The Role of Gibberellins in Regulation of Nitrogen Uptake and Physiological Traits in Maize Responding to Nitrogen Availability

Modified gibberellin (GA) signaling leads to semi-dwarfism with low nitrogen (N) use efficiency (NUE) in crops. An understanding of GA-mediated N uptake is essential for the development of crops with improved NUE. The function of GA in modulating N uptake capacity and nitrate (NO3−) transporters (NRTs) was analyzed in the GA synthesis-deficient mutant zmga3ox grown under low (LN) and sufficient (SN) N conditions. LN significantly suppressed the production of GA1, GA3, and GA4, and the zmga3ox plants showed more sensitivity in shoots as well as LN stress. Moreover, the higher anthocyanin accumulation and the decrease of chlorophyll content were also recorded. The net NO3− fluxes and 15N content were decreased in zmga3ox plants under both LN and SN conditions. Exogenous GA3 could restore the NO3− uptake in zmga3ox plants, but uniconazole repressed NO3− uptake. Moreover, the transcript levels of ZmNRT2.1/2.2 were downregulated in zmga3ox plants, while the GA3 application enhanced the expression level. Furthermore, the RNA-seq analyses identified several transcription factors that are involved in the GA-mediated transcriptional operation of NRTs related genes. These findings revealed that GAs influenced N uptake involved in the transcriptional regulation of NRTs and physiological responses in maize responding to nitrogen supply.


Introduction
Gibberellins (GAs) are phytohormones containing a tetracyclic diterpenoid structure that controls diverse aspects of plant growth and development in higher plants in response to environmental stimuli [1,2]. GA biosynthesis is catalyzed by a series of key enzymes, among which GA 20-oxidase (GA20ox) and GA 3-oxidase (GA3ox) are involved in the synthesis of bioactive GAs, and its deactivation is catalyzed by GA 2-oxidase (GA2ox) [1]. Mutations in GA biosynthesis genes (GA20ox and GA3ox) decrease the levels of endogenous GAs and lead to dwarfism, whereas exogenous GA treatment can restore normal growth [3,4]. DELLA proteins (DELLAs) are the major component in the GA signaling pathway. GAs promote plant growth by stimulating the degradation of the growth-repressing DELLAs [5]. Most importantly, dwarfism in rice and wheat cultivars was selected for enhancing global Thereby, the present study demonstrates the role of the GA in regulating the physiological responses in maize in correspondence to N supply.

Characterization of zmga3ox Mutant in Maize
The GA3ox enzymes participate in the synthesis of bioactive GAs and catalyze the conversion of GA9 to bioactive GA4 and GA20 to GA1 and GA3 (Figure 1a). In order to verify the role of GAs in response to N supply in maize, the knockout mutants were generated using a type II CRISPR-Cas9 system [29]. A knockout mutant named zmga3ox was obtained, which conferred a 34-bp deletion causing a frameshift in ZmGA3ox (Figure 1b-d). Compared to the wild-type plant, the expression level of ZmGA3ox was hardly detected in the shoots and roots of zmga3ox plant ( Figure S1). Meanwhile, the zmga3ox plant displayed growth inhibition, and exogenous GA3 could restore normal growth ( Figure  1e). The concentrations of GA1, GA3, and GA4 in zmga3ox plant were significantly lower than those in wild-type plants ( Figure 1f). Also, the zmga3ox plant had greater levels of GA9 and GA20 compared to the wild-type (Figure 1g). These results suggest that zmga3ox is an endogenous GA-defective mutant. The contents of GA9 and GA20 in shoots of zmga3ox and wild-type seedlings. Values were the means ± SD (n = 3). The asterisks indicated significant difference between wild-type and zmga3ox plants, as evaluated by Student's t-tests ** p < 0.01.

GAs Altered the Anthocyanin Accumulation and Chlorophyll Content in Leaves in Response to NO3 − Supply
The anthocyanin and chlorophyll contents were determined in wild-type and zmga3ox seedlings by growing in LN and SN solutions with or without GA3 and Ucz. The zmga3ox plants were more sensitive to LN supply than the wild-type plant, and also the zmga3ox leaves were more purple. As  3 , and GA 4 in shoots of zmga3ox and wild-type seedlings. ND: not detected. (g) The contents of GA 9 and GA 20 in shoots of zmga3ox and wild-type seedlings. Values were the means ± SD (n = 3). The asterisks indicated significant difference between wild-type and zmga3ox plants, as evaluated by Student's t-tests ** p < 0.01.

GAs Altered the Anthocyanin Accumulation and Chlorophyll Content in Leaves in Response to NO 3 − Supply
The anthocyanin and chlorophyll contents were determined in wild-type and zmga3ox seedlings by growing in LN and SN solutions with or without GA 3 and Ucz. The zmga3ox plants were more sensitive to LN supply than the wild-type plant, and also the zmga3ox leaves were more purple. As expected, the GA 3 -treated leaves were less purple, but Ucz strengthened such effect. Moreover, the application of GA 3 and Ucz had no effect on the anthocyanin accumulation under the SN condition ( Figure 2a). LN significantly enhanced the accumulation of the anthocyanin in zmga3ox leaves, and the anthocyanin content was increased by 3-fold under the LN condition ( Figure 2b). In addition, the chlorophyll content was decreased in zmga3ox leaves compared to the wild-type leaves under LN condition, and exogenous GA 3 increased the chlorophyll content in both wild-type and zmga3ox leaves; however, Ucz treatment slightly decreased the chlorophyll content in both wild-type and zmga3ox leaves (Figure 2c). GA 3 treatment declined the chlorophyll content in zmga3ox leaves under SN condition, while Ucz treatment enhanced the chlorophyll content. In addition, the photosynthetic rate, leaf areas, and O 2 − production rate were also measured in the wild-type and zmga3ox leaves treated with or without GA 3 application. The zmga3ox leaves showed higher the photosynthetic rate than the wild-type leaves, while the photosynthetic rate was decreased by GA 3 treatment ( Figure S2a). Moreover, the leaf areas of the zmga3ox plants were significantly less than those of the wild-type plants ( Figure S2b). Furthermore, the O 2 − production rate in zmga3ox leaves showed slightly lower than that in wild-type leaves ( Figure S2c). expected, the GA3-treated leaves were less purple, but Ucz strengthened such effect. Moreover, the application of GA3 and Ucz had no effect on the anthocyanin accumulation under the SN condition ( Figure 2a). LN significantly enhanced the accumulation of the anthocyanin in zmga3ox leaves, and the anthocyanin content was increased by 3-fold under the LN condition ( Figure 2b). In addition, the chlorophyll content was decreased in zmga3ox leaves compared to the wild-type leaves under LN condition, and exogenous GA3 increased the chlorophyll content in both wild-type and zmga3ox leaves; however, Ucz treatment slightly decreased the chlorophyll content in both wild-type and zmga3ox leaves (Figure 2c). GA3 treatment declined the chlorophyll content in zmga3ox leaves under SN condition, while Ucz treatment enhanced the chlorophyll content. In addition, the photosynthetic rate, leaf areas, and O2 − production rate were also measured in the wild-type and zmga3ox leaves treated with or without GA3 application. The zmga3ox leaves showed higher the photosynthetic rate than the wild-type leaves, while the photosynthetic rate was decreased by GA3 treatment ( Figure  S2a). Moreover, the leaf areas of the zmga3ox plants were significantly less than those of the wild-type plants ( Figure S2b). Furthermore, the O2 − production rate in zmga3ox leaves showed slightly lower than that in wild-type leaves ( Figure S2c).

Supply of NO 3 − Modulated the Transcript Expression of GA Biosynthesis-and Metabolism-Related Genes
In order to understand the role of GAs in maize corresponding to NO 3 − supply, the content of GA 1 , GA 3, and GA 4 was determined by using the HPLC-MS/MS method. LN significantly decreased the content of GA 1 , GA 3, and GA 4 in roots as compared to SN (Figure 3a). The GA 1 content could not be detected in the roots under the LN condition, while the GA 3 and GA 4 levels were lower under LN condition than those under the SN condition. The expression levels of ent-kaurene synthase genes ZmKS2 and ZmKS4 in GA biosynthesis pathway were significantly downregulated by LN in comparison to SN (Figure 3b). Similarly, LN significantly decreased the expression levels of ZmGA20ox1 and ZmGA20ox4 (Figure 3c). LN enhanced the expression levels of catabolic genes ZmGA2ox1, ZmGA2ox5, ZmGA2ox6, ZmGA2ox7, and ZmGA2ox12 (Figure 3d). The GA 3 contents, as well as the expression of genes encoding enzymes of GA metabolism, were also determined in the zmga3ox mutant. As shown in Figure S3, the levels of GA 3 were less under LN condition than those under the SN condition ( Figure  S3a). LN significantly down-regulated the expression levels of GA biosynthetic enzyme genes ZmKS2, ZmKS4, ZmGA20ox1, and ZmGA20ox4 ( Figure S3b,c). Inversely, LN enhanced the expression levels of catabolic enzyme genes ZmGA2ox1, ZmGA2ox5, ZmGA2ox6, ZmGA2ox7, and ZmGA2ox12 ( Figure S3d). The results indicated that LN caused a reduction in the bioactive GA S levels in roots by regulating the transcripts encoding GA 20-oxidases and GA 2-oxidases. zmga3ox shoots treated with GA3 or Ucz at 7 d after LN or SN treatment. Values were the means ± SD (n = 3). Different letters indicated significant difference calculated by Fisher's LSD (p < 0.05).

Supply of NO3 − Modulated the Transcript Expression of GA Biosynthesis-and Metabolism-Related Genes
In order to understand the role of GAs in maize corresponding to NO3 -supply, the content of GA1, GA3, and GA4 was determined by using the HPLC-MS/MS method. LN significantly decreased the content of GA1, GA3, and GA4 in roots as compared to SN (Figure 3a). The GA1 content could not be detected in the roots under the LN condition, while the GA3 and GA4 levels were lower under LN condition than those under the SN condition. The expression levels of ent-kaurene synthase genes ZmKS2 and ZmKS4 in GA biosynthesis pathway were significantly downregulated by LN in comparison to SN (Figure 3b). Similarly, LN significantly decreased the expression levels of ZmGA20ox1 and ZmGA20ox4 (Figure 3c). LN enhanced the expression levels of catabolic genes ZmGA2ox1, ZmGA2ox5, ZmGA2ox6, ZmGA2ox7, and ZmGA2ox12 (Figure 3d). The GA3 contents, as well as the expression of genes encoding enzymes of GA metabolism, were also determined in the zmga3ox mutant. As shown in Figure S3, the levels of GA3 were less under LN condition than those under the SN condition ( Figure  S3a). LN significantly down-regulated the expression levels of GA biosynthetic enzyme genes ZmKS2, ZmKS4, ZmGA20ox1, and ZmGA20ox4 ( Figure S3b,c). Inversely, LN enhanced the expression levels of catabolic enzyme genes ZmGA2ox1, ZmGA2ox5, ZmGA2ox6, ZmGA2ox7, and ZmGA2ox12 ( Figure S3d). The results indicated that LN caused a reduction in the bioactive GAS levels in roots by regulating the transcripts encoding GA 20-oxidases and GA 2-oxidases.

GAs Involved in Manipulating the NO 3 − Uptake
In order to characterize the function of GAs in regulating the N uptake, the net NO 3 − fluxes were measured by using the NMT technique in wild-type and zmga3ox roots. As shown in Figure 4a, the uptake of NO 3 − was evidently decreased in zmga3ox root compared to wild-type under SN and LN conditions. The mean NO 3 − fluxes were lower by 266% and 23.8% in zmga3ox root than wild-type under SN and LN conditions, respectively ( Figure 4b). Moreover, LN repressed the 15 N content in both wild-type and zmga3ox roots. The amount of 15 N in zmga3ox roots was less than that in the wild-type roots under LN and SN conditions (Figure 4c). LN reduced the N content in root and shoot, and the zmga3ox plants had less N content than the wild-type plants under LN and SN conditions (Figure 4d). The N contents in zmga3ox root and shoot were less by 89.5% and 77.2% than those in wild-type under LN condition, and similar results were recorded for SN.

GAs Involved in Manipulating the NO3 − Uptake
In order to characterize the function of GAs in regulating the N uptake, the net NO3 − fluxes were measured by using the NMT technique in wild-type and zmga3ox roots. As shown in Figure 4a, the uptake of NO3 − was evidently decreased in zmga3ox root compared to wild-type under SN and LN conditions. The mean NO3 − fluxes were lower by 266% and 23.8% in zmga3ox root than wild-type under SN and LN conditions, respectively ( Figure 4b). Moreover, LN repressed the 15 N content in both wild-type and zmga3ox roots. The amount of 15 N in zmga3ox roots was less than that in the wildtype roots under LN and SN conditions (Figure 4c). LN reduced the N content in root and shoot, and the zmga3ox plants had less N content than the wild-type plants under LN and SN conditions ( Figure  4d). The N contents in zmga3ox root and shoot were less by 89.5% and 77.2% than those in wild-type under LN condition, and similar results were recorded for SN. In order to further investigate the role of GAs in the regulation of N uptake, exogenous GA3 and Ucz were applied. GA3 treatment could increase the amount of 15 N in wild-type and zmga3ox plants ( Figure 4e). On the other hand, Ucz treatment significantly decreased the amount of 15 N in wild-type and zmga3ox plants. The combined GA3 and Ucz treatments restored the amount of 15 N in Ucz-treated plants while decreased that in GA3-treated plants.
To clarify the effects of GA on N uptake whether derived from differences in plant growth and nutritional status, the dry weight and total N content in the wild-type and zmga3ox plants were analyzed under LN or SN condition. Although the dry weight of shoots and roots were lower in zmga3ox plants than the wild-type plants under both LN and SN conditions, the inhibition rate of LN to SN in dry biomass of shoot or root presented no significant difference between wild-type and zmga3ox plants ( Figure S4a-d). Similarly, the zmga3ox plants had lower total N content in shoots and roots compared to the wild-type plants. However, the inhibition rate of LN to SN in total N content of shoot or root showed significant difference between wild-type and zmga3ox plants ( Figure S4e-h).  In order to further investigate the role of GAs in the regulation of N uptake, exogenous GA 3 and Ucz were applied. GA 3 treatment could increase the amount of 15 N in wild-type and zmga3ox plants ( Figure 4e). On the other hand, Ucz treatment significantly decreased the amount of 15 N in wild-type and zmga3ox plants. The combined GA 3 and Ucz treatments restored the amount of 15 N in Ucz-treated plants while decreased that in GA 3 -treated plants.
To clarify the effects of GA on N uptake whether derived from differences in plant growth and nutritional status, the dry weight and total N content in the wild-type and zmga3ox plants were analyzed under LN or SN condition. Although the dry weight of shoots and roots were lower in zmga3ox plants than the wild-type plants under both LN and SN conditions, the inhibition rate of LN to SN in dry biomass of shoot or root presented no significant difference between wild-type and zmga3ox plants ( Figure S4a-d). Similarly, the zmga3ox plants had lower total N content in shoots and roots compared to the wild-type plants. However, the inhibition rate of LN to SN in total N content of shoot or root showed significant difference between wild-type and zmga3ox plants ( Figure S4e-h).

GAs Modulated the Transcript Expression of NO 3 − Uptake-Related Genes
In order to ascertain the mechanism of GA affecting the N uptake, the expression levels of NO 3 − uptake-related genes were detected in wild-type and zmga3ox roots subjected to SN and LN. As shown in Figure 5, the expression levels of ZmNRT2.1 and ZmNRT2.2 in zmga3ox roots were lower than those of wild-type roots under LN and SN conditions (Figure 5a,b). The transcript expression of ZmNRTs was also detected in the GA 3 -treated wild-type and zmga3ox plants. GA 3 treatment could significantly enhance the expression levels of ZmNRT2.1 and ZmNRT2.2 in wild-type and zmga3ox plants under LN and SN conditions. Meanwhile, LN downregulated the expression of ZmNPF6.3b in wild-type and zmga3ox roots, while the expression level of ZmNPF6.3a showed no significant difference between wild-type and zmga3ox roots (Figure 5c). The zmga3ox root had higher expression levels of ZmNPF6.3b than the wild-type root under the LN condition ( Figure 5d). In addition, the expression levels of ZmNPF6.3a and ZmNPF6.3b showed no significant difference in GA 3 -treated both wild-type and zmga3ox plants under both LN and SN conditions.

GAs Modulated the Transcript Expression of NO3 − Uptake-Related Genes
In order to ascertain the mechanism of GA affecting the N uptake, the expression levels of NO3 − uptake-related genes were detected in wild-type and zmga3ox roots subjected to SN and LN. As shown in Figure 5, the expression levels of ZmNRT2.1 and ZmNRT2.2 in zmga3ox roots were lower than those of wild-type roots under LN and SN conditions (Figure 5a,b). The transcript expression of ZmNRTs was also detected in the GA3-treated wild-type and zmga3ox plants. GA3 treatment could significantly enhance the expression levels of ZmNRT2.1 and ZmNRT2.2 in wild-type and zmga3ox plants under LN and SN conditions. Meanwhile, LN downregulated the expression of ZmNPF6.3b in wild-type and zmga3ox roots, while the expression level of ZmNPF6.3a showed no significant difference between wild-type and zmga3ox roots (Figure 5c). The zmga3ox root had higher expression levels of ZmNPF6.3b than the wild-type root under the LN condition (Figure 5d). In addition, the expression levels of ZmNPF6.3a and ZmNPF6.3b showed no significant difference in GA3-treated both wild-type and zmga3ox plants under both LN and SN conditions.

RNA-Sequencing Revealed Differentially Expressed Genes (DEGs) in the Wild-Type and zmga3ox Plants in Response to NO3 − Supply
In order to gain molecular insights into the roles of GAs in the regulation of N uptake, the RNAseq assay was conducted. This helped identify the differentially expressed genes (DEGs) in the wildtype and zmga3ox roots at 12, 60, and 108 h after the LN or SN treatment. The statistics of the clean reads

RNA-Sequencing Revealed Differentially Expressed Genes (DEGs) in the Wild-Type and zmga3ox Plants in Response to NO 3 − Supply
In order to gain molecular insights into the roles of GAs in the regulation of N uptake, the RNA-seq assay was conducted. This helped identify the differentially expressed genes (DEGs) in the wild-type and zmga3ox roots at 12, 60, and 108 h after the LN or SN treatment. The statistics of the clean reads in the RNA-seq are shown in Table S1. A total of 3054 DEGs were identified after filtering with a threshold of |log 2 FC| ≥ 1 under the FDR (false discovery rate) <0.05 (Table S2). The Venn diagrams showed the number of DEGs in different samples. A greater number of DEGs were downregulated compared to the number of upregulated DEGs under both the LN and SN conditions. Furthermore, a greater number of DEGs were identified under the SN treatment compared to the LN treatment. The samples collected after 108 h showed the maximum number of DEGs under the LN as well as SN conditions (Figure 6a). Besides, the heat map suggested that the level of change (fold change) in the expression of these DEGs was higher after 60 and 108 h than that after the 12 h treatment ( Figure S5).
in the RNA-seq are shown in Table S1. A total of 3054 DEGs were identified after filtering with a threshold of |log2FC| ≥ 1 under the FDR (false discovery rate) <0.05 (Table S2) (Figure 6a). Besides, the heat map suggested that the level of change (fold change) in the expression of these DEGs was higher after 60 and 108 h than that after the 12 h treatment ( Figure S5). The GO (Gene Ontology) database and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis were employed to investigate the function of the 3054 DEGs and to understand the regulatory pathways that respond to the N supply in the zmga3ox mutant (Tables S3 and S4). The biosynthesis of phenylpropanoid and secondary metabolites was highly enriched in these DEGs (Figure 6b and Table S4). Moreover, the KEGG analysis revealed that the most significantly enriched pathway was the N-metabolism pathway, and this included the genes related to the NO3 − transporter, NO3 − reductase, and glutamate synthase (Figure 6b and  (Figure 6c and Table S5). Also, two ZmNRT2s genes were identified and these ZmNRT2.1 and ZmNRT2.2 genes were downregulated. Moreover, the ZmNRT3.1a and ZmNRT3.1b genes also showed downregulated expression (Figure 6c and Table S5  The GO (Gene Ontology) database and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis were employed to investigate the function of the 3054 DEGs and to understand the regulatory pathways that respond to the N supply in the zmga3ox mutant (Tables S3 and S4). The biosynthesis of phenylpropanoid and secondary metabolites was highly enriched in these DEGs (Figure 6b and Table S4). Moreover, the KEGG analysis revealed that the most significantly enriched pathway was the N-metabolism pathway, and this included the genes related to the NO 3 − transporter, NO 3 − reductase, and glutamate synthase (Figure 6b and  (Figure 6c and Table S5). Also, two ZmNRT2s genes were identified and these ZmNRT2.1 and ZmNRT2.2 genes were downregulated. Moreover, the ZmNRT3.1a and ZmNRT3.1b genes also showed downregulated expression (Figure 6c and Table S5). The other eight important DEGs were identified as the N assimilation-related genes, including ZmNIA1, ZmNIA2, ZmGln2, ZmGln1, ZmNIR1, ZmNIA3, ZmGOGAT1 and the NH 4 + transporter gene ZmAMT1.1b ( Figure 6d and Table S5); most of these important DEGs were downregulated in the zmga3ox plants.
In order to further investigate the role of GAs in the transcript regulation of N uptake and assimilation, the transcripts of the important DEGs involved in this process were evaluated by the RT-qPCR analysis. As expected, all the selected genes had similar expression patterns in the RNA-seq results (Figure 7 and Table S5). LN decreased the expression levels of ZmNRT2.1 and ZmNRT2.2 compared to SN at 60 and 108 h after treatment, and the expression levels of ZmNRT2.1 and ZmNRT2.2 in the zmga3ox plant were lower than those in the wild-type plant, except for 12 h after LN treatment (Figure 7a,b). Similar trends were observed in the expression of ZmNPF6.4 under SN condition (Figure 7c). However, the expression levels of ZmNPF7.3a were upregulated by LN compared to that by SN, and the zmga3ox plant showed lower expression levels of ZmNPF7.3a than the wild-type plant after 60 and 108 h of the LN or SN treatments (Figure 7d). Inversely, LN significantly downregulated the expression levels of ZmNPF6.3b compared to SN, and the expression levels of ZmNPF6.3b in the zmga3ox plant were higher than those in the wild-type plant after 60 and 108 h of the LN treatment ( Figure 7e). Meanwhile, the LN treatment downregulated the expression levels of ZmGOGAT1 and ZmNIR1 compared to the SN treatment, and the expression levels of ZmGOGAT1 and ZmNIR1 in the zmga3ox plant were lower than those in the wild-type plant at 60 and 108 h after the LN and SN treatments ( Figure S6a,b). Similarly, the expression of ZmGln1 and ZmNIA1 was repressed at 60 and 108 h after the LN treatment, while the zmga3ox plant showed downregulated trend of ZmGln1 and ZmNIA1 than the wild-type plant ( Figure S6c,d). Correspondingly, LN reduced the activities of nitrate reductase (NR) and glutamine synthase (GS); the activities of NR and GS in the zmga3ox plant were significantly lower than those in the wild-type plant ( Figure S7).

Transcription Factors Involved in GA-Mediated NO3 − Uptake
The transcriptional control of NO3 − uptake is well-documented over the past decade, in which transcription factors (TFs) play an important regulatory role [30]. The RNA-seq analysis revealed that a total of 287 important DEGs were identified as TFs (Table S6). To predict the targets of transcriptional regulation, the TF-binding sites were predicted from the putative promoter sequences (2 kb upstream from the transcriptional start site) of the nine identified nitrate transporter genes. As shown in Figure  8a and Table S7, 25 TFs were predicted to be involved in the regulation of nitrate transporter genes. The number of binding sites of the MYB, WRKY, and ERF families was the greatest, followed by the LBD, TCP, and bZIP family TFs.

Transcription Factors Involved in GA-Mediated NO 3 − Uptake
The transcriptional control of NO 3 − uptake is well-documented over the past decade, in which transcription factors (TFs) play an important regulatory role [30]. The RNA-seq analysis revealed that a total of 287 important DEGs were identified as TFs (Table S6). To predict the targets of transcriptional regulation, the TF-binding sites were predicted from the putative promoter sequences (2 kb upstream from the transcriptional start site) of the nine identified nitrate transporter genes. As shown in Figure 8a and Table S7, 25 TFs were predicted to be involved in the regulation of nitrate transporter genes. The number of binding sites of the MYB, WRKY, and ERF families was the greatest, followed by the LBD, TCP, and bZIP family TFs. In order to investigate the potential correlations between the DEGs, a weighted-genes coexpression network analysis (WGCNA) was performed with the 3054 DEGs. As shown in Figure 8b, the genes related to nitrate transport were clustered along with the TFs displaying a predicted potential interaction into different colored modules; the genes with the same color suggest a strong co-expression relationship. For instance, ZmNRT2.1 and ZmNRT2.2 strongly co-expressed with ZmMYB14/24/37/48, ZmEREB93, ZmWRKY22, ZmHSF18, and ZmHB107. Thereafter, ZmTCP33, ZmLBD24, ZmbZIP160, ZmWRKY34, ZmERF98, and ZmMYB41 were selected for the RT-qPCR assay. The expression patterns of these genes showed a similar trend with the RNA-seq data. On the contrary, the expression levels of these TFs showed a downregulated trend in the zmga3ox plants under the LN and SN conditions, except for ZmbZIP160, which showed a marked upregulation ( Figure 9). These results suggest that the GA-mediated N-uptake could occur by the transcriptional regulation of the expression of the ZmNRTs.

ZmbZIP160
ZmGATA32 In order to investigate the potential correlations between the DEGs, a weighted-genes co-expression network analysis (WGCNA) was performed with the 3054 DEGs. As shown in Figure 8b, the genes related to nitrate transport were clustered along with the TFs displaying a predicted potential interaction into different colored modules; the genes with the same color suggest a strong co-expression relationship. For instance, ZmNRT2.1 and ZmNRT2.2 strongly co-expressed with ZmMYB14/24/37/48, ZmEREB93, ZmWRKY22, ZmHSF18, and ZmHB107. Thereafter, ZmTCP33, ZmLBD24, ZmbZIP160, ZmWRKY34, ZmERF98, and ZmMYB41 were selected for the RT-qPCR assay. The expression patterns of these genes showed a similar trend with the RNA-seq data. On the contrary, the expression levels of these TFs showed a downregulated trend in the zmga3ox plants under the LN and SN conditions, except for ZmbZIP160, which showed a marked upregulation (Figure 9). These results suggest that the GA-mediated N-uptake could occur by the transcriptional regulation of the expression of the ZmNRTs.

Discussion
The manipulation of the GA biosynthetic and signaling genes can modulate the plant stature and lead to a "green revolution" in the cereal crops [7,8]. However, the dwarfism contributed by the GA-related mutants occurs a lower NUE in rice and wheat production [6,18]. Actually, the mutant sd1 rice plants show a lower NH4 + uptake rate than the wild-type plants, and a lower 15 NO3 − uptake is also observed in the mutant Rht-B1b wheat plants [10]. In the present study, a mutant maize zmga3ox was constructed and these plants showed the dwarfed phenotype, while the addition of exogenous GA3 could restore normal growth (Figure 1). Moreover, the mutant zmga3ox plants presented significantly lower NO3 − fluxes and 15 NO3 − uptake, which could be restored by exogenous GA3 (Figure 4). These results suggest that the N uptake was decreased by the GA deficiency in maize plant, and GA accumulation might affect the maize plant for the adaptation to the N supply in the soil. Although the zmga3ox plants had less dry weight of shoots and roots than the wild-type plants, GA deficiency could not change the inhibition rate of LN to SN in the plant growth compared to the wild-type plants

Discussion
The manipulation of the GA biosynthetic and signaling genes can modulate the plant stature and lead to a "green revolution" in the cereal crops [7,8]. However, the dwarfism contributed by the GA-related mutants occurs a lower NUE in rice and wheat production [6,18]. Actually, the mutant sd1 rice plants show a lower NH 4 + uptake rate than the wild-type plants, and a lower 15 NO 3 − uptake is also observed in the mutant Rht-B1b wheat plants [10]. In the present study, a mutant maize zmga3ox was constructed and these plants showed the dwarfed phenotype, while the addition of exogenous GA 3 could restore normal growth ( Figure 1). Moreover, the mutant zmga3ox plants presented significantly lower NO 3 − fluxes and 15 NO 3 − uptake, which could be restored by exogenous GA 3 (Figure 4). These results suggest that the N uptake was decreased by the GA deficiency in maize plant, and GA accumulation might affect the maize plant for the adaptation to the N supply in the soil. Although the zmga3ox plants had less dry weight of shoots and roots than the wild-type plants, GA deficiency could not change the inhibition rate of LN to SN in the plant growth compared to the wild-type plants ( Figure S4a-d). However, the inhibition rate of LN to SN in the N content was significantly modulated by GA deficiency (Figure S4e-h). These indicate that GA deficiency could not change the effects of NO 3 − -mediated plant growth, while it might have a significant effect on N uptake in maize exposed to LN or SN conditions. Anthocyanin accumulation is a typical characteristic of N starvation, and GAs regulate the N deficiency-induced anthocyanin accumulation in Arabidopsis or tomato [11,12,14]. Meanwhile, low anthocyanin content in leaves is observed under normal N or Pi condition, and GA 3 application exhibits little effects on anthocyanin accumulation. Similar results were observed in this study, where LN significantly induced anthocyanin accumulation in the zmga3ox plants, and this phenomenon could be weakened by GA 3 treatment (Figure 2). Furthermore, GA deficiency had a slight effect on anthocyanin accumulation under SN condition. These results suggest that GA signals are involved in altering the LN-induced anthocyanin accumulation in maize. Meanwhile, the zmga3ox leaves showed higher the photosynthetic rate and chlorophyll concentration than the wild-type leaves (Figure 2c and Figure S2a); however, the leaf areas of the zmga3ox plant were significantly less than those of the wild-type plant ( Figure S2b), which mainly caused low plant growth in the zmga3ox plant ( Figure S4a). Contrary to the anthocyanin accumulation, LN significantly decreased the chlorophyll content and N content in the zmga3ox plants. Thus, the zmga3ox plants presented a significant difference in N content compared to the wild-type plants ( Figure S4). Combined with these results, it was possible that NO 3 − supply modulated the N accumulation for affecting the anthocyanin accumulation and chlorophyll content in the zmga3ox leaves. In addition, the anthocyanin accumulation is closely associated with the activities of the antioxidant enzyme system in tobacco [31], and higher anthocyanin concentration can decrease ROS accumulation in potato leaves [32]. Here, the O 2 − content in zmga3ox leaves slightly lower than that in wild-type leaves, while it showed a little higher than the anthocyanin content under SN condition ( Figure S2c), which suggests that the anthocyanin accumulation might be involved in affecting the ROS accumulation in zmga3ox leaves. Further studies could research whether anthocyanin accumulation is involved in altering photosynthesis and ROS balance in zmga3ox leaves. The NO 3 − is the main form of available inorganic N that is utilized by most plants from the aerobic soils, taken up by the roots and transported to the shoots before assimilation; this process is mediated by the NO 3 − transporters [19]. To date, many NPF/NRT families in Arabidopsis and rice have been functionally identified for their involvement in the root NO 3 − uptake [33]. Moreover, the expression of AtNPF6.3 and AtNRT2.1 is regulated by auxin, ethylene, and cytokinin in plants in response to N supply [27]. However, the process by which the GAs regulate the expression of NRTs in plants combined with sufficient or insufficient N is still unclear. In maize, only ZmNPF6.3a and ZmNPF6.3b genes have been cloned, ZmNPF6.3b is a dual-affinity nitrate transporter, while ZmNPF6.3a displays a low-affinity nitrate transport activity [34]. In the present study, the expression of ZmNPF6.3a was not significantly different in the wild-type and zmga3ox roots. Correspondingly, GA 3 treatment could not alter the expression levels of ZmNPF6.3a and ZmNPF6.3b in wild-type and zmga3ox plants ( Figure 5). These indicate that GAs might have a little effect on the transcriptional regulation of ZmNPF6.3a and ZmNPF6.3b in maize responding to NO 3 − supply. Several studies have demonstrated that the ZmNRT6.3a and OsNRT1.1a display NO 3 − transport activity in maize and rice, and overexpression of OsNRT1.1a in rice greatly improved N uptake, but the expression levels of ZmNRT6.3a and OsNRT1.1a are not regulated by N supply [34,35]. Meanwhile, the post-translational modifications of NRT1s, such as phosphorylation and ubiquitination, are also essential for NO 3 − uptake or transport [36,37].
Here, the expression levels of ZmNPF6.3a and ZmNPF6.3b in the zmga3ox plant had no significant difference with those of the wild-type plants under SN condition (Figure 5c,d), but the zmga3ox plant had lower the N content compared to the wild-type plants (Figure 4c). These indicate whether GA could be involved in regulating the post-translational modifications of ZmNRT6.3s for altering the LATS-mediated NO 3 − uptake, which would be further studied.
The NRT2.1 is the major HATS-type gene involved in the root NO 3 − uptake in Arabidopsis [25,26].
The AtNRT2.1/NRT2.2 double-knockout mutations result in up to 80% loss of the NO 3 − -inducible HATS (iHATS) activity and show severe growth restriction in the absence of the sole NO 3 − source [38].
Both the constitutive HATS (cHATS) and iHATS activities were impaired under the LN condition in the knock-outs of NRT2.1 in cucumber [39]. In the present study, the zmga3ox roots showed lower expression levels of ZmNRT2.1 and ZmNRT2.2, and this phenomenon was also found by RNA-seq analysis, while the exogenous GA 3 could upregulate the expression of these genes. Moreover, the zmga3ox roots had lower expression levels of ZmNRT3.1a and ZmNRT3.1b, and the functionality of NRT2.1 relies on its interacting protein NRT3.1/NAR2.1, as observed in the studies on the rice plant [40]. Thus, the NO 3 − fluxes and 15 [42]. Similarly, the HvNRT2 transcript is upregulated by 20% to 30% in the barley plants exposed to 10 mM NO 3 − with tungstate (an inhibitor of nitrate reductase), while the NO 3 − influx is decreased by 50% [43]. In addition, the phosphorylation is crucial for NRT2 expression abundance in response to NO 3 − supply [44,45]. These indicate that the post-transcriptional modifications on iHATS-mediated NO 3 − might be exerted in the plants exposed to NO 3 − supply. In addition, the expression level of NRTs is not only regulated by the NO 3 − but also by the N metabolites, such as NH 4 + or amino acids [40,44,46]. In addition to N metabolites, photosynthate (e.g., sucrose) may also influence NRTs expression [47]. Here, the KEGG analysis revealed that the most significantly enriched pathway included the N-metabolism and sucrose metabolism pathways (Figure 6b). These indicate that GAs significantly affect the N-metabolism and sucrose metabolism pathways in maize plants responded to NO 3 − supply. Together with the findings of GAs modulating the NO 3 − uptake, GAs could regulate the transcript expression of ZmNRT2.1 and ZmNRT2.2 in maize in response to N supply, and might also be involved in modulating the iHATS-mediated NO 3 − uptake, which will be further investigated.
The GAs also positively regulate other nutrient-acquisition related genes, such as the Pi transporter genes SlPT2 and SlPT7 [12], the K transporter gene AtHAK5 [13], and the iron-uptake regulated genes AtIRT1 (iron-regulated transporter) and AtFRO2 (iron-regulated ferric chelate reductase) [48]. Moreover, the DELLA proteins interact with the transcription factor AtFIT and AtbHLH38/39 for modulating FIT-regulated iron-uptake regulated genes [48]. Furthermore, the transcription factor OsGRF4 drives the expression of OsNRT2.3 and OsGS1.2 in spite of being suppressed by the DELLA protein SLR1 in rice [10]. In the present study, the RNA-seq analysis revealed the presence of 287 TF genes responding to GA in the maize roots under the LN and SN conditions (Table S6). Both the binding-site prediction and co-expression analyses suggested that the ERF, MYB, WRKY, TCP, bZIP, LBD, Dof, HsF, HB, and GAGT family TFs might govern the expression of nitrate transporter genes (Figure 8). In Arabidopsis, several TFs governing the NRT2.1 and NRT1.1 expression have been identified, including the transcription factor LBD family genes (LBD37/38/39), NLP family genes (NLP7), TCP family genes (TCP20), SBP-box family genes (SPL9), and bZIP family genes (TGA1/4) [30]. Among these TFs, the transcript expressions of ZmTCP33 and ZmLBD24 were significantly repressed in the zmga3ox plant under the LN and SN conditions (Figure 9a,b). The DELLA proteins could interact with class I TCP factors and block the TCP function by binding to their DNA-recognition domains in Arabidopsis [49]. However, the transcript expression of ZmbZIP160 was upregulated in the zmga3ox plant (Figure 9c). More interestingly, the number of MYB, WRKY, and ERF family binding sites was enriched to the greatest extent, and the co-expression analysis indicated that ZmNRT2.1 and ZmNRT2.2 co-expressed with ZmMYB14/24/37/48, ZmEREB93, and ZmWRKY22. This aspect needs further exploration; therefore, subsequent research would be conducted to explore the transcriptional regulation of GAs modulating the expression of ZmNRT2.1 or ZmNRT1.1 in N uptake of maize.

Plant Materials, Growth Conditions, and Treatment
The CRISPR/Cas9 vector used to produce the zmga3ox (GRMZM2G036340) was obtained from the Maize Functional Genomic Project of China Agricultural University. The lines were produced in the maize (Zea mays L.) inbred line B73-329 background. For hydroponic experiments, the seeds were surface-sterilized in a 10% (v/v) H 2 O 2 solution for 20 min and washed 5 times with distilled water. After that, the seeds germinated on the sand in a growth chamber, at 28/22 • C with a 16/8 h light/dark cycle, relative humidity controlled to approximately 70-80%. After germinated on the sand for 7 d, the uniform seedlings with two visible leaves were transferred to a hydroponic box (12 seedlings per pot) containing 5 L nutrient solution after the endosperm of each seedling was removed. 5 µm above the root surface) were determined by moving the flux microsensor between two positions (30 µm in distance) in the direction perpendicular to the root axis. The recording rate of ion flux was one reading per 6 s, and each measurement point was monitored for 10 min. The N uptake assay in roots of the zmga3ox seedling was performed as described previously [10]. After LN or SN treatment for 5 d, the uniform seedlings were selected and rinsed the roots with 0.1 mM CaSO 4 solution for 1 min, and then incubated in the nutrient solution containing 0.05 or 2 mM K 15 NO 3 with a 99% atom excess of 15 N, respectively, for 10min. After that, washed for 1 min in 0.1 mM CaSO 4 solution before sampling. The samples were harvested and dried at 120 • C for 30 min, and then 65 • C for 72 h before being ground. The power was used for total 15 N determination by isotope ratio mass spectrometry (Vario PYRO cube ISOprime 100, Cheadle Hulme, UK).
The roots and shoots of maize seedlings at 5 d after LN or SN treatment were separated for assay of N concentration following the Kjeldahl method [52]. Plant total N content was calculated as the product of N concentration and corresponding dry weight.

RNA Isolation and Reverse Transcription Quantitative Polymerase Chain Reaction (RT-qPCR) Analysis
Total RNA was isolated from each sample using the Plant RNeasy Mini kit (Tiangen, Beijing, China). Then the full-length cDNA was synthesized from 2 ug of RNA using Oligo d (T) primer and M-MLV reverse transcriptase (Takara, kusatsu, Japan). Quantitive PCR was conducted in an Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, CA, USA) using SYBR ® Premix Ex Taq™ (Takara, Japan) following the manufacturer's instructions. ZmUBC (ubiquitin C) was used to normalize gene expression [53]. Fold change of gene expression values were calculated using the 2 −∆∆Ct method [54]. The primers for qRT-PCR are listed in Table S8.

RNA-Seq Analysis
Three biological RNA replicates were obtained from the root of wild-type and zmga30x plants at 12,60, and 108 h after LN or SN treatment, each biological replicate contained roots from three plants. Library construction for transcriptome sequencing was performed using the NEBNext UltraTM RNA Library Prep Kit for Illumina (NEB, Ipswich, MA, USA). The clustering of the index-coded samples was performed on a cBot Cluster Generation System using a TruSeq PE Cluster Kit v4-cBot-HS (Illumia) according to the manufacturer's instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq 4000 platform, and paired-end 150 bp reads were generated. Clean data (clean reads) were obtained by removing reads with adapter reads containing ploy-N and low quality reads from raw data. At the same time, Q20, Q30, GC-content, and sequence duplication levels of the clean data were calculated. All the downstream analyses were based on clean data with high quality. The filtered reads from each sample were aligned to the maize reference genome (B73 RefGen_v3, http://www.maizegdb.org/assembly/) using TopHat2 [55]. Gene expression levels were estimated by fragments per kilobase of transcript per million reads (FPKM) to compare among different samples. EdgeR software package [56] was used to detect genes differentially expressed between wild-type and zmga3ox mutant. K-means clustering analysis was conducted by cluster 3.0 between LN and SN at three time points and visualized by using Java TreeView. A gene was regarded as significantly differentially expressed if the false discovery rate (FDR) adjusted to a p-value < 0.05 and log 2 (fold change) ≥1.
Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotation and enrichment analyses of DEGs were performed by the online agriGO software [57] and KOBAS 3.0 software [58], respectively.

Weighted Gene co-Expression Network Analysis (WGCNA)
The FPKM of DEGs was adopted to process the raw files as the input file for WGCNA. The expression data from all three time points were analyzed combined. The co-expression networks were constructed with R package WGCNA (version 1.66) [59]. The soft thresholding power was 14, minModuleSize was 20, deepSplit was 2, networkType was signed, and others were default settings following previous descriptions [60]. The co-expression network was visualized with Cytoscape 3.7.1 [61].

Transcription Factor-binding Site Prediction
Putative promoter sequences (2 kb upstream from the transcriptional start site) of DEGS were downloaded from Ensemble Plants (http://plants.ensembl.org). Then the TF-binding site prediction was performed with online software PlantRegMap (http://plantregmap.cbi.pku.edu.cn), according to Jin et al. [62].

GAs Concentration Analysis
The roots or shoots of wild-type and zmga3ox seedlings were collected at 5 d after LN or SN treatment. The endogenous concentration of GAs was determine by HPLC-MS/MS. Briefly, approximately 500 mg of tissue was ground in liquid nitrogen and then extracted with 5 mL of 90% aqueous MeOH. Simultaneously, 2 ng of each D-labeled gibberellin compound was added to each sample. The subsequent operation was performed as described by [63].

Assay of Physiological and Biochemical Properties
After LN treatment for 5 d, 0.2 g fresh leaves or roots of wild-type and zmga3ox seedlings were collected for analyzing the activities of GS and NR. Chlorophyll content was measured following Wu et al. [64]. The anthocyanin content was determined, as described previously [65]. The activities of GS and NR were determined, respectively, with the Glutamine Synthetase and Nitrate Reductase Kit (Solarbio LIFE SCIENCES, BC0910/ BC0080, Beijing, China) following the manufacturer's instructions. The O 2 − production rate was determined with the protocol described by Chen et al. [66].
The photosynthetic rate was assessed by using an LI-6400 XT portable photosynthetic system (LI-COR, Lincoln, NE, USA). Leaf area was calculated by using a Li-3000C portable leaf meter.

Statistical Analysis
The data were statistically analyzed using SAS 9.0 (SAS Institute Inc., Cary, NC, USA). The Student's t-test was used for comparisons between two groups of data. For the data sets of more than two groups, one-way ANOVA with LSD (p < 0.05) was used.

Conclusions
Taken together, the results of the present study indicate that GA plays a significant role in the regulation of nitrogen uptake in the maize plant in response to N supply. LN significantly repressed the level of the bioactive GAs in the roots. Moreover, the shoots of the zmga3ox plants showed more sensitivity toward the LN stress, corresponding to the contribution of anthocyanin accumulation and the decrease of chlorophyll content. Also, the zmga3ox plants had low net NO 3 − fluxes leading to lesser 15 N content compared to the wild-type plants. The transcript expression of ZmNRTs was downregulated in the zmga3ox roots under the LN and SN conditions. These results suggested that GAs regulated N uptake involved in transcriptional regulation of NRTs expression and the physiological responses in maize responding to nitrogen availability. This research thus provides a valuable theoretical basis for improving N efficiency in crop production.