To character the NLP members, the protein sequences of NLP family in Arabidopsis and rice were used as query sequences to perform BLASTp in 34 species in plant genomics resource¹. The BLASTp were used with an e-value cutoff set to 1e-003. The isolated protein sequences were examined using the domain analysis program Hidden Markov Model (HMM) of SMART² and Pfam³ with the default cutoff parameters (Bateman et al., 2002; Letunic et al., 2012). We obtained the sequences with the Pfam number PF02042.10 and PF00564.19, which contained a typical RWP-RK and PB1 domains, from the maize genome sequences using a Perl-based script.

The full length NLP protein sequence of Arabidopsis, maize, sorghum, and rice were used for constructing phylogenetic tree with the neighbor-joining (NJ) method of the MEGA7 program⁴ using the p-distance and complete default option parameters. The reliability of the obtained trees was tested using a bootstrapping method with 1000 replicates. The images of the phylogenetic trees were drawn using MEGA7. The genomic and CDS of NLP members were downloaded from the plantGDB database⁵. The gene structures of NLP genes were generated using the GSDS⁶. The position of GAF-Like, RWP-RK and PB1 domain was indicated from the Pfam results.

The amino acid sequences of ZmNLPs were aligned using the ClustalX program with BLOSUM30 as the protein-weight matrix. The chromosomal location was retrieved from the maize genome data⁷, and the genes were mapped to the chromosomes using MapDraw program. The isoelectric points and molecular weights of these proteins were obtained with the help of the proteomics and sequence analysis tools on the ExPASy Proteomics Server⁸.

The conserved motif was discovered by Motif discovery in MEME⁹ using 248 NLP proteins in 34 species. The logos were created and downloaded by Multiple Em for Motif Elicitation.

Maize B73 inbred seeds were grown in the fields every year (2014–2016) and a part of 0–20 cm underground roots and ear leaves(whole) were collected at the stage vegetative 3 (V3), vegetative 5 (V5), vegetative 9 (V9), vegetative 13 (V13), reproductive 1 (R1), and vegetative 18 (V18) (Sekhon et al., 2011). These are important stages for the growth and development of maize. A part of 0–20 cm underground roots, ear leaf (whole), ear stem (whole), ear sheath (whole), tassels (whole), and grains in reproductive 1 stage (15 days after pollination) were collected for testing the gene expression every year. Maize plants used for testing the induction of nitrate responsive genes were grown on matrix watered with B5 solution to 3-leaf stage, and then the endosperm were removed from the seedlings. The uniform seedlings were transferred to hydroponics system (Liu W. et al., 2017) with 2.5 mM ammonium succinate for 2 days, followed by the treatments with 0.2 mM or 10 mM KNO₃/KCl for 2 h. The roots were collected for testing the expression of ZmNLP genes. Maize seeds were grown on matrix watered with different nitrate concentrations (0, 0.2, 2.5, 5, and 10 mM KNO₃) for 2 weeks and then the shoots and roots were collected for detecting the expression of ZmNLP genes.

Arabidopsis plants used for testing the YFP fluorescence intensity were grown on plates with nitrate medium (10 mM KNO₃) for 4 days and observed under a fluorescence microscope (Nikon Eclipse Ti-s). The fluorescence of roots was quantified using Image J. For Arabidopsis plants used for qPCR analysis of nitrate responsive genes, seedlings were grown in aseptic hydroponics (initial medium with 2.5 mM ammonium succinate) as previously described (Xu et al., 2016) for 7 days and then treated with 10 mM KNO₃ or KCl for 2 h. RNA was extracted from the roots for testing the expression of nitrate responsive genes.

Total RNA was isolated using a total RNA miniprep kit (CWBIO). Template cDNA samples were prepared using the EasyScript One-Step gDNA Removal and cDNA Synthesis SuperMix kit (TRAN). qPCR reaction was prepared with UltraSYBR Mixture (With Rox I) followed by performing qPCR using QuantStudio^TM 6 (Thermo Fisher Scientific). ZmActin1 (J01238), ZmUBCE (GRMZM2G027378) (Manoli et al., 2012), AtTUB2 (At5g62690), and AtActin8 (At1g49240) were used as the internal reference genes.

The coding regions of ZmNLP6 and ZmNLP8 containing the stop codon were cloned into pMDC43 (Invitrogen). The pMDC43-GFP-ZmNLP6 and pMDC43-GFP-ZmNLP8 were transiently transformed into Nicotiana benthamiana epidermal cells and Arabidopsis mesophyll protoplasts. The GFP fluorescence was observed using a confocal microscope (Leica TCS SP5II).

Yeast one-hybrid (Y1H) assay was performed using the system as described previously (Lin et al., 2007). The full length of cDNA for ZmNLP6 and ZmNLP8 were ligated into the pJG4 vector and the putative NRE from the 5′ and 3′ flanking sequences of nitrate uptake and assimilation genes in maize were ligated into the pLacZ-2U vector. The constructed vectors were co-transferred into yeast EGY48 competent cells and grew the cells on SD/-Trp/-Ura medium for 4 days, and then positive clones were transferred into a new plate of SD/-Trp/-Ura with X-gal.

The nlp7-4 mutant was isolated from EMS-treated seeds of transgenic lines containing a nitrate-regulated promoter fused to a yellow fluorescent protein (NRP::YFP) (Wang et al., 2009). ZmNLP6 and ZmNLP8 were amplified from the B73 inbred line by PCR. The constructs p35S::ZmNLP6 and p35S::ZmNLP8 in binary vector pSuper1300 were transformed into the nlp7-4 mutant (Xu et al., 2016) using the Agrobacterium-mediated floral dip method. Homozygous transgenic lines were isolated and two of these lines for each ZmNLP gene were selected for further investigation. All primers were listed in Supplementary Table S1.

For nitrate content detection, seedlings of WT, nlp7-4, and transgenic lines were grown in 1/2 MS medium (pH = 5.7, 10 mM KNO₃ and 10 mM NH₄NO₃) for 7 days and then collected 0.05–0.1 g of the whole seedlings. The samples were milled into powder using a RETCH MM400 and then added into l mL ddH₂O followed by boiling at 100 °C for 30 min. The cooled sample was centrifuged at 13000 rpm for 10 min and 400 μL supernatant was collected to the flow cell. The nitrate content was detected by hydrazine reduction method using AutoAnalyzer 3 continuous flow analytical system (SEAL Analytical).

Nitrate reductase activity was detected by sulfanilamide colorimetric method as described previously (Ferrario-Mery et al., 1998). Amino acid content in tissue was determined using ninhydrin colorimetric analysis as described previously (Rosen, 1957).

For root architecture, WT, nlp7-4, ZmNLP6 and ZmNLP8 transgenic lines grown vertically on agar plates containing 0.2, 2.5, and 5 mM KNO_3, respectively, for 10 days. The primary root length was measured by Image J and the lateral root number was counted under a dissecting microscope (Optec). For biomass measurement, seedlings were grown on KNO₃ solid medium 0.2, 2.5, and 5 mM KNO₃, respectively, for 10 days and fresh weight was measured or seedlings were dried under 65°C for 3 days to obtain the dry weight.

For yield measurement, seedlings were grown in pots with matrix at 22°C, 16 h light/8 h dark photoperiod in growth room and watered with 0.2 mM or 5 mM KNO₃. The seeds were collected when plants were mature and dried at 30°C for 2 weeks.

To find the NLP homologous genes in maize, BLASTp of 34 species using NLP family amino acid sequences of Arabidopsis and rice as queries were performed from the plant genomics resource¹⁰. The search results were further verified by HMM of SMART and Pfam tools, and the proteins containing RWP-RK and PB1 domains were considered as NLP homologs. By using this strategy, 248 proteins were defined as NLP homologs in 34 species (Supplementary Figure S1). NLP proteins exist in all land plants we searched. Nine NLP members were found in the maize genome (Table 1). We named these nine members, ZmNLP1 to 9, mainly following their chromosomal locations except GRMZM2G475305 as ZmNLP6 to be consistent with nomenclatures proposed in the previous study (Burdo et al., 2014) and GRMZM2G176655 on chromosome 6 alternately named ZmNLP4. The genes of the ZmNLP family are distributed over eight of the 10 maize chromosomes (except Chr.4 and Chr.9) (Supplementary Figure S2). Pairwise comparison of the ZmNLP proteins revealed levels of identity ranging from 16.67% between ZmNLP5 and ZmNLP7 to 66.08% between ZmNLP6 and ZmNLP8 (Supplementary Figure S3). Three sister pairs (ZmNLP3/7, ZmNLP2/9, and ZmNLP6/8) of paralogs with high identity (>60%) and strong bootstrap support (>90%) (Figure 1A) were found, implicating a paralogous pattern of ZmNLP divergence by gene duplication in maize. The alignment results of the nine full length ZmNLP amino acid sequences were shown in Supplementary Figure S4. Two conserved domains, RWP-RK and PB1, were found to be separated by an area of low similarity in the C-termini of ZmNLPs.

To analyze the phylogenetic relationship of NLP proteins, the unrooted phylogenetic tree was generated based on the full length amino acid sequences of NLP proteins from Arabidopsis (9), maize (9), sorghum (5), and rice (6) using MEGA7.0. As shown in Figure 1A, the NLPs were divided into 3 groups (Group I, II, and III with different colors). In each group, the NLP proteins from maize, sorghum, and rice were closely clustered. Exon/intron structures of these NLPs, shown in Figure 1B, revealed an obvious difference among groups: a module with four exons and three introns was found in most group I members with the smallest size in the second exon. The group II members mainly contain five exons and four introns, and two smaller exons separate three bigger ones. A module of five exons and four introns also found in group III members, but the third exon is the smallest one. Thus, the length and position of introns make great contributions to the diversity of gene structure in each group. For the ZmNLP proteins, the RWP-RK and PB1 domains were found to be well conserved in all tested members, and GAF-like domain were also found in the N-termini of ZmNLPs (Figure 1C). In addition, the results revealed two novel motifs, GAF motif and GSL motif, in the NLP family by using MEME analysis (Figures 1D,E and Supplementary Figure S4). The GAF motif (SX₄FX₁₀GLPGR), which is located in the N-termini of NLPs, is the main part of the GAF-Like domain and critical for nitrate signaling transduction in Arabidopsis (Wang et al., 2009; Liu K.H. et al., 2017). GSL motif (LX₂LX₃FXGSLKD) exists just in front of the RWP-PK domain and with conserved GSL residues as a core. The GSL motif was well conserved in 236 NLP proteins in 34 species, but not found in the RKD family proteins, so that it can be regarded as a marker for distinguishing the NLP proteins from RKD proteins.

To better understand the function of each gene in the ZmNLP family, we examined their temporal and spatial expressions in maize B73 inbred lines by using qPCR. The results showed that different ZmNLP genes exhibited various expression patterns (Figures 2A,B). Among these genes, ZmNLP6 and ZmNLP8 showed significant higher expression levels in almost all tested stages and tissues, while the expression levels of ZmNLP1 and ZmNLP7 were much lower. ZmNLP4 and ZmNL9 exhibited the lowest expression levels in diverse organs and stages. Remarkably, the expression of ZmNLP6 and ZmNLP8 was the highest in the vegetative 13 (V13) stage in leaves and reproductive 1 (R1) stage in roots. V13 is the stage of tasseling and R1 is grouting stage, both of which are essential for maize production, suggesting that ZmNLP6 and ZmNLP8 may play outsized roles in these stages critical to maximizing maize yield. The expression of ZmNLP6 and ZmNLP8 was higher in roots, tassels, and grains, while ZmNLP2 and ZmNLP3 was higher in leaves than in other organs (Figure 2B). The expression of ZmNLP1 was relatively uniform, while ZmNLP7 was more highly expressed in both roots and tassels. The expression of ZmNLP5 was relatively low, mainly expressed in leaves, tassels, and grains. We also used the second housekeeping gene ZmUBCE for testing the expression of these genes and similar results were found (Supplementary Figure S5). These results suggest that ZmNLP6 and ZmNLP8 may play more important roles in different developmental stages and tissues compared with the other members of ZmNLPs in maize.

To test if ZmNLPs are responsive to different nitrate conditions, we investigated the expression of all ZmNLP genes in roots. Seedlings were grown on 1/2MS medium (10 mM KNO₃ and 10 mM NH₄NO₃) for 2 weeks, and then transferred into 2.5 mM ammonium succinate for 2 days followed by treatments with 0.2 or 10 mM KNO₃ for 2 h. As shown in the Figure 2C, ZmNLP3 and ZmNLP4 were induced by 0.2 mM KNO₃ in roots. After 10 mM KNO₃ treatment, the expression of most ZmNLPs was not induced except ZmNLP2 (Figure 2C and Supplementary Figure S5). Thus, most ZmNLP genes were not induced by nitrate.

To test the expression patterns of ZmNLP genes under various nitrate conditions, 2-week-old seedlings were grown under 0, 0.2, 2.5, 5, and 10 mM KNO₃ conditions and RNAs were extracted from the shoots and roots for qPCR analysis. The results showed that the expression levels of the ZmNLP4, ZmNLP6, and ZmNP8 were generally increased with the higher concentrations of nitrate in shoots, while the highest peak was observed at 2.5 mM nitrate concentration for ZmNLP6 and ZmNLP8 in the roots (Figure 3). However, the expression of ZmNLP3 and ZmNLP7 generally decreased in both shoots and roots with the increase of nitrate concentrations. For ZmNLP1, the expression was stable in shoots while decreasing with increased nitrate concentrations (Figure 3). These data indicate that different ZmNLPs exhibit various expression profiles. The expression of ZmNLP6 and ZmNLP8 in shoots was the highest and increased with increasing nitrate concentrations, indicating that both genes are the strongest candidates for playing important roles in regulating nitrate metabolism in maize. The second housekeeping gene ZmUBCE were used for testing the expression of these genes and the results were similar to those obtained by ZmActin1 (Supplementary Figure S6). Therefore, we focused on ZmNLP6 and ZmNLP8 in the following studies.

AtNLP7 has been reported as a key regulator in nitrate signaling (Marchive et al., 2013). The Arabidopsis mutant nlp7-4 with a mutation (C to T) at the position 62, converting Gln to a stop codon, was identified by a forward genetic screen (Xu et al., 2016). The nlp7-4 plants contain NRP::YFP (a nitrate-regulated promoter fused to a yellow fluorescent protein) construct and are defective in nitrate regulation of nitrate responsive genes. To investigate if ZmNLP6 and ZmNLP8 are involved in nitrate signaling, we cloned them from the cDNAs of B73 inbred line and overexpressed them in the mutant nlp7-4, respectively, using a 35S promoter. Transgenic plants overexpressing the ZmNLP6 and ZmNLP8 were generated and two independent homozygous lines (ZmNLP6/nlp7-4-1 and ZmNLP6/nlp7-4-12 for ZmNLP6 transgenic lines, ZmNLP8/nlp7-4-8 and ZmNLP8/nlp7-4-16 for ZmNLP8 transgenic lines) were selected for following investigation (Figure 4A). Nitrate-responsive expression was monitored using a construct containing the YFP fused to a nitrate-responsive synthetic promoter (NRP) (Wang et al., 2009).

The transgenic lines were grown on nitrate medium and YFP fluorescence of these lines was found to be partially or completely restored to WT levels (Figures 4B,C), indicating that the ZmNLP6 and ZmNLP8 may function in nitrate signaling. Furthermore, we investigated the expression of endogenous nitrate-responsive genes including NIA1, NiR, and NRT2.1 and found that the induction levels of all three genes were recovered in ZmNLP6 and ZmNLP8 transgenic lines after nitrate treatments (Figure 4D). The expressions of these genes were also normalized to the percentage of second housekeeping gene AtActin8 and similar results were found (Supplementary Figure S7). These results suggest that ZmNLP6 and ZmNLP8 can complement the Atnlp7 mutant and thus play important roles in nitrate signaling.

To investigate the subcellular localization of ZmNLP6 and ZmNLP8 proteins, the full length CDS of the two genes driven by 35S promoter were fused with GFP in N-terminal. The 35S::GFP-ZmNLP6 and 35S::GFP-ZmNLP8 constructs were transiently transformed into Nicotiana benthamiana epidermal cells and Arabidopsis mesophyll protoplasts, the results showed that both ZmNLP6 and ZmNLP8 were mainly localized in the nucleus and slightly in cytosol (Figures 5A,B). As AtNLP7 has been reported to be regulated by nitrate via a nuclear retention mechanism (Marchive et al., 2013), we next tested the subcellular localization of ZmNLP6 and ZmNLP8 proteins under nitrate starvation and nitrate re-addition conditions. Both ZmNLP6 and ZmNLP8 were found to be localized in cytosol after nitrate starvation (Figures 5C,D) while in the nucleus when nitrate was resupplied (Figures 5E,F). Thus, ZmNLP6 and ZmNLP8 are mainly localized in the nucleus and slightly in cytosol in the presence of nitrate while both proteins are localized in cytosol when nitrate is absent.

Previous studies have shown that nitrate assimilation is impaired and the nitrate content is increased in nlp7 mutants (Castaings et al., 2009). To test if ZmNLP6 and ZmNLP8 affect nitrate assimilation, we measured nitrate content in transgenic lines and found that the increased nitrate content in nlp7-4 was recovered to WT levels (Figure 6A). To investigate if this recovery is associated with nitrate reduction, we examined the NR activity and found that this activity was restored in the transgenic lines (Figure 6B). Furthermore, the deficiency of amino acid in nlp7-4 mutant was completely rescued when ZmNLP6 and ZmNLP8 were overexpressed in nlp7-4 mutant (Figure 6C). Taken together, these data indicate that ZmNLP6 and ZmNLP8 are involved in regulating nitrate assimilation when overexpressed in Arabidopsis.

In order to further explore the underling mechanism whereby ZmNLP6 and ZmNLP8 affect nitrate content, we examined nitrate accumulation and the expression of several genes involved in nitrate assimilation. Plants were grown on 2.5 mM ammonium succinate for 7 days and then treated with 5 mM KNO₃ for 0, 0.25, 0.5, 1, 2, and 4 h in the presence of 2.5 mM ammonium succinate. The nitrate content in whole seedlings was investigated and the results showed that no difference was found in nitrate accumulation among the WT, nlp7-4, and transgenic lines (Supplementary Figure S8). However, the expression of Gln1.1, Gln1.3, NIA2, and NiR in transgenic lines was recovered to the levels even higher than in WT (Supplementary Figure S9). These findings suggest that ZmNLP6 and ZmNLP8 can regulate the nitrate assimilation more strongly than nitrate accumulation.

It has been found that Arabidopsis NLP proteins can bind nitrate regulatory elements (NREs) to regulate the nitrate responsive genes (Konishi and Yanagisawa, 2013). To investigate whether ZmNLP6 and ZmNLP8 can bind NREs in maize, Y1H assay was performed. Firstly, we searched for NREs with a module (tGACcCTTN₁₀AAGagtcc) from the 5′ and 3′ franking sequences of nitrate uptake and assimilation genes. Nine putative NREs were obtained as shown in Supplementary Table S2. Then, the candidate NREs were used for testing the binding activity of ZmNLP6 and ZmNLP8. The results showed that ZmNLP6 and ZmNLP8 could bind the NRE-like motifs of ZmNRT1.2 (3′UTR, with position 7299–7326) and ZmNiR2 (promoter, -149-124), respectively (Figures 7A–C). But no binding activity was found between ZmNLP6/ZmNLP8 and other NREs (Supplemental Figure S10). These data suggest that the ZmNLP6 and ZmNLP8 proteins can bind putative NREs similar to what was reported for Arabidopsis.

Low NUE in agriculture system is a global problem and therefore we assessed the potential of both genes to affect root architecture and NUE of in Arabidopsis. We first examined the primary root length and lateral root number in plants grown vertically on the media with different nitrate concentrations (0.2, 2.5, and 5 mM KNO₃). The results showed that the length of primary roots and number of lateral roots were higher in ZmNLP6/nlp7-4 and ZmNLP8/nlp7-4 transgenic lines than in WT and nlp7-4 mutant under these three nitrate conditions (Figures 8A–C). To determine whether ZmNLP6 and ZmNLP8 can enhance NUE in plants, we investigated the biomass of WT, nlp7-4, and ZmNLP6/nlp7-4 and ZmNLP8/nlp7-4 transgenic lines under different nitrate concentrations. The results showed that the transgenic seedlings grew bigger than WT and nlp7-4 and the biomass of the whole seedlings increased by 15, 35, and 40% more than WT under 0.2, 2.5, and 5 mM KNO₃ conditions (Figures 8D–F), indicating that the ZmNLP6 and ZmNLP8 can rescue the deficient growth phenotype of nlp7-4 mutant and promote plant growth under both low and high nitrate conditions.

Seed yield is an important trait for agricultural production and also for assessing the NUE of plants. Thus, we investigated the seed yield of ZmNLP6/nlp7-4 and ZmNLP8/nlp7-4 transgenic lines grown under both high and low nitrate conditions. The results showed that the yield per plant was higher in ZmNLP6/nlp7-4 and ZmNLP8/nlp7-4 transgenic lines by 44 and 45%, respectively, than in WT under low nitrate conditions (Figures 8G,H). However, no significant difference was found between the transgenic lines and WT when grown under high nitrate conditions (Supplementary Figure S11). These data suggest that ZmNLP6 and ZmNLP8 may improve plant NUE under low nitrate conditions.