The NIN-like protein 5 (ZmNLP5) transcription factor is involved in modulating the nitrogen response in maize.

Maize exhibits marked growth and yield response to supplemental nitrogen (N). Here, we report the functional characterization of a maize NIN-like protein ZmNLP5 as a central hub in a molecular network associated with N metabolism. Predominantly expressed and accumulated in roots and vascular tissues, ZmNLP5 was shown to rapidly respond to nitrate treatment. Under limited N supply, compared with that of wild-type (WT) seedlings, the zmnlp5 mutant seedlings accumulated less nitrate and nitrite in the root tissues and ammonium in the shoot tissues. The zmnlp5 mutant plants accumulated less nitrogen than the WT plants in the ear leaves and seed kernels. Furthermore, the mutants carrying the transgenic ZmNLP5 cDNA fragment significantly increased the nitrate content in the root tissues compared with that of the zmnlp5 mutants. In the zmnlp5 mutant plants, loss of the ZmNLP5 function led to changes in expression for a significant number of genes involved in N signalling and metabolism. We further show that ZmNLP5 directly regulates the expression of nitrite reductase 1.1 (ZmNIR1.1) by binding to the nitrate-responsive cis-element at the 5' UTR of the gene. Interestingly, a natural loss-of-function allele of ZmNLP5 in Mo17 conferred less N accumulation in the ear leaves and seed kernels resembling that of the zmnlp5 mutant plants. Our findings show that ZmNLP5 is involved in mediating the plant response to N in maize.

Introduction

Maize (Zea mays L.) is known to exhibit one of the highest yield responses to supplemental nitrogen (N), leading to a significant amount of N fertilizers being applied to its production (Glass, 2003; Bi et al., 2014). Intensive use of N fertilizers not only increases crop input costs but also negatively impacts the environment (Gutierrez, 2012). A better understanding of how maize senses N and fine tunes its physiological and developmental processes to fluctuating N concentrations in the soil is needed to improve N use efficiency (NUE) in maize production.

Nitrate is the primary N source for land plants. It also serves as an essential physiological signal in initiating and regulating the N response of plants (Crawford, 1995; Crawford and Forde, 2002). Plants absorb nitrate through specific nitrate transporters, such as NRT1.1 (Ho et al., 2009). The variation in the nitrate transporter gene (OsNRT1.1B) contributes to NUE divergence between rice subspecies (Hu et al., 2015). Once taken up by the plant, nitrate is reduced to ammonium by two key enzymes, namely, nitrate reductase (NR) and nitrite reductase (NIR) (Crawford and Forde, 2002; Gojon et al., 2009). Analysis of the NIR promoters from Arabidopsis and several higher plants revealed a conserved nitrate‐responsive cis‐element that has been shown to be essential for the nitrate‐dependent activation of the promoter to direct N‐responsive transcription (Konishi and Yanagisawa, 2010).

Plants have evolved a complex of physiological, morphological, and developmental mechanisms to respond and adapt to nitrate fluctuation (Chardin et al., 2014). Several studies have been performed to uncover the mechanism underlying N response. High nitrate‐triggered lateral root initiation in the shoot‐borne roots of maize by modulating auxin‐related cell cycle regulation (Yu et al., 2015). A merged transcriptomic and proteomic survey in the maize root apex transition zone revealed nitrate sensing linked to the biosynthesis and signalling of several phytohormones, and suggested that cytoskeleton activation and cell wall modification occurred in response to nitrate (Trevisan et al., 2015).

Several important N‐responsive transcription factors (TFs) that modulated the expression of genes involved in N uptake and assimilation have been identified, such as CCA1 (a Myb‐related TF) and LBD37/38/39 found in Arabidopsis (Gutierrez et al., 2008; Rubin et al., 2009). NIN‐like protein (NLP), is a plant‐specific TF family carrying two major conserved domains. The RWP‐RK domain for DNA binding consists of c. 60 amino acid residues containing an RWPXRK motif (Schauser et al., 2005; Chardin et al., 2014). The PB1 domain consists of c. 80 amino acid residues and is involved in protein–protein interactions associated with nitrate‐inducible gene expression in higher plants (Sumimoto et al., 2007; Chardin et al., 2014; Konishi and Yanagisawa, 2019). NLPs have been suggested to be involved in mediating the early N response. For example, AtNLP7 has been shown to bind to the nitrate‐responsive cis‐element in the promoter region of NIR1, resulting in increased expression of NIR1 (Konishi and Yanagisawa, 2013). In addition, AtNLP7 binds to many genes involved in nitrate signalling and assimilation, including ANR1 (Zhang and Forde, 1998), CIPK8 (Hu et al., 2009), NRT1.1 (Ho et al., 2009), and NR1 (Konishi and Yanagisawa, 2011). Additionally, NLP proteins can work as heterodimers by interacting with each other. For example, under deficient nitrogen (DN) conditions, the AtNLP6/AtNLP7 heterodimer interacts with the transcription factor TCP20 and coordinates plant responses to nitrate availability (Guan et al., 2014; Guan et al., 2017). Furthermore, the nitrate‐CPK (Ca2+‐sensor protein kinase)‐NLP regulatory network has been found to be an important component in the nutrient‐growth network in Arabidopsis (Liu et al., 2017).

Previous studies have detected genome‐wide transcriptional changes occurring in response to nitrate in maize and demonstrated that c. 7% of the maize transcriptome is nitrogen responsive (Liu et al., 2008; Trevisan et al., 2011; Yang et al., 2011; Zhao et al., 2013; Zamboni et al., 2014). It has been reported that ZmNRT2.1, ZmNRT2.2, and ZmGln1‐3 play an important role in N use (Martin et al., 2006; Garnett et al., 2013). Moreover, a study showed that treating the maize plants with nitrate and urea simultaneously enhances the expression of genes associated with nitrate transport and assimilation, compared with that exposed to nitrate alone (Zanin et al., 2015). However, the signals involved in nitrate signalling and assimilation in maize are largely unknown. Recently, studies of genome‐wide identification of NLPs in maize have been reported (Ge et al., 2018; Wang et al., 2018a), where nine maize NLPs were identified (ZmNLP1‐ZmNLP9) and their gene structures were characterized. It was shown that the expression of two NLPs, ZmNLP4 and ZmNLP5, was most significantly upregulated upon application of nitrate (Ge et al., 2018). In this study, we found that ZmNLP5 is one of the central hub genes in the molecular network for mediating N signalling and metabolism. Thus, we investigated the functional roles of maize NLPs by characterizing a key family member, ZmNLP5, in depth.

Results

is a central hub in a molecular network for mediating N signalling and metabolism

A genome‐wide survey identified nine maize NLP genes based on a homology‐based analysis of the conserved RWP‐RK and PB1 domains (Ge et al., 2018). To investigate how these NLPs interact with other genes and pathways in maize, we conducted a gene regulatory network (GRN) analysis using a set of maize pan‐transcriptome RNA‐seq data on 503 diverse maize inbred lines (Hirsch et al., 2014). In total, 30 127 transcripts and 6 174 719 gene–gene links were identified. The regulatory network associated with maize NLPs was extracted. The resulting maize NLP network included 110 genes and 489 links, where genes are denoted as nodes connected by links/edges representing potential regulatory interactions (Figure 1a and Table S1).

Figure 1

Molecular network associated with ZmNLPs. (a) Genome‐wide co‐expression network associated with ZmNLPs. Network was constructed from pan‐transcriptome expression data of 503 maize inbred lines using DeGNServer. Nodes represent genes. Edges connecting the nodes represent co‐expression patterns indicative of regulatory relations. ZmNLPs were used as seed genes for network construction, and are highlighted here in red. Known N‐responsive transcription factors are highlighted in yellow. Gene involved in N metabolism are highlighted in green. Other genes are shown in grey, wherein the prefix ‘GRMZM’ in a gene ID is omitted for visual clarity (e.g., 2G141636 has a gene ID GRMZM2G141636). (b) Subnetwork associated with ZmNLP5. Subnetwork was extracted using community‐find algorithm GeNa with the nine ZmNLPs as query genes. ZmNLP5 is highlighted in red, whereas the two other ZmNLPs identified in this subnetwork, ZmNLP3 and ZmNLP7, are highlighted in purple. Known N‐responsive transcription factors are highlighted in yellow. Genes involved in N metabolism are highlighted in green.

The results revealed a high degree of interaction between the maize NLPs and genes known to be involved in N sensing, signalling and metabolism, including NRT, AAT, and ASN1 (Ho et al., 2009; Zamboni et al., 2014), as well as N‐regulatory TFs such as MYB‐related (ZmMYBR47, 115), MYB (ZmMYB149), GLK (ZmGLK10, 27) (Gutierrez et al., 2008), according to GRASSIUS (http://grassius.org/). Gene set enrichment analysis on the network genes showed that the maize NLP network is significantly enriched in genes involved in N metabolism and amino acid biosynthesis pathways (Table S2, P ≤ 0.05). Among the maize NLP family members, ZmNLP3, ZmNLP5, and ZmNLP7 have the highest degrees of connectivity (>30) in the network (Table S1).

Compared with that of ZmNLP3 and ZmNLP7, the expression of ZmNLP5 showed a higher variation in response to nitrate (upregulated over two‐fold, Figure S1, Ge et al., 2018). Furthermore, ZmNLP5 has a closer phylogenetic relationship with AtNLP7 than ZmNLP3 and ZmNLP7 (Castaings et al., 2009; Ge et al., 2018; Figure S2). Therefore, we selected ZmNLP5 for further analysis. In the subnetwork associated with ZmNLP5, the ZmNLP5 was found to be significantly linked with 39 genes, many of which are known to be important in N signalling or metabolism in maize plants, including NRTs and AATs. ZmNLP5 was also found to be connected with ZmNLP3 and ZmNLP7 (Figure 1b and Table S3). These results suggest that ZmNLP5 is potentially a central hub in a molecular network for mediating N signalling and metabolism in maize.

is predominantly expressed in root and vascular tissues

To examine the tissue‐specific expression pattern of the gene, we investigated ZmNLP5 expression in maize root, stem, and leaf tissues by semiquantitative reverse transcription polymerase chain reaction (PCR). ZmNLP5 transcripts were detected in all three tissues, but significantly more in roots than in stems or leaves (Figure 2a). Next, a ZmNLP5‐specific antibody (Figure S3) was generated to investigate the protein level of ZmNLP5. The results of the immunoblot assay showed that ZmNLP5 was mainly detected in roots at the protein level (Figure 2b). These results suggest that ZmNLP5 is preferentially expressed in roots.

Figure 2

ZmNLP5 is predominantly expressed in root and vascular tissues. (a, b) Expression of ZmNLP5 in maize root, stem, and leaf tissues. (a) mRNA levels were measured by reverse transcription PCR with maize housekeeping gene ZmUPF1 (GRMZM2G163444) used as the reference gene. (b) Protein levels were measured by immunoblot assay, with anti‐UDPGP used as a sample loading control. (c–h) RNA in situ hybridization of ZmNLP5 in root and leaf. (c) Transverse section of a root tip with sense probe as a negative control; (d–g) root tips; (h) leaf. Hybridization signals detected by labelled antisense probes are shown by arrows in root stele (d); lateral root primordia (e–g); leaf vascular bundle (h). Bar = 100 μm.

An RNA in situ hybridization assay was conducted to detect ZmNLP5 transcripts in maize root and leaf tissue sections (Figure 2c–h). In roots, ZmNLP5 transcripts were shown to be primarily localized in the epidermis and stele of the root tip (Figure 2c,d), as well as in the lateral root primordia (Figure 2e–g). In leaves, ZmNLP5 mRNA was mainly detected in vascular bundles (Figure 2h). These findings suggest the involvement of ZmNLP5 in nutrient uptake and transport pathways.

expression is responsive to N supply

To examine whether ZmNLP5 is responsive to nitrate, 1‐week‐old seedlings were subjected to N starvation for 2 weeks and then supplied with nitrate. Samples of mRNA and protein were collected from seedling roots at a series of time points after induction of nitrate and were used for quantitative PCR (qPCR) and immunoblot assays, respectively. The transcription level of ZmNLP5 was significantly upregulated shortly after the supply of nitrate on the nitrate‐deprived plants and peaked at 90 min after treatment (Figure 3a). The protein level of ZmNLP5 increased 30 min after nitrate supply, followed by a gradual decrease from 60 to 90 min after induction (Figure 3b). These results demonstrated that the expression of ZmNLP5 responds rapidly and strongly to nitrate, suggesting that ZmNLP5 may be involved in orchestrating early stages of the N response.

Figure 3

ZmNLP5 expression is responsive to nitrate. (a, b) Expression response of ZmNLP5 to nitrate, samples were collected from roots of N‐deprived maize seedlings at 0, 15, 30, 60, 90, 120, and 240 min after resupply of nitrate. (a) mRNA levels of ZmNLP5 were measured by quantitative PCR (qPCR), significant differences were indicated by letters (anova; P ≤ 0.05); (b) Protein levels of ZmNLP5 after nitrate treatment were analyzed by immunoblot assay. (c, d) Expression of ZmNLP5 in maize seedling root tissues under sufficient N (SN) and deficient N (DN) conditions, samples were collected from roots of maize seedlings grown in hydroponic culture with SN or DN supply for 21 days, respectively. (c) mRNA levels of ZmNLP5 were measured by qPCR, asterisk represents a P ≤ 0.05 statistical significance using Student’s t‐test [n = 3]. (d) Soluble (S) and nuclear (N) proteins levels of ZmNLP5 were measured by immunoblot assay, UDPGP protein was used as a cytosolic marker and histone H3 was used as a nuclear marker. The histogram corresponds to the ratios of ZmNLP5 (S or N) to control (UDPGP or H3) content shown in the right, and the protein contents were evaluated by ImageJ. One‐way anova analysis was used to test treatment effects. Significant differences were indicated by letters (anova; P ≤ 0.05). (e) Subcellular localization pattern of ZmNLP5 in response to N fluctuation in maize root. Soluble (S) and nuclear (N) proteins were extracted from samples collected from N‐starved seedlings supplied with nitrate for indicated time (+N 0.5 h and +N 1.5 h), followed by N‐deprivation (−N 24 h). Immunoblot assay was used to ascertain ZmNLP5 soluble and nuclear protein levels using antibodies against ZmNLP5. Anti‐UDPGP and anti‐H3 were used as soluble and nuclear protein sample loading controls. The histogram corresponds to the ratios of ZmNLP5 (S or N) to control (UDPGP or H3) content shown in the right, the protein content was evaluated by ImageJ. One‐Way anova analysis was used to test treatment effects. Significant differences were indicated by letters (P ≤ 0.05).

The change in the transcript abundance of ZmNLP5 in response to long‐term N availability was also investigated. The qPCR results showed that the transcript level of ZmNLP5 was significantly higher in seedlings grown under sufficient N (SN) conditions compared with seedlings grown under deficient N (DN) conditions (Figure 3c, P ≤ 0.05). At the protein level, under both N conditions, ZmNLP5 was predominantly localized in the nucleus, with relatively low abundance in the cytoplasm (Figure 3d), a distribution pattern in agreement with ZmNLP5’s role as a transcription factor. Under SN conditions, both nucleus and cytoplasm ZmNLP5 levels were higher than those under DN conditions (Figure 3d, P ≤ 0.05). Nevertheless, it appears that under DN conditions, the fraction of ZmNLP5 in the cytoplasm is much lower than that under SN conditions (6.13% versus 40.67%, P ≤ 0.05). In other words, ZmNLP5 is more concentrated in the nucleus upon N shortage.

We further investigated the subcellular localization pattern of ZmNLP5 in response to N fluctuation (Figure 3e). In the cytoplasm, ZmNLP5 protein levels (normalized against UDP‐glucose pyrophosphorylase (UDPGP) loading control; Figure 3e) rapidly increased following N treatment and diminished following N‐deprivation, while in the nucleus, ZmNLP5 protein levels (normalized against H3 loading control; Figure 3e) increased following N treatment and remained relatively stable even after 24 h of N‐deprivation. The above results suggest the nuclear localization of the newly synthesized ZmNLP5 in accordance with its role as a transcription factor.

Together, these results suggest that ZmNLP5 responds to N influx with increased expression and maintains a higher level of expression in response to a higher N availability in the long term. Moreover, there is more ZmNLP5 in the nucleus than in the cytoplasm during N deficiency.

mutant plants exhibit impaired N assimilation under N deficiency

To better characterize the function of ZmNLP5, we obtained a mutant for further analysis (Maize Genetics Cooperation Stock Center, line UFMu‐01175). A Mu insertion was found in the fourth exon of the ZmNLP5 gene (Figure 4a). Quantitative PCR and immunoblot results revealed that the expression of ZmNLP5 was significantly reduced in zmnlp5 mutant seedlings (P ≤ 0.05, Figure 4b, c).

Figure 4

ZmNLP5 mutation impairs the maize N response under N deficiency. (a) Gene structure of ZmNLP5. Boxes present exons. Location of the Mu element in zmnlp5 mutant is shown at triangle. Homozygous mutant plants were identified by the specific primers: P1 + P2 and P3 + P2. (b) ZmNLP5 transcript levels in wild‐type (WT) and zmnlp5 mutant by qPCR. Double asterisks represent a P ≤ 0.01 statistical significance using Student’s t‐test [n = 3]. (c) Protein levels in in WT and zmnlp5 mutant quantified by immunoblot analysis. Anti‐UDPGP was used as a sample loading control. (d) Nitrate, nitrite, and ammonium contents in root and shoot tissues of WT and zmnlp5 mutant plants grown for 21 days in hydroponic culture on SN (15 mm KNO3) solution and DN (0.15 mm KNO3) solution. Asterisk represents a P ≤ 0.05 statistical significance using Student’s t‐test [n = 3]. (e, f) Phenotype detection of WT and zmnlp5 plants in mature period on sufficient nitrogen (SN) and deficient nitrogen (DN) conditions. (e) Phenotype of ear leaves in mature plants of WT and zmnlp5; (f) The chlorophyll (designated as SPAD value) contents in ear leaves; The total N contents in ear leaves; The total N contents in seed. Asterisk represents a P ≤ 0.05 statistical significance using Student’s t‐test [n = 3]. (g) Functional complementation tests of zmnlp5. The construct pHB‐ZmNLP5 was transformed into immature embryos from a maize inbred line (B104). Transgenic lines (T1) were crossed to the homozygous mutant (zmnlp5) and then self‐pollinated (F2). ZmNLP5/zmnlp5 (1–8) and zmnlp5/zmnlp5 (9–16) seedlings were identified from an F2 population by P1 + P2, P3 + P2, and primers for the bar gene, B104, WT, and zmnlp5 were used as controls. mRNA levels of ZmNLP5 in seedling roots were measured by qPCR. Nitrate contents in root tissues of plants grown for 21 days in hydroponic culture on DN (0.15 mm KNO3) solution. *P ≤ 0.05, **P ≤ 0.01 using Student’s t‐test [n = 3].

We investigated the contents of nitrate, nitrite, and ammonium in 3‐week‐old seedlings of WT and zmnlp5 mutants under both SN and DN conditions. Under SN conditions, WT and zmnlp5 exhibited no significant difference in the nitrate, nitrite, or ammonium contents for both root and shoot tissues. Under DN conditions, zmnlp5 accumulated significantly less nitrate and nitrite in the root tissues, along with more nitrite and less ammonium in the shoot tissues when compared with those of WT (Figure 4d, P ≤ 0.05). These results of altered N compositions in the zmnlp5 mutant seedlings suggest that ZmNLP5 is involved in modulating N uptake and assimilation under deficient N conditions (Figure 4d).

To test the effect of ZmNLP5 on N assimilation in adult plants, we examined the total N content in ear leaves and seed kernels of WT and zmnlp5 plants at the mature stage (Figure 4e,f). Under SN conditions, WT plants had accumulated higher N content in ear leaf but similar N content in seed kernels compared with that of zmnlp5 plants (Figure 4f), whereas under DN conditions, zmnlp5 exhibited significantly lower N content for both ear leaf and seed kernels compared with that of WT (Figure 4f, DN: 18.388 ± 0.619 mg/g in zmnlp5 versus 23.367 ± 2.702 mg/g in WT (ear leaf); DN: 12.950 ± 0.321 mg/g in zmnlp5 versus 16.411 ± 0.170 mg/g in WT (seed kernels), P ≤ 0.05). Visually, zmnlp5 plants exhibited paler leaf colour compared with WT, especially under DN conditions (Figure 4e), which is supported by the SPAD meter values (Figure 4f). These results of altered N contents in the zmnlp5 mutant plants indicate that ZmNLP5 is involved in modulating N assimilation, especially under deficient N conditions.

To confirm that ZmNLP5 is involved in the modulation of N assimilation in maize under N limitation, a functional complementation test was performed. We transformed the cDNA fragment encoding ZmNLP5 into a maize inbred line (B104). The transgenic lines (ZmNLP5) were crossed to the homozygous mutant plants (zmnlp5) and then self‐pollinated. Based on genotyping analysis, 16 seedlings (Figure 4g, 1–16) containing homozygous zmnlp5 alleles were identified from an F2 population, among which eight plants were transgene positive (ZmNLP5/zmnlp5; Figure 4g) and eight were transgene negative (zmnlp5/zmnlp5; Figure 4g). Then, we investigated the transcriptional levels of ZmNLP5 and nitrate contents in 3‐week‐old seedling roots of 16 plants under DN conditions. The qPCR results showed that the expression of ZmNLP5 was significantly increased in ZmNLP5/zmnlp5 compared with that of zmnlp5, whereas the expression in zmnlp5/zmnlp5 exhibited little difference. Furthermore, in comparison with that of zmnlp5, ZmNLP5/zmnlp5 accumulated higher nitrate contents in root tissues (increased 3.5–9.7‐fold; Figure 4g, 1–8), whereas nitrate contents in zmnlp5/zmnlp5 root tissues showed no significant difference (Figure 4g, 9–16). These results revealed that the transgenic ZmNLP5 cDNA fragment could recover from the defective ability of N assimilation in the zmnlp5 mutant.

Transcriptional landscape of N signalling and metabolism is altered in mutant plants

To investigate the molecular mechanism of the effect of zmnlp5 mutation on maize N response, we conducted transcriptome profiling using WT and zmnlp5 seedlings (V3 stage) under N starvation and 30 min after nitrogen supply (each sample has three biological replicates). The RNA‐seq analysis of 136 223 gene models revealed that roughly one‐third of maize gene models were transcribed in each sample (Figure 5a). These gene models were used for further analysis. To compare the response of genes to the change of N levels between WT and zmnlp5 seedlings, the transcript abundance of genes in each sample was visualized using the ggplot2 package of R software (Figure 5d). The results showed that WT plants had not only a higher number of expressing genes after N treatment (Figure 5a) but also a higher number of differentially expressed genes after N treatment (Figure 5b,d).

Figure 5

Transcriptome profiling of WT and zmnlp5 mutants in response to nitrate. (a) Number of expressed and non‐expressed genes in WT and zmnlp5 mutants before (−N) and after (+N) nitrate treatment by transcriptome profiling of WT and zmnlp5 mutants in response to nitrate. (b) Venn diagram showing numbers of differentially expressed genes (DEGs) in response to nitrate in WT and zmnlp5 mutants. (c) qPCR validation of 16 N‐related genes that showed altered expression in zmnlp5 mutant plants in response to N. 1: ZmNRT1.5a (GRMZM2G044851); 2: ZmNRT2.1 (GRMZM2G010280); 3: ZmNRT2.2 (GRMZM2G010251); 4: ZmNRT3.1a (GRMZM2G179294); 5: ZmPTAL (GRMZM2G441347); 6: ZmIAA16 (GRMZM2G159285); 7: ZmIAA3 (GRMZM2G167794); 8: ZmARF (GRMZM2G066219); 9: ZmAP2 (GRMZM2G174834); 10: ZmANR1 (GRMZM2G147716); 11: ZmDof (GRMZM2G089850); 12: ZmLBD37 (GRMZM2G132693); 13: ZmLBD39 (GRMZM2G386674); 14: ZmMYB (GRMZM2G104551); 15: ZmMYBR (GRMZM2G097788); 16: ZmWRKY (GRMZM2G304573). (d) Volcano plots of log2 fold changes of gene expression in N‐starved WT and zmnlp5 mutant plants after 30 min of nitrate treatment. The red or green points indicate that both large‐magnitude fold changes (x‐axis) as well as high statistical significance (‐lg of P‐value, y‐axis). (e) GO classification of the 779 functional annotated genes that were responsive to nitrate supply in WT but not in zmnlp5. A list of top significant GO terms and number of genes classified within each GO term is shown.

Differentially expressed genes (DEGs) were identified with the criteria of log2 expression ratios being either ≥1 or ≤−1 and q ≤ 0.05 (Storey, 2003). After nitrate treatment, 1507 and 1128 DEGs were identified in WT and zmnlp5 plants, respectively (Figure 5b). There were 170 DEGs shared between WT and zmnlp5, while 89% (1337 out of 1507; Table S4) of DEGs in WT were not differentially expressed upon nitrate supply in the zmnlp5 mutant (Figure 5b). In total, 1337 genes responsive to nitrate supply in WT but not in zmnlp5 were then subjected to a gene ontology (GO) term enrichment analysis, where 779 functionally annotated genes were characterized and significantly enriched for GO terms in N metabolism and related pathways (Figure 5e), including regulation of the nitrogen compound metabolic process (GO: 0051171) and G‐protein coupled receptor protein signalling pathway (GO: 0007186). Moreover, putative homologues modulating N signalling and assimilation pathways in Arabidopsis have been reported as DEGs exclusively found in WT, including ZmANR1; ZmLBD37; ZmLBD39; ZmNRT1.1; ZmNRT2.1; ZmNRT3.1; ZmNR1.2; ZmNIR1.1; and ZmARF (Gutierrez, 2012; Zhang and Forde, 1998; Rubin et al., 2009; Ho et al., 2009; Konishi and Yanagisawa, 2010, 2011; Rinaldi et al., 2012; Zamboni et al., 2014; Zanin et al., 2015; Trevisan et al, 2015; Table S4). Figure 5(c) shows qPCR validation of the log2 ratio of transcript abundance after/before N‐resupply (log2 (+N/−N)) on 16 N‐related genes selected based on RNA‐seq results. We observed similar patterns between results from qPCR and RNA‐seq. These results indicate that the loss of the ZmNLP5 gene function affects the transcription of more than one thousand genes regulated by N supply.

To further elucidate the effect of the zmnlp5 mutation on N assimilation, we examined the expression levels of four genes encoding key enzymes associated with N assimilation pathways in zmnlp5 versus WT (ZmNR1.1: GRMZM5G878558, ZmNR1.2: GRMZM2G428027, ZmNIR1.1: GRMZM2G079381, ZmNIR1.2: GRMZM2G102959, ZmNR1.2, and ZmNIR1.1 were included in the set of DEGs; Table S4). One‐week‐old seedling roots of zmnlp5 and WT plants were subjected to N starvation for 2 weeks and then treated with nitrate. We examined the transcript levels of ZmNR1.1, ZmNR1.2, ZmNIR1.1, and ZmNIR1.2 at a series of time points (0, 30, 60, and 120 min). The results showed that while the expression of these genes was highly induced by nitrate in both WT and mutant plants, this N‐induced upregulation of gene expression was significantly impeded in the zmnlp5 mutant plants (Figure S4), suggesting that ZmNLP5 plays a role in modulating N response and assimilation. Furthermore, significant differential expression was detected between WT and mutant plants after 30 min for ZmNR1.2. In contrast, this difference in expression was not apparent for other genes until after 60 min. These results indicated that these genes were not uniformly influenced by ZmNLP5. In addition, ZmNIR1.1 was the only gene that demonstrated significant differential expression between WT and mutant plants under N deficiency (Figure S4, 0 min), suggesting that ZmNLP5 is critical in the modulation of ZmNIR1.1 upon N starvation.

ZmNLP5 directly regulates by binding to the nitrate‐responsive ‐element

Quantitative PCR analysis revealed that the nitrate‐induced expression of ZmNIR1.1 and ZmNIR1.2, homologues of AtNIR1, were disturbed in zmnlp5 (Figure S4), suggesting that ZmNIR1.1 and ZmNIR1.2 are candidate targets directly regulated by ZmNLP5. To test this hypothesis, we performed chromatin immunoprecipitation (ChIP) assays followed by qPCR to determine whether the promoter regions (P1, P2, and P3) of ZmNIR1.1 and ZmNIR1.2 would be enriched by ChIP with anti‐ZmNLP5 antibody. The ChIP‐qPCR results revealed that the P1 region of the ZmNIR1.1 promoter was significantly enriched in immunoprecipitated DNA compared with that of the negative control (Figure 6a,b). Weak enrichment of the P2 and P3 regions of the ZmNIR1.1 promoter, as well as the P1 region of the ZmNIR1.2 promoter was also detected (Figure 6a,b). Because of the relatively stronger enrichment results, ZmNIR1.1 was selected and used as the exemplary target gene in this study.

Figure 6

ZmNLP5 activates the expression of ZmNIR1.1 through NRE binding. (a) Schematic representation of the structures of the promoters (lines) and 5′UTRs (boxes) of ZmNIR1.1 and ZmNIR1.2. Nitrogen response cis‐elements are shown as black stripes in 5′ UTR. (b) ChIP‐qPCR analysis of ZmNLP5 putative target genes in different regions of promoters. (c) Sequences of wild‐type NREs (AtNRE of AtNIR1 and ZmNRE of ZmNIR1.1) as well as mutated NRE (mZmNRE). (d, e) EMSA determination of complex formation between ZmNLP5 and NREs. (d) Wild‐type NREs (AtNRE and ZmNRE) shown in (c) as probes to bind to the recombinant 6× His‐tagged RWP‐RK domain of ZmNLP5 (6× His‐RWP‐RK); (e) Non‐labelled wild‐type and mutated NREs were used at an excess molar ratio (50×, 100×) as competitor DNA. (f, g) ZmNLP5 activates transcription of ZmNIR1.1. Constructs used in the transcription activation assay. (f) The 35S::REN‐ProZmNIR1.1::LUC reporter construct was transiently expressed in onion epidermal cells together with the negative control vector and the 35S::ZmNLP5 effector vector. The reporter construct contains two reporter genes, the firefly luciferase (LUC) gene and the Renilla luciferase (REN) gene. Expression level of REN was used as an internal control. (g) Expression of reporter genes. The LUC/REN ratio represents the relative activity of ZmNIR1.1 promoters. The grey bar shows the LUC/REN ratio from the assay with only the reporter construct expressed. The black bar shows the LUC/REN ratio from the assay with both effector construct and reporter construct expressed. Three biological replicates were used. Double asterisks represent statistical significance of P ≤ 0.01 using Student’s t‐test [n = 3].

To further characterize the binding specificity of ZmNLP5 to its target, we first tested whether the RWP‐RK domain can bind to the same nitrogen response cis‐elements (NRE) as AtNLP7 (Konishi and Yanagisawa, 2010; Konishi and Yanagisawa, 2013). An electrophoretic mobility shift assay (EMSA) showed that the RWP‐RK domain of ZmNLP5 is able to bind to AtNRE in vitro (Figure 6c,d), suggesting that ZmNLP5 could potentially target similar sequences of AtNRE in vivo. Although no sequence was found in the P1 region of ZmNIR1.1 that exactly matched the classical NRE pattern with 10‐bp spacer, a similar pseudopalindromic NRE sequence separated by a 9‐bp spacer ‘CTTN9AAG’ (e.g. CTTGGGGAGTTCAAG, designated as ZmNRE, Figure 6c) was found. Subsequently, we investigated the binding specificity of ZmNLP5 using the sequence containing ZmNRE in the P1 region of ZmNIR1.1 (Figure 6a,c). The results showed that the RWP‐RK domain fused to a 6 × His‐tag successfully binds to ZmNRE (Figure 6d). A sequence containing mutated distal half sites of ZmNRE was generated (designated as mZmNRE) and used to test the binding specificity of ZmNLP5 (Figure 6c). The results showed that mZmNRE was unable to compete for biotin‐labelled wild‐type (WT) NREs (Figure 6e). These results indicate that the RWP‐RK domain of ZmNLP5 is capable of directly and specifically binding to NREs.

To further assess the function of ZmNLP5 in transcriptional activation in vivo, we conducted a dual‐luciferase transient transcriptional activity assay. The construct 35S::ZmNLP5, with ZmNLP5 ORF from B73 was generated as an effector. The reporter construct contains two luciferase cassettes, the Renilla luciferase gene (REN) driven by the cauliflower mosaic virus (CaMV) with the 35S promoter (35S::REN) used as an internal control, and the firefly luciferase gene (LUC) driven by the ZmNIR1.1 promoter that was used as a reporter (Figure 6f). Co‐expression of ProZmNIR1.1::LUC with 35S::ZmNLP5 resulted in a 2.2‐fold (t‐test, P ≤ 0.01) increase in LUC activity compared with the control (Figure 6g), implying that ZmNLP5 can activate the transcription of ZmNIR1.1.

It is worth mentioning that the ZmNLP5 allele in the ZmNIR1.1 binding study is from the maize reference inbred line B73, whose genome has been sequenced, assembled and annotated (Jiao et al., 2017). Compared with that of B73, the cDNA sequence of ZmNLP5 from W22 contains 24 SNPs and a 3‐bp insertion (Table S5). No amino acid differences were observed in the RWP‐RK domain between the B73‐ and W22‐derived ZmNLP5 cDNA sequences, which is consistent with the observation that both lines are sensitive to nitrate supply (Ge et al., 2016).

Natural loss‐of‐function allele of ZmNLP5 is identified in the maize inbred line Mo17

Phenotypic variation has been observed in maize germplasm regarding N response and sensitivity. For instance, it has been previously reported that Mo17 is much less sensitive to fluctuations in external N levels than B73 (Ge et al., 2016). We compared the ZmNLP5 cDNA sequences between Mo17 and B73 to identify potential sequence polymorphisms. The results showed that the Mo17 ZmNLP5 cDNA sequence contains major mutations when compared with the B73 reference sequence (Figure 7a and Table S5). There are 20 SNPs and five insertion/deletions between the cDNA sequences. Particularly, there is a complete intron retention (139‐bp) at +661 bp in Mo17, accounting for the largest insertion compared with the cDNA sequences in B73 (Figure S5 and Table S5). Alignment between genomic DNA and cDNA revealed that the insertion is caused by retention of the third intron in Mo17, which leads to a frame shift and a premature stop codon 104 bp after the insertion. Thus, the resulting ZmNLP5 protein product in Mo17 is truncated with two‐thirds of the coding region absent, including the conserved RWP‐RK domain that confers the DNA binding function. In addition, there is a 211‐bp deletion at +1796 bp, overlapping with the PB1 domain near the C terminal of the protein. Thus, Mo17 seems to have a natural loss‐of‐function allele of ZmNLP5.

Figure 7

ZmNLP5 natural mutation impairs N assimilation under N deficiency. (a) The genomic structure of the ZmNLP5 gene in B73 and Mo17, the filled box presents the exon, and In = Insertion, De = Deletion. (b) Schematic representation of constructs using in transcriptional activity assay. (c) Transcription activity of ZmNLP5 (from Mo17 or B73) were tested in tobacco leaves using a GAL4/UAS‐based system. Double asterisks represent statistical significance of P ≤ 0.01 using Student’s t‐test [n = 3]. (d–f) Phenotype detection of RHL1 and RHL2 plants in mature period on sufficient nitrogen (SN) and deficient nitrogen (DN) conditions. The chlorophyll (designated as SPAD value) contents in ear leaves (d), total N contents in ear leaves (e) and seed (f), significant differences from the corresponding control values, *P ≤ 0.05, **P ≤ 0.01 using Student’s t‐test (n = 3).

To test whether the Mo17 allele of ZmNLP5 has a transactivation function, we conducted a transcriptional activity assay. The full length of ZmNLP5 cDNA from Mo17 or B73 was fused to the GAL4 DNA binding domain (G4DBD) under the control of the 35S promoter to generate the effector. The firefly luciferase (LUC) gene was placed under the control of six copies of the GAL4 binding site (UAS) to generate the reporter (Figure 7b). We co‐expressed the effector and reporter in tobacco (Nicotiana benthamiana) leaves. The expression of GAL4DBD‐ZmNLP5 (B73) significantly increased the LUC activity driven by the GAL4 promoter compared with that of the control (t‐test, P ≤ 0.01), whereas GAL4DBD‐ZmNLP5 (Mo17) cannot increase LUC activity compared with that of the control (Figure 7c), implying that the Mo17 allele of ZmNLP5 lacks a transactivation function.

To test whether the Mo17 allele of ZmNLP5 is associated with the mutant phenotype as we observed in the zmnlp5 mutant plants, we obtained two residual heterozygous IBM lines (derived from the intermated B73 × Mo17 recombinant inbred line population, Lee et al., 2002). Genotyping of those two residual heterozygous lines (RHL1: the Mo011 IBM line, RHL2: the Mo379 IBM line) confirmed that their genetic backgrounds are homogeneous except for the small region harbouring ZmNLP5, which is derived from B73 or Mo17, respectively (Table S6). Seeds within each residual heterozygous line were genotyped and classified as either a B73‐allele or Mo17‐allele depending on their allelic result at the ZmNLP5 locus (Figure S6). These near‐isogenic seeds differing at the ZmNLP5 locus within each line were then grown into mature plants and subjected to nitrate treatment. Phenotypic analysis showed that, for both residual heterozygous IBM lines under deficient nitrate conditions, the ear leaves and seed kernels of near‐isogenic plants with the Mo17 allele of ZmNLP5 had significantly lower N content than those having the B73 allele of ZmNLP5 (Figure 7d–f). As such, the Mo17 allele of ZmNLP5 seemingly confers a phenotype resembling that of the zmnlp5 mutant plants. These results showed that loss of function of ZmNLP5, either through natural or artificial mutation, impairs N assimilation in response to nitrogen deficiency.

Discussion

Improving the NUE of crops is crucial for minimizing N loss and reducing environmental pollution, which is very important for sustainable agriculture. For example, the overexpression of the nitrate transporter gene OsNRT1.1A confers high production and early maturation in rice (Wang et al., 2018b). Maize is an essential food and cash crop in the world, and a better understanding of how maize plants respond to fluctuating environmental N levels is therefore a critical step towards deciphering the molecular mechanism of N use in maize (Zanin et al., 2015; Trevisan et al., 2015; Yu et al., 2015). In this study, we report the functional characterization of ZmNLP5 for its role in modulating the N response, suggesting that ZmNLP5 is a potential candidate for improving NUE in maize production.

Given that the maize NLP family has nine family members, it is expected that there is a certain level of functional redundancy, which may explain why zmnlp5 mutant plants did not show severe growth defects. However, the fact that substantial numbers of genes were differentially regulated between WT and zmnlp5 mutant plants indicates that ZmNLP5 is a unique member of the maize NLP family. Notably, the mature plants of zmnlp5 accumulate less N in the ear leaves and kernels under N limitation, suggesting the involvement of ZmNLP5 in acquiring N nutrients from the soil under deficient N conditions. This is in agreement with the phylogenetic analysis where ZmNLP5 is located in a separate clade on the phylogenetic tree from the rest of the maize NLPs (Ge et al., 2018).

In Arabidopsis, AtNLP7 is considered a major regulator of the N response and has been shown to regulate many N‐responsive genes (Marchive et al., 2013). Phylogenetic analysis showed that ZmNLP5 is one of the close homologues of AtNLP7 (Ge et al., 2018). Although they are similar, in this study, we show that AtNLP7 and ZmNLP5 exhibit two differences in modulating the N response. First, AtNLP7 and ZmNLP5 respond to nitrate in different fashions. AtNLP7 is shown to be regulated by nitrate at the post‐translational level via a nuclear retention mechanism (Marchive et al., 2013), whereas we demonstrate that ZmNLP5 responds to nitrate by elevating expression transcriptionally and translationally (Figure 3a,b). Second, atnlp7 plants exhibit stronger mutant phenotype under sufficient N conditions than under deficient N conditions (Castaings et al., 2009); whereas the zmnlp5 mutant phenotype is more evident under deficient N conditions than under sufficient N conditions.

ZmNLP5 impacts a set of N‐responsive genes, including the NR and NIR genes, which are critical for reducing nitrate to ammonium. The upregulation of ZmNR1 and ZmNIR1 is mitigated in the zmnlp5 mutant plants in response to the nitrate supply (Figure S4). Furthermore, RNA‐seq analysis reveals that several key genes involved in N signalling and assimilation lose their N response in the zmnlp5 mutant plants in comparison to that of in the WT plants. Thus, we speculate that the N response and N assimilation pathways are negatively affected by the loss of ZmNLP5 and eventually cause less N accumulation, especially under DN conditions.

AtNIR1 (At2G15620) in Arabidopsis is the key enzyme that catalyzes the reduction of NO2 − to NH4 + (Crawford, 1995; Stitt, 1999; Crawford and Forde, 2002; Gojon et al., 2009). The nitrate‐induced expression of ZmNIR1.1 and ZmNIR1.2 in WT seedlings is impaired in zmnlp5 mutant seedlings. The sequence analysis of the promoters of ZmNIR1.1 and ZmNIR1.2 show that both contain the NRE (CTTN9AAG) candidate, respectively. However, it appears that ZmNLP5 can only bind to the NRE of ZmNIR1.1 (Figure 6b). Thus, ZmNIR1.2 may not be directly regulated by ZmNLP5; however, since there are nine NLP members in maize, it is possible that ZmNIR1.2 is bound by one of the other ZmNLPs, with which ZmNLP5 possibly interacts to regulate the expression of ZmNIR1.2. In Arabidopsis, AtNLP7 binds to the cis‐elements of target genes including the NREs, and such binding leads to either activation or suppression of the targets (Konishi and Yanagisawa, 2013). Our results showed that ZmNLP5 binds to NREs with either 9‐bp (ZmNRE) or 10‐bp (AtNRE) spacer, suggesting the flexibility of its binding activity. Furthermore, ZmNLP5 enhances the expression of the reporter gene fused with the ZmNIR1.1 promoter, indicating that ZmNLP5 can act as a transcription activator that transactivates gene expression.

Nitrite is a transient intermediate product in the pathway of N assimilation and is then being rapidly reduced to ammonium due to its adverse effects on plant growth. NIR is responsible for reducing nitrite to ammonium; decreased NIR activity has been shown to result in higher concentrations of nitrite and suppressed plant growth, for instance, in tobacco (Morot‐Gaudry‐Talarmain et al., 2002). In this study, we show that ZmNLP5 directly activates ZmNIR1.1 expression (Figure 6), consisting with that ZmNIR1.1 expression is significantly reduced in zmnlp5 mutants (Figure S4). It has been demonstrated that plants tend to transfer the most of nitrate to the shoots for assimilation (Hachiya et al., 2016). An in situ RNA assay showed that ZmNLP5 was located not only in the roots, but also in the vascular bundle of leaves. Together with the observation that higher content of nitrite at above‐ground parts of zmnlp5 seedlings under N limitation (Figure 4d), we speculate that the higher accumulation of nitrite in the shoot tissues under N deficiency is contributed, at least in part, by the suppression of ZmNIR1.1 expression in the mutant plants.

In addition to its elevated expression under N supply, it is worth mentioning that there is a subtle difference in terms of the cellular localization of ZmNLP5 in response to N availability in the environment. When plants grown under N deficiency are transferred to sufficient N conditions, the majority of the increased ZmNLP5 proteins are localized in the cytoplasm (Figure 3e). Certainly, this could be attributed to the fact that it takes some time for newly synthesized proteins to be transported from the cytoplasm to the nucleus. However, this does not explain why under long‐term sufficient N conditions, there is still a considerable amount of ZmNLP5 in the cytoplasm compared with less ZmNLP5 in the cytoplasm under N‐deficient conditions (Figure 3d). We propose that this is part of the strategy for ZmNLP5 to modulate the N response. The relatively ample ZmNLP5 proteins in the cytoplasm under normal conditions actually serve as ‘sentinel enzymes’. Once the plant is subject to N starvation, these ZmNLP5 proteins would be quickly moved into the nucleus to maintain ZmNLP5 at an adequate level in the nucleus without the necessity of synthesizing new proteins (Figure 3e). In this case, the adverse effect for NUE is minimized under N limitation. Collectively, we speculate that ZmNLP5 is an important element of the N signalling and assimilation pathway, especially under N limitation.

Experimental procedures

Plant materials and growth conditions

The zmnlp5 mutant (UFMu‐01175, http://www.maizegdb.org/uniformmu) in W22 (wild‐type: WT) background was obtained from Maize Genetics Cooperation Stock Center UniformMu Transposon Resource. Homozygous mutant plants were identified by PCR using the specific primers listed in Table S7. Two residual heterozygous lines (RHLs) (RHL1: Mo011 and RHL2: Mo379) were selected from a set of 189 RILs bred from the cross between B73 and Mo17 (the data from the genetic analysis are shown in Figure S5 and Table S6).

A modified Hoagland nutrient solution (Schlüter et al., 2012) was employed, with 15 mm KNO3 as a sufficient nitrogen (SN) solution and 0.15 mm KNO3 as a DN solution. The differences in potassium supply were balanced with KCl. The solutions were changed every 2 days. For expression pattern analysis in different tissues, tissues were collected from 21‐day‐old seedlings of WT grown in the hydroponic culture on SN solution. For nitrate induction experiments, 1‐week‐old WT seedlings subjected to 14 days of N‐deprivation (0 mm KNO3) were supplied with 15 mm KNO3. For expression pattern analysis under SN and DN conditions, maize seedlings of WT were grown for 21 days in hydroponic culture on SN solution and DN solution. For subcellular localization pattern analysis in response to N fluctuation, WT were grown for 14 days in hydroponic culture on DN, followed by N treatment (15 mm KNO3) for 1.5 h, and ended with N‐deprivation for 24 h.

For phenotype analysis at the seedling stage, seedlings were grown in hydroponic culture on SN solution and DN solution for 21 days. For phenotype analysis at the mature stage, plants were grown in 20‐L pots containing sand (80%) and vermiculite (20%) in a greenhouse. The 28‐day‐old plants of the zmnlp5 mutant, WT, and RHLs were treated with SN or DN solution until to the reproductive stage was reached. For transcriptome analysis, WT and zmnlp5 mutant plants were grown in SN conditions for 7 days, N‐deprived for 14 days, and then supplied with 15 mm KNO3 or 15 mm KCl (Castaings et al., 2009; Schlüter et al., 2012). After 30 min, seedlings (WT+N or WT−N, zmnlp5+N or zmnlp5−N) were collected with three independent biological replicates (Krouk et al., 2010; Ge et al., 2018).

Transcriptome analysis

For transcriptome analyses, each sample from three independent biological replicates was performed. After the quality control process, library construction and sequencing were performed according to Illumina instructions by the Berry Genomics company (Beijing, China). The analysis of RNA‐seq data was performed according to the previous study (Trapnell et al., 2013). FPKM (fragments per kilobase of exon per million fragments mapped) was calculated with Cufflinks (v2.1.1) (Trapnell et al., 2010) representing the expression level. The comparison of gene expression in response to N between WT and zmnlp5 mutant plants was performed using the ggplot2 package of R software. To identify DEGs, a paired t‐test on the log ratios was performed, assuming the same variances of the log ratios for all genes. The raw P‐values were adjusted by the Bonferroni method to avoid false positives in a multiple‐comparison context (Ge et al., 2003). The web server agriGO (http://bioinfo.cau.edu.cn/agriGO/analysis.php, Du et al., 2010) was used to perform singular enrichment analysis (SEA). To detect different N responses in the N metabolism pathway between WT and zmnlp5 mutant plants, the transcription profile of N metabolism was analyzed by the online MapMan tool (http://mapman.gabipd.org/mapmanstore; Usadel et al., 2005; Sekhon et al., 2012). The predicted gene function in this study is based on the annotation provided in MaizeGDB (http://www.maizegdb.org).

Gene regulatory network analysis

The DeGNServer (http://plantgrn.noble.org/DeGNServer/Analysis.jsp, Li et al., 2013) was used to infer networks for the maize pan‐transcriptome data, including 503 inbred lines (Hirsch et al., 2014). Normalized RNA‐seq data were imported to the DeGNServer for genome scale network construction based on context likelihood of relatedness (CLR; Faith et al., 2007), with Spearman’s rank correlation cutoff over 3.8. The NLP regulatory network was built up using nine ZmNLPs as seed genes based on the community‐finding algorithm GeNa (Maneesha et al., 2013). The network including 110 genes and 489 gene–gene associations was imported into Cytoscape 2.8.2 for network analysis and display. The web server KOBAS 2.0 (http://kobas.cbi.pku.edu.cn, Xie et al., 2011) was used to perform enrichment analysis for genes involved in the NLP regulatory network. In total, 110 genes were annotated to KEGG maize genes by protein sequence similarity mapping, and statistically significantly enriched pathways were identified based on Fisher’s exact test. The P‐values were corrected by FDR correction (Benjamini and Hochberg, 1995). Raw pathway terms were trimmed to remove any with the P ≤ 0.05.

Quantitative PCR (qPCR) and reverse transcription PCR (RT‐PCR)

Total RNA was isolated from collected samples using the SV Total RNA Isolation System kit (Promega, Madison, WI, USA) and digested with RNase‐free DNase I. RNA was then reverse transcribed to cDNA using the Prime Script RT Reagent kit (Takara, Dalian, China). Quantitative real‐time PCR was performed using a Bio‐Rad CFX96 system with SYBR® Premix Ex Taq™ II (TaKaRa). Primer pairs of tested genes were designed as described (Lv et al., 2016), and the housekeeping gene ZmUPF1 (GRMZM2G163444, Lin et al., 2014) was used as the internal control gene (Table S7). Each sample had three biological replicates to ensure the accuracy of the results.

Preparation of antibodies

Total protein was extracted from samples using the Plant Nuclei Isolation/Extraction kit (Sigma, St. Louis, MO, USA). For anti‐ZmNLP5 antibody production, a fragment of ZmNLP5 encoding Leu190–Ala389 was cloned into pGEX‐4T‐1 with a glutathione S‐transferase (GST) tag. Peptide synthesis, protein purification, and production of antibodies in rabbits were performed according to standard protocols of ABclonal Biotech Co., Ltd (Wuhan, China). The antibodies of UDPGP were purchased from Agrisera (catalogue nos. AS05086). The H3 antibodies were gifts from Dr Ren‐Tao Song (Shanghai University).

RNA in situ hybridization

The specific fragment of ZmNLP5 (163 bp) was amplified by primers in Table S7 and inserted into the pGEM‐T Easy vector (Promega) for sequencing. The sense probe was then generated using primers T7‐F and ZmNLP5‐in situR, and the antisense probe by primers ZmNLP5‐ in situF and T7‐R.

Sense and antisense probes were transcribed in vitro from the T7 promoter with T7 RNA polymerases using the digoxigenin RNA‐labelling kit (Roche, Basel, Switzerland). Tissues for in situ hybridization were fixed overnight in 4% (wt/vol) paraformaldehyde in phosphate buffer, pH 7.0, and embedded in Paraplast Plus (Sigma‐Aldrich, St. Louis, MO, USA). Non‐radioactive RNA in situ hybridization with digoxigenin‐labelled sense and antisense probes was performed on 8‐mm sections of different root parts as described by Coen et al. (1990).

Subcellular localization of ZmNLP5

Cytoplasmic protein and nuclear protein were extracted from samples using the Plant Nuclei Isolation/Extraction kit (Sigma‐Aldrich), and then immunoblot analysis was performed as described previously. The UDPGP protein was used as a cytosolic marker, and histone H3 was used as a nuclear marker.

Determination of nitrate, nitrite, and ammonium

Nitrate, nitrite, and ammonium in samples were extracted and quantified by assay kits purchased from COMIN (Production nos. SPYXY‐2‐G, ZXTD‐2‐G, and ZATD‐2‐G; Suzhou, China). Means and standard errors were calculated from three biological repetitions of three plants each.

Chromatin immunoprecipitation (ChIP)

ChIP with ZmNLP5‐specific antibody was carried out using an improved protocol (Saleh et al., 2008; Li et al., 2015), with some modifications to accommodate the low concentration of genome DNA in maize root. Briefly, 10 grams of roots of the WT seedlings were cross‐linked in cross‐linking buffer (125 ml) with applied vacuum for 15 min. Fixation was stopped by adding glycine ([final] = 0.125 m) under vacuum infiltration for an additional 10 min, followed by three washes with sterile ddH2O and ground to a powder in liquid nitrogen. Each sample was resuspended in 90 ml (two 50 ml tubes) of cold nuclei isolation buffer. The homogenate was filtered through four layers of cheesecloth before nuclei isolation. Nuclear‐enriched extracts were resuspended in 2 ml cold nuclei lysis buffer (50 mm HEPES pH 7.5, 150 mm NaCl, 1 mm EDTA, 0.3% SDS, 0.1% sodium deoxycholate, 1% Triton X‐100, 1 μg ml−1 pepstatin A, and 1 μg ml−1 aprotinin), followed by sonication for 9 min with a Covaris M220 sonicator (200–500 bp fragments). Antibodies against ZmNLP5 and the IgG control were used for IP. The precipitated DNA was extracted with phenol:chloroform:isoamyl acetates (25:24:1) and chloroform and subsequently analyzed by qPCR with appropriate primers (Table S7) and SYBR® Premix Ex Taq™ II (TaKaRa) using a Bio‐Rad CFX96 system.

Electrophoretic mobility shift assay

The cDNA fragment encoding the RWP‐RK domain was cloned into vector pCold (TaKaRa). The construct was transformed into Escherichia coli BL21 (DE3) cells, cultured at 37°C and refrigerated at 15°C for 30 min when the OD600 was 0.4–0.5. Then, the culture solution was added isopropyl β‐d‐1‐thiogalactopyranoside (IPTG) to a final concentration of 0.4 mm and continue cultured at 15°C for 24 h. His‐tagged recombinant proteins were purified using a His‐tag Protein Purification Kit (Beyotime, Shanghai, China). Oligonucleotide probes were synthesized and labelled with biotin at the 5′‐end. Purified proteins (80 ng) were mixed with 2.5 ng of probes at 25°C for 20 min in an EMSA/Gel‐Shift Binding Buffer (Beyotime, Shanghai, China). The mixture was separated by 6% native polyacrylamide gel electrophoresis (PAGE) in 0.5× TBE buffer. The DNA in the gel was then transferred to N+ nylon membranes (0.2 μm, Millipore, USA). The DNA on the membranes was detected using a Chemiluminescent EMSA Kit (Beyotime).

Transient transcriptional activity assay

For the transient transcriptional activity assays of the promoters, the −1000 bp upstream of the start codon of ZmNIR1.1 were cloned into vector pGreenII0800‐LUC to generate reporters for the dual‐luciferase assays. The full‐length ZmNLP5 cDNA was inserted into vector pMDC83‐35S to generate a 35S promoter‐driven ZmNLP5 effector. Transient dual‐luciferase assays were performed in onion (Allium cepa) epidermal cells, tungsten particles coated with DNA were used to bombard the onion epidermal cells in the PDS‐1000 system (Bio‐Rad, Hercules, CA, USA), and then the bombarded samples were incubated for at least 8 h in the dark at 25°C.

For transcriptional activity assays, a GAL4/UAS‐based system was used. The firefly luciferase (LUC) gene was placed under the control of six copies of the GAL4 binding site (UAS) to generate the reporter, and the reference plasmid harbours the P35S::REN fusion gene. The full‐length ZmNLP5 cDNA (from Mo17 or B73) was fused to the GAL4 DNA binding domain (G4DBD) under the control of the 35S promoter to generate the effector. The reporter and the effector or the control were transformed into Agrobacterium GV3101 (pSoup‐P19) and then co‐infiltrated into leaves of N. benthamiana. After incubation in the dark for 24 h, the plants were subjected to normal conditions and growth for 48 h. The leaves were observed using a low‐light cooled charged coupled device (CCD) imaging apparatus (Tanon 5200 Multi).

The firefly luciferase (LUC) activity and Renilla luciferase (REN) activity were measured by a Dual‐Luciferase Reporter Gene Assay Kit (Beyotime, Shanghai, China) using the Tecan M200 system. The ratio between LUC and REN activities was measured three times.

Functional complementation test

The full‐length ZmNLP5 coding sequence amplified from B73 was cloned into vector pHB (Mao et al., 2005). The construct pHB‐ZmNLP5 was transformed into immature embryos of maize inbred line B104 through Agrobacterium‐mediated transformation according to Ishida et al. (2007). The transgenic lines (T1) were hybridized to the homozygous mutant plants (zmnlp5) and then self‐pollinated to produce the respective F2 population. Two hundred F2 seedlings were grown in hydroponic culture on DN solution for 21 days. Seedlings containing homozygous zmnlp5 alleles were identified by P1 + P2 and P3 + P2 primers, and seedlings containing the ZmNLP5 transgene were identified by primers for the bar gene. The B104, zmnlp5, and WT were used as controls. The transcript levels of ZmNLP5 and nitrate content in the root tissues were then detected in the identified seedlings (ZmNLP5/zmnlp5, zmnlp5/zmnlp5).

The raw sequence data from this article were submitted to NCBI (SRA accession: PRJNA494286).

Author contributions

MG and YW performed experiments; LN, HD, FL, TZ, SL, and HL provided technical assistance; BH, YL, LZ analyzed data; HZ, MG YL and YW conceived the project and wrote the article with contributions of all the authors.

Conflict of interest statement

The authors declare that they have no competing interests.

Availability of data and materials

The datasets supporting the conclusions of this article have been submitted to NCBI SRA database (https://www.ncbi.nlm.nih.gov/sra/) under the accession number PRJNA494286.

Figure S1. Expression response of ZmNLP3, ZmNLP5, and ZmNLP7 to nitrate.

Click here for additional data file.

Figure S2. Alignment of the amino acid sequence of AtNLP7 and ZmNLP5.

Click here for additional data file.

Figure S3. Preparation of ZmNLP5‐specific antibody and specificity verification.

Click here for additional data file.

Figure S4. N‐responsive expression of nitrate reductase (NR) and nitrite reductase (NIR) genes.

Click here for additional data file.

Figure S5. Alignment of the cDNA sequences of ZmNLP5 in B73 and Mo17.

Click here for additional data file.

Figure S6. Identification of two residual heterozygous lines (RHL) for ZmNLP5 from the intermated B73 × Mo17 (IBM) population.

Click here for additional data file.

Table S1 Summary of the ZmNLP network.

Click here for additional data file.

Table S2. Pathways significantly enriched in the ZmNLPs network.

Click here for additional data file.

Table S3. Summary of the ZmNLP5 subnetwork.

Click here for additional data file.

Table S4. Genes having nitrate‐induced differential expression in WT but not in zmnlp5 mutant plants.

Click here for additional data file.

Table S5. Allelic variation of ZmNLP5 ORF among B73, W22, and Mo17.

Click here for additional data file.

Table S6. Genotype analysis of RHLs selected from the intermated B73 × Mo17 recombinant inbred line (IBM) population.

Click here for additional data file.

Table S7. Primer list.

Click here for additional data file.

Click here for additional data file.