Transcriptome analysis reveals coordinated spatiotemporal regulation of hemoglobin and nitrate reductase in response to nitrate in maize roots


Author for correspondence:
Silvia Quaggiotti
Tel: +39 049 8272914


  • Given the importance of nitrogen for plant growth and the environmental costs of intense fertilization, an understanding of the molecular mechanisms underlying the root adaptation to nitrogen fluctuations is a primary goal for the development of biotechnological tools for sustainable agriculture. This research aimed to identify the molecular factors involved in the response of maize roots to nitrate.
  • cDNA-amplified fragment length polymorphism was exploited for comprehensive transcript profiling of maize (Zea mays) seedling roots grown with varied nitrate availabilities; 336 primer combinations were tested and 661 differentially regulated transcripts were identified. The expression of selected genes was studied in depth through quantitative real-time polymerase chain reaction and in situ hybridization.
  • Over 50% of the genes identified responded to prolonged nitrate starvation and a few were identified as putatively involved in the early nitrate signaling mechanisms. Real-time results and in situ localization analyses demonstrated co-regulated transcriptional patterns in root epidermal cells for genes putatively involved in nitric oxide synthesis/scavenging.
  • Our findings, in addition to strengthening already known mechanisms, revealed the existence of a new complex signaling framework in which brassinosteroids (BRI1), the module MKK2–MAPK6 and the fine regulation of nitric oxide homeostasis via the co-expression of synthetic (nitrate reductase) and scavenging (hemoglobin) components may play key functions in maize responses to nitrate.


The intensification of agricultural systems has been one of the major global changes in the 20th century. The use of artificial nitrogenous fertilizers has played a central role in this, but environmental concerns have emerged about the use of a large quantity of synthesizing ammonia products in both developing and developed countries. The Haber–Bosch process produces a total amount of reactive nitrogen (Nr) species of c. 100 Mt yr−1 (in 2000), of which > 80% is used for fertilizer production (Galloway et al., 2003).

Over 50% of the applied nitrogen is lost from the plant–soil system (Peoples et al., 1995), leading to environmental damage and to negative impacts on human health (Camargo & Alonso, 2006). In addition, the use of fertilizers with a high nitrogen concentration has been paralleled by the selection of genotypes giving acceptable yields only when grown at high nitrogen availability, but also characterized by a low efficiency of uptake and use of the nutrient (Quaggiotti et al., 2003; Hirel et al., 2007).

Maize (Zea mays) is one of the world’s major crops and is also expected to make an important contribution to human nutrition in the next century (Hirel et al., 2007). Enhancing the maize nitrogen use efficiency (NUE) may contribute to both yield improvement and to a reduction in the environmental hazards resulting from the excessive application of nitrogen fertilizers, even considering that its NUE is estimated to be well below 50% (Raun & Johnson, 1999; Zhu, 2000). Furthermore, the selection of cultivars more adaptable to low nutritional inputs is a fundamental goal in order to both diminish the cost of intensive agriculture in developed countries and to stimulate low-input sustainable agriculture in developing countries.

The genetic mechanisms regulating NUE are far from completely clarified and further studies are necessary to identify NUE components and to better investigate the signaling events underlying the global plant response to nitrogen availability. NUE depends on external factors, such as soil type and management, interactions with microorganisms, the type of nitrogen source and climate (Moll et al., 1982), but also on the intrinsic physiological efficiency of nitrogen uptake, assimilation and translocation, including the whole regulation and signaling system within the plant (Hirel et al., 2007).

Plants take up and assimilate both nitrate and ammonium, but nitrate is the main source of inorganic nitrogen for plants in aerobic soil conditions and its availability in soil solution may fluctuate strongly, depending on environmental conditions.

In addition to its role as a nutrient, nitrate acts as a signaling molecule regulating the expression of the genes involved in growth and developmental processes (Palenchar et al., 2004; Wang et al., 2004). However, the molecular aspects related to nitrate sensing and signaling pathways have only been partially unraveled.

Gene regulation has long been known to be a key step in plant adaptation to fluctuating nitrogen environments (Krouk et al., 2010). In the last 10 yr, a number of genome-wide expression studies have been performed on nitrate responses, mainly in Arabidopsis (Wang et al., 2000, 2003, 2004; Palenchar et al., 2004; Scheible et al., 2004; Gutiérrez et al., 2007b; Peng et al., 2007b; Gifford et al., 2008; Krouk et al., 2009), leading to the conclusion that nitrate control of gene expression can be dramatically context dependent as only c. 300 genes have been found to be consistently nitrate regulated in a meta-analysis of these transcriptome sets (Gutiérrez et al., 2007a). This context/environment dependence of nitrate transduction pathways may have slowed the discovery of the signaling and molecular mechanisms through which plants adapt to nitrogen fluctuations in soil (Krouk et al., 2010).

Nevertheless, the discovery of target genes and the characterization of their role in NUE continue to represent an ambitious challenge for plant scientists and a goal for the achievement of low-input crops.

In this study, a wide-spectrum approach was applied to maize roots to identify the key genes putatively involved in the response to nitrate availability. With this aim, 661 maize genes, differentially transcribed by nitrate supply or nitrate deprivation, were identified by means of cDNA-amplified fragment length polymorphism (AFLP). Most transcripts showed an altered expression in response to a prolonged (days) nitrate supply or deprivation, whereas fewer were up-/down-regulated by a shorter period (hours), thus being more likely to be involved in earlier signaling events. Quantitative real-time PCR (qPCR) was performed to validate the AFLP profiles and to better depict the expression profiles of a selected set of genes. In addition, in situ localization of mRNA was carried out for five of these genes. The results obtained allowed us to better characterize the molecular events involved in the maize root response to nitrate fluctuations, and their potential physiological implications are discussed in relation to the existing literature.

Materials and Methods

Plant material and experimental design

Seedlings of maize (Zea mays L.) hybrid 5783, supplied by DEKALB (Monsanto, Italy), were sown, grown and supplied with 1 mM KNO3 as described by Quaggiotti et al. (2003) and Trevisan et al. (2008). The nitrate was supplied at a concentration of 1 mM. In the nitrogen-depleted nutrient solution, KNO3 was replaced by 1 mM KCl. After 5 d, plants were transferred for 6 h into a different nutrient solution, according to the plot reported in Fig. 1. Root samples from each treatment were collected 6 h after the transfer. For reverse transcription-polymerase chain reaction (RT-PCR) analysis, additional time points, ranging from 0 to 8 h, were also included. Tissues used for gene expression analyses were collected and immediately frozen in liquid nitrogen and kept at −80°C for subsequent RNA extraction.

Figure 1.

Schematic representation of the experimental plan: 3 d after sowing, maize (Zea mays) seedlings were collected into four groups and transferred to 2-l tanks containing different aerated nutrient solutions, according to the presence or absence of a nitrogen source. After 5 d, plants were transferred for 6 h into a different nutrient solution, according to the treatment (1, 2, 3 and 4). Treatment 1 corresponds to the positive control (long-term 1 mM nitrate (NO3)-supplied seedlings) and treatment 3 is the negative control (long-term nitrate-depleted seedlings). The comparison between treatment 2 and treatment 1 (6 h of nitrate depletion after 5 d of 1 mM nitrate supply) enables the detection of the short-term NO3 deficiency effects. Treatment 3 vs 4 (6 h of 1 mM nitrate supply after 5 d of nitrate depletion) shows the short-term NO3 supply effects. Treatment 3 vs 1 (long-term 1 mM nitrate supply vs long-term nitrate shortage) shows the long-term NO3 deprivation effects. The NO3 concentration was selected according to our previous findings (Quaggiotti et al., 2003; Trevisan et al., 2008), demonstrating that 1 mM is a sufficient concentration to induce a plant response to NO3. For all experiments, each treatment was started at 09:00 h in order to complete it before 15:00 h (6-h treatments).

Twenty seedlings were used per sample in each experiment, which was conducted in triplicate for cDNA-AFLP and in six independent biological replicates for qPCR.

RNA extraction and cDNA synthesis

Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol. RNA was quantified by spectrophotometric reading, and its quality was assayed by agarose gel electrophoresis. First-strand cDNA was synthesized from 1 μg of total RNA, using 200 U of MMLV Reverse Transcriptase (Promega, Milan, Italy) and oligodT as a primer, in 20-μl reactions, as described in Quaggiotti et al. (2003).

cDNA-AFLP, band purification and sequencing

Double-strand cDNA was used for cDNA-AFLP analyses as described previously (Botton et al., 2008). The analyses were conducted using 336 different combinations of EcoRI and MseI primers (Supporting Information Table S1). The autoradiogram films were scanned and analyzed with KODAK 1D v 3.6 software (Scientific Imaging Systems, Eastman Kodak Company, Rochester, New York, USA). Amplicons showing at least a three-fold difference in terms of intensity between the two samples in all the replicates were excised from the blots, purified as described previously (Quaggiotti et al., 2007; Botton et al., 2008), and used as template in a standard PCR with the same EcoRI and MseI primers as adopted in the respective cDNA-AFLP experiment. The re-amplified DNA was directly sequenced at BMR Genomics (University of Padova, as PCR-derived product purified by the EXO-SAP enzymatic system (Amersham Biosciences, GE Healthcare, Little Chalfont, Buckinghamshire, UK).

Bioinformatics and gene ontology (GO) analysis

The BlastN algorithm was adopted to retrieve the full-length coding sequences corresponding to the cDNA-AFLP-derived sequences, to be used as a query to extract the GO information from retrieved database matches. The whole analysis was performed with Blast2GO software (Conesa et al., 2005). InterProScan (Zdobnov & Apweiler, 2001) was performed by enabling all the possible motif database searches and the resulting GO terms added to the annotation. The annotation analysis was then implemented with the ‘Augment Annotation by ANNEX’ function. Basically, this approach uses univocal relationships between GO terms from the different GO categories to add implicit annotation (for a detailed description, see Myhre et al., 2006).

Results were visualized by count-based histograms using a sequence cut-off = 0 to include all the available annotations.


Relative quantification of transcripts by qPCR was performed in a StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) as described by Nonis et al. (2007). Experiments were conducted using SYBR Green chemistry (5Prime, Hamburg, Germany) according to the manufacturer’s instructions. For each reaction, 5 ng of retrotranscribed RNA were used as template. Three technical replicates were performed on six independent biological replicates using the following thermal cycling conditions: a first denaturation step at 95°C for 3 min, followed by 40 cycles (94°C, 15 s; 56°C, 20 s; 68°C 30 s). Gene-specific primers (Table S2) were designed with Primer3 software version 0.4.0 ( Melting curve analysis was performed to confirm the absence of multiple products and the formation of primer dimers. Data were exported and analyzed according to the Livak & Schmittgen (2001) method using GADPH (primers: forward, 5′-CCTGCTTCTCATGGATGGTT-3′; reverse, 5′-TGGTAGCAGGAAGGGAAACA-3′; Bolduc & Hake, 2009) and Actin (primers: forward, 5′-GATTCCTGGGATTGCCGAT-3′; reverse, 5′-TCTGCTGCTGAAAAGTGCTGAG-3′; Lin et al., 2009) as reference genes according to Vandesompele et al. (2002). For each transcript, the ratio between the expression measured for a given treatment and that of its own control was used to estimate the up- or down-regulation of genes. The ratios obtained were then expressed as the base-2 logarithm to build the graphs.

In situ hybridization (ISH)

Roots were fixed in 4% formaldehyde, dehydrated in a graded series of ethanol, infiltrated with paraffin (Paraplast X-tra; Sigma-Aldrich, St Louis, MO, USA) and sectioned (7 μm). Templates for probe synthesis were selected by PCR from the cDNA of maize roots, and sense and antisense riboprobes were transcribed in vitro using T7 and SP6 RNA polymerases (Roche, Basel, Switzerland) and labeled with digoxygenin (DIG) RNA labeling mix (Roche) following the manufacturer’s protocol. Primer sequences used for probe selection are listed in Table S2. Paraffin was removed in Histo Clear II (National Diagnostics, Atlanta, GA, USA) and sections were hydrated in a decreasing ethanol series. Sections were digested with 1 μg ml−1 proteinase K (Roche) at room temperature for 10 min, and the reaction was stopped by the addition of phosphate-buffered saline. Hybridization was conducted overnight in 50% formamide at 50°C. DIG detection and signal visualization were carried out with nitroblue tetrazolium/5-bromo-4-chloro-3′-indolylphosphate (NBT/BCiP) (Roche) following the manufacturer’s instructions. After staining, slides were observed with an Olympus BX50 microscope (Olympus Corporation, Tokyo, Japan). Images were captured with an Axiocam Zeiss MRc5 color camera (Carl Zeiss, Oberkochen, Germany), and processed with Adobe Photoshop 6.0. In order to achieve the highest reproducibility and throughput, all liquid handling procedures were modified and adapted to be performed in a liquid handling robot (Freedom EVO100, Tecan, Männedorf, Switzerland). Using this liquid handling platform, tissue permeabilization, hybridization and washing steps until detection were performed in an automated manner on at least six slides (including at least five sections each) per probe concentration (at least two probe concentrations were tested for each hybridization combination) per experimental condition.


cDNA-AFLP analysis, identification of differentially expressed transcript-derived fragments (TDFs) and GO analysis

The cDNA-AFLP technique allowed the identification of an average of five differentially expressed transcripts for each of the 336 primer combinations. Among these, 661 TDFs showing the most interesting profiles were selected. A comparison of the effects of the four different treatments on the expression profiles allowed us to display the transcriptional responses of maize roots to both short- and long-term nitrate supply or deprivation (Fig. S1). Homologies for 457 TDFs have been found in public databases using BLASTn and tBLASTx applications, whereas none have been found for the remaining 204 sequences (a full list of BLAST results for the 661 TDFs isolated in this work is given in Table S3). The identified genes were subsequently classified, according to their expression patterns, into six groups (A, B, C, D, E or F), as represented in Fig. 2. Group A included transcripts (14% of the TDFs) either up- or down-regulated only in response to treatment 1 (5 d of NO3-) when compared with the three other nutritional conditions.

Figure 2.

Graphical representation of amplified fragment length polymorphism (AFLP) transcript-derived fragment (TDF) expression patterns in maize (Zea mays). Group A (transcripts either up- or down-regulated only in response to 5 d of nitrate (NO3) supply) represents the cluster of genes implicated in the long-term response to NO3 (treatment 1). Group B includes genes that exclusively respond to long-term NO3 deficiency (treatment 3). Transcripts that are regulated by short-term NO3 deficiency (treatment 2) are contained in group C, whereas group D includes the 9% of TDFs differentially regulated by treatment 3 (as observed for TDF belonging to group B), but unresponsive to a 6-h NO3 resupply (treatment 4). Genes responding to short-term NO3 presence are included in group E. Group F includes the genes that are regulated by both short- and long-term NO3 deficiency. White blocks represent TDFs differentially regulated (either up- or down-regulated) in the considered treatment, and gray blocks represent TDFs whose expression was not influenced by nitrate availability. The profile distribution percentages are also reported in the figure.

More than one-half of all selected TDFs (52.42%) belonged to group B, namely including the genes that exclusively responded to long-term NO3 deficiency (treatment 3), but that were readily regulated after a few hours of nitrate resupply (treatment 4). Such a high percentage of TDFs included in group B shows that long-term NO3 deficiency is probably the condition that more strongly affects the root transcriptome. Nevertheless, the response to short-term NO3 deficiency was mediated by the genes belonging to group C (c. 10%), which showed an altered expression profile only in response to treatment 2. Group D includes the TDFs (c. 9%) that were differentially regulated by long-term NO3 deficiency (treatment 3), as also observed for TDFs belonging to group B, but, unlike the latter, they did not respond to a 6-h NO3 resupply (treatment 4). Approximately 6% of all identified TDFs belonged to group E, which includes the few genes responding specifically to the short-term presence of NO3, thus being putatively involved in the earlier events of NO3 perception. Finally, group F includes the genes that were regulated by both short- and long-term NO3 deficiency. They represent c. 8% of all identified TDFs.

In order to identify the major biological processes involved in the response to nitrate, all sequences of the identified TDFs were used to automatically recover the related GO annotations by means of Blast2GO software (Conesa et al., 2005) (Fig. S2). Different levels of ontology were chosen according to the number of sequences described. With regard to the biological processes, most annotations analyzed were related to ‘cellular process’ (30%) and ‘metabolic process’ (30%). Among the molecular functions, the majority of TDFs were related to ‘binding’ (44%) and ‘catalytic activity’ (38%). Finally, with regard to the cellular component ontology, the majority of TDFs were related to the ‘cell part’ (38%) and ‘membrane-bound organelle’ (31%).

Gene expression analyses

The expression levels of 39 nitrate-regulated transcripts were further analyzed by qPCR to validate the cDNA-AFLP results and to better characterize their expression profiles. Genes were selected according to their AFLP profiles and presumed signaling role. Profiles detected by cDNA-AFLP were generally confirmed by means of normalized expression measured by qPCR. It should be pointed out that only a few genes showed a very strong transcriptional regulation in response to the treatments, whereas the majority displayed only slight alterations after nitrate supply/exclusion. However, all 39 transcripts showed reproducible expression profiles for each of the six biological replicates. The data shown in Table 1 and Figs 3–5 include the expression profiles of 28 of the overall 39 transcripts, which displayed increases or decreases of transcription of at least 20% in comparison with their own control. Data obtained for the remaining 11 genes, which showed very slight variations in expression level among treatments, are not included in the figures.

Table 1.   List of 39 transcript-derived fragments (TDFs) analyzed by quantitative real-time PCR
 IDGenBank accession numberAnnotation2 vs 14 vs 33 vs 1
  1. TDFs are progressively numbered (Gene Identification Number, ID) and their GenBank accession numbers, together with the associated annotations, are reported. The expression profiles of the 28 TDFs that displayed increases (+) or decreases (−) in transcription of at least 20% in comparison with their own control are described in the last three columns. TDFs were grouped according to their treatment responsiveness. Transcripts from 29 to 39 (in italic) showed an increase/decrease of expression lower than 20% and were not included in Figs 3 and 5.

I1AJ579383.1Glutathione synthetase (GS)  
2EU952447.1Pre-rRNA-processing protein TSR2  +
3EU956602.1OsWAK receptor-like protein kinase  
4EU977123.1Receptor-like protein kinase (RK20-1)  
5M60215.1Protein phosphatase-1 (PP1)  
6EU952181.1AKIN gamma  
7AF077372.1Cytochrome b5 reductase (NFR)  +
II8AY787873.1Zinc-finger protein (znf1)+  
10EU971385.1SNARE-interacting protein (KEULE)  
11EU956325.1Root border cell-specific protein  
12EU969809.1RWD domain-containing protein  
13EU962873.1Ubiquitin-protein ligase/zinc ion-binding protein  
14AF104924.2Unconventional myosin heavy chain (MYO1)  
IV15EU963548.1Farnesyltransferase/geranylgeranyltransferase type I alphasubunit+ 
16EU966165.1Corn cystatin I+ 
17EU959126.1Cell division control protein 2+ 
V18EU966069.1ASC1-like protein +
19AY109304.1Calcium-dependent protein kinase +
20EU965735.1Glycogen synthase kinase-3 MsK-3 mRNA 
VII22AF153448.1Nitrate reductase (NR)+
25BT065734.1MAPKK (1/2)+
26EU967163.1Peroxisomal multifunctional enzyme type 2++
27EU957250.1LysM receptor-like kinase++
28EU974732.1Stachyose synthase precursor++
VIII29EU956629.1Heat shock 70-kDa protein 4   
30EU957579.1Aldehyde dehydrogenase   
31X98245.1Annexin p35   
32AB007502.1Squalene synthase   
33AF144079.1l-Methionine S-methyltransferase   
34EU975516.1Calmodulin-binding protein   
35EU970228.1CCR4-NOT transcription complex subunit 7   
36AY301062.2Calcium-dependent protein kinase (CPK11)   
37EU962348.1Calcineurin B-like protein 3   
38AB016802.1MAP kinase 5   
39U82481.1Kinase interacting kinase1 (kik1)   
Figure 3.

Short-term effects of nitrogen starvation on the maize (Zea mays) transcriptome: cDNA-amplified fragment length polymorphism (AFLP) gene expression validation by quantitative real-time PCR (qPCR). The changes in transcript abundance are shown for 12 genes. Data were expressed as log2 ratios of the normalized expression levels (for details, please refer to the Materials and Methods section) measured at 6 h after nitrogen depletion (treatment 2) with respect to the control (treatment 1). The results are averages ± SE of six independent biological replicates, each performed in three technical repetitions.

Figure 4.

Short-term effects of nitrogen supply on the maize (Zea mays) transcriptome: validation of cDNA-amplified fragment length polymorphism (AFLP) gene expression by quantitative real-time PCR (qPCR). The changes in transcript abundance are shown for 19 genes. Data were expressed as log2 ratios of the normalized expression levels (for details, refer to the Materials and Methods section) measured in nitrogen-depleted seedlings 6 h after nitrogen supply (treatment 4) with respect to the nitrogen-depleted seedlings (treatment 3). The results are averages ± SE of six independent biological replicates, each performed in three technical repetitions.

Figure 5.

Long-term effects of nitrate (NO3) deficiency on the maize (Zea mays) transcriptome: validation of cDNA-amplified fragment length polymorphism (AFLP) gene expression by quantitative real-time PCR (qPCR). The changes in transcript abundance are shown for 18 genes. Data were expressed as log2 ratios of the normalized expression levels (for details, see the Materials and Methods section) measured in nitrogen-depleted seedlings (treatment 3) with respect to the control (treatment 1). The results are averages ± SE of six independent biological replicates, each performed in three technical repetitions.

Genes responding to short-term NO3 deficiency (treatment 2 vs treatment 1)  This section includes transcripts displaying an increased or decreased expression after 6 h of nitrate depletion, and represents candidates putatively involved in the response to short-term NO3 deficiency. Fig. 3 shows the increase/decrease in expression measured for 12 genes that were selected as early responsive to nitrate shortage. Seven (8, 15, 16, 17, 26, 27, 28) showed an increase in transcript amount ranging from 25% to 75%, whereas five (21–25) exhibited a repression after 6 h of nitrate deficiency, as also indicated in Table 1 (‘2 vs 1’ column). The most relevant change was measured for two genes, encoding a maize hemoglobin (HB, 23, AF236080.1) and a nitrate reductase (NR, 22, AF153448.1), which showed a six- and three-fold down-regulation of transcription respectively. Of the 12 genes, only one appeared to be exclusive to this cluster. It encoded a zinc finger protein (8, znf1, AY787873.1) and showed a 75% up-regulation of transcription when compared with the positive control.

Genes responding to short-term NO3 supply (treatment 4 vs treatment 3)  This cluster includes 19 maize root transcripts early responsive to nitrate provision (Fig. 4), six of which were induced (19, 22–25, 28) and the remaining 13 showed a down-regulation of expression (9–18, 20, 26, 27) in response to nitrate application. This assemblage corresponds to genes that are likely to be implicated in the early sensing of nitrate. The expression rate measured ranged from values only slightly (35%) different from those detected in nitrate-depleted roots (negative control, 3) to a 20- or 90-fold higher transcript accumulation, as in the case of the genes encoding NR (22) and HB (23), respectively. Six genes among the 19 (9–14) belonged exclusively to this group and all displayed a decreased expression after 6 h of nitrate provision.

Genes responding to long-term NO3 deprivation (treatment 3 vs treatment 1) Fig. 5 shows the expression profiles of 18 genes that were differentially regulated in response to prolonged nitrate deprivation (5 d, treatment 3). They are assumed to take part in the whole response to nitrogen deprivation in maize. Thirteen (1, 3–6, 19–25, 28) showed a decreased transcription in roots grown in a minus-nitrate solution when compared with the positive control, whereas only five (2, 7, 18, 26, 27) were up-regulated in response to treatment. The most significant changes in expression were detected for the down-regulated genes. These included BRI1 (21, BRASSINOSTEROID INSENSITIVE 1, EU954960.1) and mitogen-activated protein kinase 6 (24, MAPK6, putative MAPK, EU965114) for which the mRNA abundance did not exceed one-quarter of that detected in roots supplied with nitrate, and both NR (22) and HB (23), which displayed expression levels 16- and 30-fold lower than the positive control, respectively.

Seven genes (1–7), together with the 18 selected, resulted as belonging specifically to this cluster and to be largely (five to over seven times) down-regulated by a long-lasting lack of nitrate.

Time-course expression of HB- and NR-encoding genes  To better characterize the expression profiles of the genes encoding HB and NR, we analyzed their time course of transcription 30 min and 2, 4, 6 and 8 h after nitrate resupply (treatment 4 vs treatment 3). The data reported in Fig. 6(a,b) show a marked accumulation of transcripts after just 30 min of nitrate provision for both, with an increasing trend being observed, reaching a maximum after 6 h of treatment. After 8 h of nitrate provision, both genes were still significantly up-regulated.

Figure 6.

Time-course analysis of early hemoglobin (a) and nitrate reductase (b) gene induction transcription after 30 min and 2, 4, 6 and 8 h of nitrate (NO3) supply. Maize (Zea mays) seedlings were grown for 5 d in a nitrate-depleted solution (treatment 3) and then transferred to a nutrient solution supplied with nitrate, as described in the Materials and Methods section. For each time point shown, data were expressed as log2 ratios of the normalized expression levels measured in nitrogen-supplied seedlings (treatment 4) with respect to the control (treatment 3). The results are averages ± SE of three independent biological replicates.

Localization of nitrate-responsive genes by ISH

To investigate the involvement of specific cell types in nitrate-mediated transcriptional regulation, ISH experiments were conducted to localize the mRNAs of five selected nitrate-responsive genes in the apex and along the differentiated zone of primary root sections.

In order to reach the highest reproducibility of results in terms of spatial distribution and, at the same time, a semi-quantitative evaluation of expression levels, an automated ISH approach was set up based on liquid handling procedures. This enabled us to obtain highly reproducible signals not only for localization, but also for the visualization of different levels of specific mRNAs. Genes selected for ISH were those encoding HB (AF236080.1), NR (AF153448.1) and BRI1 (EU954960.1), which showed significant differences in transcript accumulation, as quantified by qPCR, and peroxisomal multifunctional enzyme type 2 (EU967163.1) and lysM receptor-like kinase (EU957250.1), which, conversely, showed only slight variations in expression.

Tissues were probed with either an antisense (Figs 7, 8) or a sense RNA probe, avoiding any nonspecific signal detection (data not shown).

Figure 7.

In situ hybridizations of hemoglobin (a–c), nitrate reductase (d–f), BRASSINOSTEROID INSENSITIVE 1 (BRI1) (g–i), peroxisomal multifunctional enzyme (j–l) and lysM receptor-like kinase (m–o) mRNAs in apices of maize (Zea mays) primary root. The first two columns (from left to right) show the probe localization in two different treatments. The third column represents a greater magnification, showing the detail of signal localization. cc, central cylinder; cx, cortex; ep, epidermis; pc, pericycle. Bars, 200 μm.

Figure 8.

In situ hybridizations of hemoglobin (a–e), nitrate reductase (f–j), BRASSINOSTEROID INSENSITIVE 1 (BRI1) (k–o), peroxisomal multifunctional enzyme 2 (p–t) and lysM receptor-like kinase (u–y) mRNAs in lateral root primordia (LRP) and lateral roots (LR) of maize (Zea mays). First two columns (from left to right) show the probe localization in two different treatments in LRP. The last two columns show the probe localization in two different treatments in LR. Central panels indicate greater magnification, showing the detail of signal localization in lateral root apices. Bars, 200 μm.

In the case of HB, mRNA was detected in the epidermis and cortex cells of the root apical meristem for both treatments analyzed (3 and 4), but appreciable quantitative differences were visualized between them. Indeed, the root apex and, most evidently, the root epidermal cells of seedlings supplied with nitrate for 6 h (treatment 4, Fig. 7b,c) always showed a significantly higher signal in comparison with nitrate-deprived roots (treatment 3, Fig. 7a), consistent with the qPCR results.

NR transcripts were localized in the root apex, in both the epidermis and central cylinder (Fig. 7d–f). In this case, however, rather than quantitative differences, the two treatments analyzed (3 and 4) showed a different localization of the transcripts. Indeed, in the case of roots of nitrate-resupplied plants (treatment 4, Fig. 7e,f), messages were also detected in the cells proximal to the vascular cylinder in the elongation zone of the primary root, whereas, in nitrate-depleted seedlings (treatment 3, Fig. 7d and data not shown), the signal was restricted to the differentiation zone.

Concerning BRI1, mRNA was preferentially localized in epidermal cells of the root apex, although a less intense signal (even undetectable in certain sections) was also observed in vascular bundles (Fig. 7h). Long-term exposure to nitrate (treatment 1, Fig. 7h,i) and prolonged nitrate deprivation (treatment 3, Fig. 7g) shared the same expression pattern, although root sections derived from nitrate-supplied seedlings (treatment 1) always showed a stronger signal.

The gene encoding peroxisomal multifunctional enzyme type 2 appeared to be expressed constitutively across the primary root apex (Fig. 7k), but seemed to accumulate more clearly in epidermal cells of control roots (treatment 1, Fig. 7k) than in nitrogen-deprived tissues (treatment 3, Fig. 7j). In the case of the gene encoding lysM receptor-like kinase, a lower and not very specific signal was detected in both positive (treatment 1) and negative (treatment 3) controls with no significant differences (treatments 1 and 3, Fig. 7m–o).

The same transcripts were localized in both emerging lateral root primordia (LRP), at early and late developmental stages (Fig. 8, left), and in mature lateral roots (Fig. 8, right) in plants grown in different nitrate conditions.

HB and NR mRNA localization was analyzed in both LRP (at early and late developmental stages) and in mature lateral roots in response to prolonged nitrate deprivation (treatment 3) and nitrate resupply (treatment 4) (Fig. 8a–e). HB transcripts showed a strong signal in LRP, with no marked differences found between the two considered treatments (Fig. 8a–c). Moreover, the accumulation of transcripts of this gene spread out across the cortex cells of the primary root. The gene encoding NR was detected only in mature LRP of nitrate-resupplied seedlings (treatment 4) (Fig. 8f), whereas prolonged nitrate deprivation did not give rise to any staining (Fig. 8g). mRNAs of both genes were also localized in the apices of lateral roots, with a linear increase in signal after treatment 4 (Fig. 8h–j).

In the case of the transcript localization of BRI1, peroxisomal multifunctional enzyme type 2 and lysM receptor-like kinase, long-term exposure (treatment 1) and prolonged nitrate deprivation (treatment 3) effects were analyzed. Long-term nitrate supply induced BRI1 mRNA accumulation in early-developed LPR (Fig. 8k). However, a weaker staining was also detectable in prolonged nitrate-deprived roots (Fig. 8l).

The localization of BRI1 transcripts was also investigated in mature lateral roots. In these structures, the signal was predominantly localized at the apices, always showing stronger staining in roots of seedlings supplied with nitrate than in deprived seedlings (Fig. 8, right).

With regard to lateral roots and LRP, peroxisomal multifunctional enzyme type 2 and lysM receptor kinase mRNAs were detected at very low levels, even undetectable in certain sections, and did not show evident differences in transcript accumulation between the treatments analyzed, in both LRP (treatment 1, Fig. 8p,u; treatment 3, Fig. 8q,v) and mature lateral roots (treatment 1, Fig. 8r,s,w,x; treatment 3, Fig. 8t,y). ISH data, in general, demonstrated, at least for some genes, that the apparently small variations in expression measured by qPCR on RNAs extracted from whole roots were a result of the highly localized transcription, restricted for these genes to specific tissue/cell types (e.g. epidermal cells, BR1, Fig. 7g–i).


Inorganic nitrogen is a vital nutrient for plants. Crop productivity strongly depends on intense fertilization programs throughout the world, with an annual nitrogen fertilizer consumption currently close to 80 × 1012 g N (Miller & Cramer, 2004).

Nonetheless, the excessive application of nitrogen fertilizers has a great impact on environmental safety and human health (Galloway, 1998), and has led to the selection of genotypes with a low efficiency of uptake and use of this nutrient (Hirel et al., 2001). Therefore, an understanding of the physiological and molecular basis of a plant’s response to nitrogen is fundamental in order to select cultivars better adapted to low nutritional inputs, and is a first step towards the development of biotechnological tools for sustainable agriculture.

In order to acclimatize to various nutritional soil conditions, plants have evolved different strategies of nitrogen acquisition, but, in general, in a typical agricultural soil, the main absorbed form is nitrate (Crawford & Glass, 1998), which, in addition to serving as a nutrient, also acts as a signal.

In this study, both short- and long-term responsive genes to the absence/presence of nitrate were identified. The majority (52%) of transcripts identified belonged to the group of genes that responded to prolonged nitrate deficiency, implying that long-term nitrate deficiency is probably the condition that mostly affects the maize root transcriptome.

With regard to the transcripts already differentially expressed after just 6 h of nitrate depletion and possibly involved in the early perception of nutritional stress, seven (group VII, Table 1) of the 12 genes belonging to this group were also recovered in the other two clusters, implying that they may play a more transverse function in nitrate response regulation. Conversely, three (group IV, Table 1) of the 12 fell into both the first (early responsive to nitrate deprivation) and second (early responsive to nitrate supply) clusters, showing an opposite regulation and, more precisely, being up-regulated in response to 6 h of nitrate depletion and down-regulated after the same time of nitrate supply. These included corn cystatin I, belonging to the phyto-cystatins, taking part in the defense against biotic and abiotic stresses (Pernas et al., 2000; Massonneau et al., 2005), and the farnesyltransferase/geranylgeranyltransferase type I alphasubunit homolog of Arabidopsis pluripetala (PLP, AT3G59380), which is involved in the prenylation of many proteins important for environmental signaling (Running et al., 2004) and in the regulation of nutrient allocation into actively growing tissues (Zhou et al., 1997), and a third gene encoding cell division control protein 2 (CDC2), the homolog of the Arabidopsis gene CDKA;1. CDKA;1 is the major driver of G2/M transition after its association with A-, B- and D-type cyclins (Gutierrez, 2009). The regulation of a gene involved in the cell cycle in response to nitrate is unsurprising, considering the plethora of effects exerted by nitrate on root development, as reported by Stitt (1999), Zhang et al. (2000, 2007), Forde (2002) and Walch-Liu et al. (2006). Only one transcript, missing in the other two clusters, was exclusively involved in the early response to nitrate removal. It encoded a zinc finger protein (znf1), belonging to the A20/AN1 zinc-finger domain-containing proteins, which are known to be involved in the stress response in plants and in the immune response in animals (Heyninck & Beyaert, 2005; Vij & Tyagi, 2008). The second group analyzed contained genes encoding proteins likely to be involved in early nitrate sensing. It contained predominantly transcripts showing a decreased expression after 6 h of nitrate provision, although the most notable profiles were those of two genes encoding a nonsymbiotic HB and NR, which showed nitrate-induced up-regulation of expression of 90- and 25-fold, respectively. Both of these fit into the group of seven (group VII) genes that seem to be more extensively implicated in the nitrate response, displaying a differential expression profile in each of the three comparisons made.

Six (group III, Table 1) of 19 are unique to this pool, thus being engaged specifically in the early signaling governing the root response to nitrate. Two encoded proteins were involved in cytoskeleton re-arrangement, that is the motor protein MYO1, belonging to the plant myosin XI family implicated in organelle transport, F-actin organization, and cell and plant growth (Peremyslov et al., 2011), and a cytokinesis-related Sec1 protein, KEULE, a key regulator of vesicle trafficking, capable of integrating a large number of intra- and/or intercellular signals (Waizenegger et al., 2000; Assaad et al., 2001). A gene encoding AN14, another A20/AN1 zinc-finger domain-containing protein, fits into this same group, further confirming the above hypothesized role of this gene family in the nitrogen response. Two additional maize transcripts specific to this subgroup encoded a ubiquitin-protein ligase/zinc ion-binding protein and an RWD (RING finger and WD repeat-containing protein) domain-containing protein (Carbia-Nagashima et al., 2007), respectively. Protein degradation by the proteasomal pathway is an important mechanism of plant adaptation to nitrogen stress (Vidal et al., 2010). Peng et al. (2007b) identified an E3 ubiquitin ligase NLA, that is involved in Arabidopsis adaptation to nitrogen limitation. The identification of a ubiquitin-protein ligase and of an RWD domain-containing protein among the transcripts turned off early by nitrate provides further evidence of the regulatory role of proteasome degradation via ubiquitination on the whole plant response to nitrogen scarcity. Finally, a gene encoding a root border cell-specific protein specific to the early response to nitrate was also identified. This is not surprising given that border cells are considered to be differentiated tissues of the root system that modulate the plant root environment by producing specific substances to be released into the rhizosphere (Brigham et al., 1995; Vicréet al., 2005).

The third cluster analyzed contained 18 transcripts showing differential expression in response to prolonged nitrate deficiency. Seven genes were restricted to this cluster, hence being involved in the response to a prolonged lack of nitrate. Four were slightly repressed by a 5-d nitrate depletion and encoded glutathione synthetase (GS), AKIN gamma, OsWAK receptor-like protein kinase and protein phosphatase-1 (PP1), respectively, all of which seem to be implicated in the plant response to abiotic and biotic stresses. Indeed, GS is the second enzyme in the biosynthesis of glutathione, a key molecule in response to environmental stresses (May et al., 1998; Noctor et al., 2002), whereas AKIN gamma and WAK111-OsWAK receptor-like protein kinases appear to play a role in both pathogen resistance and abiotic stress tolerance (Zhang et al., 2005; Gissot et al., 2006). Furthermore, PP1, a major member of the PPP family of serine/threonine protein phosphatases, seems to be involved in various processes, such as the regulation of gene expression, phytohormone signaling, regulation of ion channels and pathogen resistance (Smith & Walker, 1996; Luan, 2003).

Seven of the 18 genes in this cluster also fell into the other two groups, being, in all probability, transversally implicated in the plant’s adaptation to nitrate fluctuations. This subgroup included a lysM receptor-like kinase, homolog to the Arabidopsis chitin elicitor receptor kinase 1 (CERK1), playing functions in fungal resistance and, possibly, in a more general activation of a complex signaling cascade(s) leading to various defense responses (Miya et al., 2007; Wan et al., 2008). Among these, we also found the stachyose synthase precursor responsible for the synthesis of stachyose, which belongs to the raffinose family of oligosaccharides (RFOs, α1,6-galactosyl extensions of sucrose). Increases in stachyose and raffinose in plant tissues are associated with cold acclimation (Peters & Keller, 2009). The physiological role of the synthesis of RFOs in response to nitrate is still unknown and must be studied further in depth. Similarly, the function of peroxisomal multifunctional enzyme type 2, which showed a differential expression in both early and late responses to nitrate provision and shortage, is still under investigation. Type 1 and 2 peroxisomal multifunctional enzymes are implicated in fatty acid degradation via the β-oxidation cycle. Recent studies in the model plant Arabidopsis revealed novel roles for β-oxidation, for example in the synthesis of auxins and jasmonic acid (Poirier et al., 2006). This gene was selected for ISH localization, together with that encoding the lysM receptor-like kinase, which showed a similar trend of expression in response to both early and late nitrate supply and shortage. For these two transcripts, ISH revealed an almost constitutive expression in the root apical meristem, with a narrow signal decline in nitrate-depleted roots. In the differentiated root zone, transcript accumulation was clearly distinguished in LRP from early stages of development. Further studies would clarify their involvement in the plant general response to nitrate availability.

The most interesting profiles were observed for two genes encoding a maize NR and a nonsymbiotic HB, respectively. Both displayed a very strong increase in transcription after just 30 min of nitrate supply. Conversely, they were significantly down-regulated in roots after 6 h of nitrate depletion, but also after prolonged (5 d) starvation. ISH experiments fully confirmed the results obtained from qPCR. ISH localization showed a shared expression pattern for these two genes, with a clear marked transcript accumulation in epidermal cells and a detectable signal in central cylinder cells of the root apical meristem, LRP and mature lateral roots.

HBs are hemeproteins that reversibly bind molecular oxygen (O2) and were first identified as symbiotic HBs (sHb) in legume root nodules, whereas the discovery of nonsymbiotic HBs (nsHbs) is a more recent finding (as reviewed by Garrocho-Villegas et al., 2007). In addition to O2 transport, nsHbs seem to play a key role in nitric oxide (NO) scavenging (Perazzolli et al., 2004, 2006; Crawford & Guo, 2005). NO detoxification by the nonsymbiotic HB AHb1 has been demonstrated under hypoxic conditions in Arabidopsis (Dordas et al., 2003; Perazzolli et al., 2004). Furthermore, nsHbs have been shown to be induced by nitrate in Arabidopsis, Lotus japonicus and in cultured rice cells (Wang et al., 2000; Ohwaki et al., 2005; Shimoda et al., 2005). The conversion of nitrite to NO by NR was documented in 1988 (Dean & Harper, 1988) in Leguminosae and has been successively demonstrated in detail in other plant species (Wildt et al., 1997), including maize (Yamasaki et al., 1999; Yamasaki & Sakihama, 2000), implying a putative responsibility for NR as signal emitter, similar to that of mammalian NO synthase (NOS). Hence, in addition to its physiological function in nitrogen assimilation, NR may be required for the control of the production of active nitrogen species (Yamasaki & Sakihama, 2000). nsHbs may thus play a protective role against NR-derived NO generated by nitrogen fertilization (Perazzolli et al., 2006), but may also contribute to the modulation of NO-mediated signaling (Hebelstrup et al., 2007). NO is a bioactive molecule that functions in numerous physiological and developmental processes in plants, including lateral root development and response to abiotic stress (Siddiqui et al., 2010). It therefore cannot be excluded that this molecule could also operate at the interface between nitrate perception and transduction, taking part in the overall physiological and developmental plant adaptation to nitrate. The detection of NR and HB transcripts in tissues devoted to nutrient uptake and their spatial distribution in epidermal cells, the first layer of living cells in the root in contact with the external environment, strongly suggests that they could play an important role during the early perception and signaling of nitrate in the rhizosphere. Moreover the co-localization of mRNA for NR and HB observed in the root apex matches with the major sites of NO accumulation, as shown in Arabidopsis (Stöhr & Stremlau, 2006), suggesting that these two genes may represent the pivotal elements of a fine-tuning system for NO homeostasis and signaling.

A gene encoding a MAPK kinase, homolog to ATMKK2 of Arabidopsis, together with a gene coding for MAPK6, were also recovered among our genes. They showed transcription profiles similar to those measured for NR and HB, although less evident. ATMKK2 is a member of a signal transduction module consisting of the MAPKKK MEKK1, the upstream activator of MKK2 and the downstream MAPKs MAPK4 and MAPK6, mediating responses to cold and salt stresses (Teige et al., 2004). In particular, MAPK6 appears to act as a universal regulator in plant stress responses, as well as during growth and development (Feilner et al., 2005). In Arabidopsis, MAPK6 is also involved in the regulation of NO synthesis, via NR, in response to hydrogen peroxide during lateral root development (Wang et al., 2010). These authors demonstrated that the phosphorylation of NIA2 by MAPK6 considerably increased the activity of NIA2 and production of NO, and led to morphological changes in the root system of Arabidopsis.

Finally, a gene encoding BRI1, a brassinosteroid (BR) receptor-like kinase (Clouse et al., 1996), was also identified in our screens, the expression of which was slightly down-regulated after 6 h of nitrate depletion and more significantly after 5 d. The involvement of BRs in the regulation of the response to nitrate is not surprising given their role in the regulation of development (Bajguz, 2007). Transcript localization strengthened the involvement of BRI1 in root growth and, in particular, in root meristem development. The specific localization in the apical meristem further supports the role of a BRI1-dependent signaling pathway in the control of the quiescent center identity (González-García et al., 2011). Moreover, the epidermal localization of BRI mRNA is in agreement with the results reported by Hacham et al. (2011), demonstrating that the specific epidermal expression of BRI1 controls gene expression in the inner cell rows and drives root meristem size. In addition, BRs play important roles in inducing plant tolerance to various abiotic and biotic stresses (Li et al., 1996; Li & Chory, 1999; Nakashita et al., 2003; Kagale et al., 2007; Bajguz & Hayat, 2009; Ren et al., 2009; Tanaka et al., 2009). Furthermore, a very recent paper underlined a novel connection between NO and BRs in the enhancement of tolerance to oxidative damage caused by water stress in maize leaves (Zhang et al., 2011).

In conclusion, this work allowed us to identify a number of genes putatively involved in the early and late responses of maize roots to nitrate deprivation/provision, possibly playing key regulatory functions in the adaptation of this crop to nitrogen fluctuations in the soil. Furthermore, through optimized reproducible ISH experiments, the first proof was obtained of the differential localization of transcripts belonging to HB, NR, BRI1, peroxisomal and lysM genes in maize roots subjected to varied nitrate availability. At least for some target genes, ISH enabled a clear and reliable localization of the different accumulation of mRNAs according to nitrate availability. This could be used to reveal the response of a network of genes to changes in environmental conditions and to investigate differences between tissues and individual cells in the same or different species. On the basis of our results, the nitrate-derived NO, together with BRs, should be considered as additional players probably governing the root overall response to nitrate. Although a number of issues are still open and further studies will be needed to better elucidate the role of each component, in the complex and intriguing scenario of the nitrate signaling responses summarized in Fig. 9, a pivotal role is played by NO homeostasis, for which NR and HB appear to represent important fine-tuning elements.

Figure 9.

Model for nitrate-induced signaling in maize (Zea mays) roots: fluctuation of nitrate (NO3) concentration in soil leads to changes in nitrate reductase (NR) and BRASSINOSTEROID INSENSITIVE 1 (BRI1) mRNA levels. NR and BRI1 increase the production of nitric oxide (NO), which, in part, could be sequestered by hemoglobin (HB). Hemoglobin may play a protecting role against NR-derived NO, and may also contribute to the modulation of the NO-mediated signaling system. BR-induced NO production mediates abscisic acid (ABA) biosynthesis (Zhang et al., 2011), and can enhance the tolerance to abiotic stress. Free NO may affect a series of downstream physiological processes (e.g. lateral root development, nutrient uptake). In this scenario, abiotic stress could induce the expression of mitogen-activated protein kinase kinase 2 (MAPKK2), an upstream activator of MAPK6. MAPK6 seems to act as a universal regulator in plant stress responses (Feilner et al., 2005), and is also involved in the regulation of NO synthesis, via nitrate reductase, in response to hydrogen peroxide (H2O2) during lateral root development (Wang et al., 2010).


This project was funded by Progetto di Ateneo, Università di Padova, 2008 – prot. CPDA088137 and ex-60% 2010 – and by European Project ‘AUTOSCREEN’ (LSHG-CT-2007-037897) (M.B. and A.N.).