_T method (Han et al., 2006). The primers used for qRT-PCR are listed in Supplementary Table S2.

The ammonium uptake deficient yeast strain 31019b (Δmep1, Δmep2, Δmep3, ura3; Marini et al., 1997) was obtained from the Frommer Laboratory (Carnegie Institution for Science). A pDRf1-GW (Xuan et al., 2013) vector harboring TaAMT1;1a, TaAMT1;1b, or TaAMT1;3a was transformed into the yeast cells. The successful transformants were screened by growing of yeast cells in the SD/-ura solid medium. Each transformant was plated on yeast N base media containing 1 mM NH₄Cl or 1 mM arginine, and yeast growth was monitored at 28°C for 3 days. The primers used for cloning the TaAMT genes are listed in Supplementary Table S1.

Nucleotides of TaAMT1;1b ORF were fused to 1.5 kb of the AtAMT1;1 promoter sequence and cloned into the pABind vector (Bleckmann et al., 2010). The pABind-pAtAMT1;1–TaAMT1;1b construct was transformed into the Arabidopsis qko mutant background via Agrobacterium-mediated transformation. qko mutant has mixed Col-0 and Ws-2 genomes; therefore, the qko+AtAMT1;1 was used as an control. The transgenic plant seeds were selected in the plates containing hygromycin, and around 20 individual plants were selected. For analyzing MeA uptake and cellular NH4+ contents, three independent lines were further examined.

Nucleotides of TaAMT1;1 ORFs were cloned into a pABindGFP (35S promoter) destination plasmid (Bleckmann et al., 2010) followed by transient expression in Nicotiana benthamiana leaves using the Agrobacterium-mediated transient expression method (Kim et al., 2009). Green fluorescent protein fluorescence was detected under a confocal microscope (SP5; Leica).

Enzymatic determination of the ammonium contents in roots was performed using an F-Kit (Roche) according to the manufacturer’s instructions (Oliveira et al., 2002).

Statistical calculations were conducted using Prism 5 (GraphPad, San Diego, CA, United States). Significant differences between two groups were analyzed by Student’s t-test. Comparisons between more than two groups were performed by using one-way ANOVA followed by Bonferroni’s multiple comparison test. P-values of <0.05 were considered statistically significant.

High affinity NH4+ transport is mediated by AMT family transporters. To isolate wheat TaAMTs sequences, 10 rice, 6 Arabidopsis, and 3 potato AMT sequences were used as baits to obtain the amino sequences of AMTs in wheat. A total of 23 AMTs (among 25 identified in Uniprot database, 2 AMTs were duplicated) were identified in T. aestivum, and were designated as TaAMTs on the basis of their similarity to OsAMTs, AtAMTs, and LeAMTs (Figure 1 and Supplementary Table S1). In order to understand the evolutionary relationships between TaAMTs and AMTs from other plant species, the amino acid sequences of 10 AMTs from O. sativa, 6 AMTs from A. thaliana, and 3 AMTs from Solanum lycopersicum were collected and aligned. An unrooted phylogenetic ML tree was constructed and the results indicated that the AMTs from four species could be classified into four sub-groups (AMT1, AMT2, AMT3, and AMT4). Among these, all the AMT1 proteins clustered together with high bootstrap support (100%), while AMT2, AMT3, and AMT4 were closely associated in the phylogenetic tree. Furthermore, the tree topology was independent of the methods used for the phylogenetic reconstruction (data not shown). Rice and wheat were clustered more closely, whereas the tomato and Arabidopsis AMTs were clustered relatively distantly (Figure 1).

To examine the expression patterns of TaAMT genes, the roots, leaves, stems, and flowers were collected for RNA extraction. qRT-PCR was performed to analyze 12 TaAMT gene expressions among the 23 TaAMT members, and the results showed that two TaAMT1;1 group genes (1;1a and 1;1b) and TaAMT1;3a were highly expressed, while the remainder of the TaAMT genes were barely detected (Figure 2). The TaAMT1;1 group genes (1;1a and 1;1b) and TaAMT1;3a exhibited similar patterns and showed the highest expression levels in the roots, while similar expression levels were found in the leaves, stems, and flower tissues. The expression levels of TaAMT1;2a and TaAMT1;2c were relatively higher in the leaves and roots than in the stems and flowers. The expression levels of the TaAMT3;2 group genes were not as high as the TaAMT1;1 group genes, and TaAMT3;3a exhibited the highest expression levels in the roots, and the lowest in the stems and flowers. The remaining genes either showed reduced expression levels or were not detected (Figure 2).

Since TaAMT1;1 group genes (1;1a and 1;1b) and TaAMT1;3a were the dominant TaAMT members, the subcellular localization of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a was monitored. GFP coding sequences were C-terminally fused to TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a and the fusion proteins were transiently expressed in N. benthamiana leaves via Agrobacterium-mediated transformation. Two days after infection, the TaAMT-GFP signal was detected on the plasma membrane (Figure 3A). TaAMTs conserved with other characterized AMTs from different plant species in a phylogenetic tree (Figure 1). Therefore, we examined their NH4+ transport activity. Since the TaAMT1;1 group of genes and TaAMT1;3a were most strongly expressed in all the tissues tested, the transport activity of these three proteins was analyzed (Figure 2). We used the yeast mutant complement approach to test the transport activity of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a when NH4+ was supplied. Coding sequences of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a were cloned into the yeast expression vector pDRf1-GW, and the three genes were expressed in the yeast strain 31019b (Δmep1, Δmep2, Δmep3, ura3), which is deficient in NH4+ uptake (Marini et al., 1997). Yeast cell growth was monitored in the media containing 1 mM NH₄Cl or 1 mM arginine. The results indicate that TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a are able to transport NH4+ into yeast cells (Figure 3B).

Two TaAMT1;1 genes (TaAMT1;1a and TaAMT1;1b) are able to complement the yeast mutant 31019b in which NH4+ uptake is deficient, and TaAMT1;1b showed higher affinity than TaATM1;1a in transport of NH4+ (Figure 3B). Therefore, TaAMT1;1b function was further analyzed in plants. The Arabidopsis quadruple mutant (qko) missing AMT1;1, AMT1;2, AMT1;3, and AMT2;1 greatly reduced NH4+ uptake ability (Yuan et al., 2007a). Since gene transformation in wheat is challenging, TaAMT1;1b was selected and expressed in qko under the control of a 1.5 kb AtAMT1;1 promoter. The RT-PCR results indicate that TaAMT1;1b was expressed in the three independent transgenic Arabidopsis lines (#1, #2, and #4), while no visible transcript of TaAMT1;1b was detected in the qko mutant (Figure 4A). MeA uptake was tested in qko, qko+AtAMT1;1 (in which AtAMT1;1 was expressed by its own promoter), and three independent qko+TaAMT1;1b plants. Three-day-old seedlings grown on modified 0.5× MS medium containing 1 mM KNO₃ as the sole N source were transferred to the same medium with or without 10 mM MeA. qko, qko+AtAMT1;1, and qko+TaAMT1;1b exhibited similar growth patterns without the addition of MeA. However, the growth of qko+AtAMT1;1 and qko+TaAMT1;1b, but not gko, was severely affected after the addition of MeA to the growth medium (Figure 4B). Root growth and seedling fresh weight of qko+TaAMT1;1b plants was more severely affected by MeA than qko+AtAMT1;1 plants (Figures 4B,C). Additionally, cellular NH4+ contents were measured from the 7-day-old qko, qko+AtAMT1;1, and qko+TaAMT1;1b plant roots, and the results indicated that the roots of qko+AtAMT1;1 and qko+TaAMT1;1b contained more NH4+ than the roots of qko (Figure 4D).

The expression of AMT genes is sensitive to exogenous N conditions. In Arabidopsis, the AMT1 genes were repressed upon the resupply of NH4+, while the AMT1;1 and AMT;12 genes in rice were highly induced by NH4+ application (Sonoda et al., 2003; Loqué et al., 2006). To examine the NH4+-mediated expression of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a, 17-day-old wheat seedlings grown under N-free conditions were treated with 1 mM NH4+ for 0, 1, 3, and 6 h, and their whole roots and leaves were sampled (Figure 5A). Prior to analyzing the NH4+-dependent expression patterns of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a, the cellular NH4+ contents were measured in the roots. After transferring the wheat plants from the N-free medium to the solution containing NH4+, the NH4+ contents were rapidly increased up to 6 h of treatment (Figure 5B). qRT-PCR results showed that TaAMT1;3a was suppressed while TaAMT1;1a and TaAMT1;1b were not suppressed by NH4+ treatment in the roots (Figure 5C). The expression levels of the remaining TaAMT members were not altered by NH4+ (data not shown). Furthermore, the NH4+ contents in the leaves were monitored before and after NH4+ treatment. The NH4+ levels increased and reached a maximum at 6 h of treatment, but the contents were much lower than in the roots (Figure 5D). However, the NH4+-mediated expression levels of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a were unaltered in the leaves (Figure 5E).

AMT expression is not only sensitive to NH4+ conditions, but also responds to N deprivation in Arabidopsis and rice (Sonoda et al., 2003; Loqué et al., 2006). Therefore, expression of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a was analyzed upon N starvation. The 14-day-old seedlings grown in water were grown in nutrient solution containing NH4+ for another 3 days. They were subsequently transferred to an N deprived solution (Figure 6A). After the transfer, the cellular NH4+ levels rapidly decreased in the roots, and the lowest level was observed after 3 days (Figure 6B). Simultaneously, qRT-PCR was used to determine the N-status-dependent expressions of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a. The results indicated that TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a tested were induced under N starved conditions. The TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a were highly induced, especially on the third day in the roots (Figure 6C), and leaves (Figure 6D).

AtAMT1;1 has been shown to alter basal defense against P. syringae and P. cucumerina (Pastor et al., 2014). In wheat, three TaAMTs levels were altered by AM fungi infection (Duan et al., 2015). To dissect whether there is a relationship between Pgt infection and expression of TaAMT1;1 genes, the stem rust-susceptible line LC (Roelfs and Martens, 1988) and resistant line Mini 2761 (Luan et al., 2013) were utilized, respectively. Following inoculation, urediniospore multiplication was detected in the leaves of the LC plants, while Mini 2761 exhibited lack of lesions in leaves (Figure 7A). Furthermore, the urediniospore inoculation-induced expression of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a genes was examined. The expression levels of TaAMT1;1a, TaAMT1;1b, and TaAMT1;3a differed insignificantly between the leaves of LC and Mini 2761. TaAMT1;3a was induced at 12 and 24 h (Figure 7B); TaAMT1;1a was induced at 36 and 48 h (Figure 7C); and TaAMT1;1b was induced at 24, 36, and 48 h after inoculation in LC only (Figure 7D). The expression levels of the three genes showed trifle difference following inoculation of the leaves of Mini 2761.

As TaAMT genes were specifically induced in a stem rust-susceptible line (LC) after inoculation of the urediniospores, the effects of N availability in the growth medium on wheat stem rust disease were examined further. The LC plants were grown without any nutrient supply for the first 2 weeks after germination, following which they were transferred to a medium containing either 0 or 1 mM NH4+. After 2 weeks of growth, the urediniospores were inoculated evenly on the surface of the leaves (Figure 8A). The symptoms were monitored after another 2 weeks of inoculation. The results indicated that the plants grown under N-free conditions exhibited lighter colored uredinia as well as chlorosis surrounding the uredinium-formed regions (Figure 8B). Disease class standards were calculated for the symptoms shown in Figure 7B following international wheat rust disease standards (Abiko et al., 2005). The results indicate that plants grown on the medium containing NH4+ were on average Class 4, while the plants grown under N-free conditions were on average Class 3 (Figure 8C). These results suggest that N deficiency inhibits stem rust disease in wheat.