Nitrate is a major nitrogen source for land plants and also acts as a signaling molecule that induces changes in growth and gene expression. To identify the cis-acting DNA element involved in nitrate-responsive gene expression, we analyzed the promoter of the Arabidopsis gene encoding nitrite reductase (NIR1). A region from positions −188 to −1, relative to the translation start site, was found to contain at least one cis-element necessary for the nitrate-dependent activation of the promoter, in which the activity of nitrate transporter NRT2.1 and/or NRT2.2 plays a critical role. To define this nitrate-responsive cis-element (NRE), we compared the sequences of several nitrite reductase gene promoters from various higher plants and identified a conserved sequence motif as the putative NRE. A synthetic promoter in which the four copies of a 43-bp sequence containing the motif were fused to the 35S minimal promoter was found to direct nitrate-responsive transcription. Furthermore, mutations within this conserved motif in the native NIR1 promoter markedly reduced the nitrate-responsive activity of the promoter, indicating that the 43-bp sequence is an NRE that is both necessary and sufficient for nitrate-responsive transcription. We also show that both the native NIR1 promoter and the synthetic promoter display a similar level of sensitivity to nitrate, but respond differentially to exogenously supplied glutamine, indicating independent modulation of NIR1 expression by NRE-mediated nitrate induction and feedback repression mediated by other cis-element(s). These findings thus define the presence of multiple cis-elements involved in the nitrogen response in Arabidopsis.
Several recent studies have identified factors involved in the regulation of nitrate-responsive gene expression in plants. A dual-affinity nitrate transporter, NRT1.1, has been suggested to be one of these key components. In Arabidopsis nrt1.1 mutants, nitrate-responsive gene expression is only reduced when the mutants were grown with culture medium containing ammonium. As this conditional phenotype could not be fully explained by a reduced nitrate uptake by NRT1.1, this has been proposed to be a factor involved in nitrate signaling (Wang et al., 2009). Furthermore, the recent characterization of an nrt1.1 mutant, chl1-9, has revealed that NRT1.1 acts as a nitrate sensor, independent of its nitrate uptake activity (Ho et al., 2009). An RWP-RK protein, NLP7, may be another factor associated with nitrate-responsive expression, as a reduction in the nitrate-induced expression of NR genes and the gene for another nitrate transporter, NRT2.1, was observed in Arabidopsis nlp7 mutants (Castaings et al., 2009). A GATA-type transcription factor has also been implicated in the regulation of nitrogen utilization in higher plants (Bi et al., 2005). Recently, a protein kinase, CIPK8, was suggested to be a regulator that modulates nitrate-responsive gene expression in the presence of high concentrations of nitrate (Hu et al., 2009). In spite of the increasing number of identified factors for nitrate signaling and response, the mechanisms underlying nitrate-inducible transcription is still largely unknown, because of a lack of information on cis-elements and trans-acting factors that directly modulate nitrate-responsive transcription.
To investigate cis-elements for nitrate-responsive transcription, promoters of nitrate-inducible genes, including NRT2.1, NR and NIR genes, have been analyzed previously. A 150-bp region within the Arabidopsis NRT2.1 promoter was found to be sufficient to mediate stimulation by nitrate and repression by N metabolites (Girin et al., 2007). This suggested an intimate linkage between this sequence and the N status. However, the involvement of this sequence in the immediate activation of transcription by nitrate has not yet been addressed. On the other hand, complex results were obtained in other previous analyses of the NR gene promoters. The activity of the tobacco (Nicotiana tobacum) NR gene promoter was found not to be enhanced, but in fact repressed, by the supply of nitrate (Vaucheret et al., 1992; Vaucheret and Caboche, 1995). In contrast, the NR gene promoter from birch (Betula pendula) was reported to be mildly activated by nitrate (Hachtel and Strater, 2000). An earlier study of the promoters of the Arabidopsis NR genes (NIA1 and NIA2) showed that they could confer nitrate-inducible expression (Lin et al., 1994), and further analysis of these promoters via linker scanning mutagenesis experiments suggested that a 5′-(A/T)7A(C/G)TCA-3′ motif might be a nitrate-responsive cis-element (NRE) (Hwang et al., 1997). However, it has not yet been addressed whether this sequence is sufficient to induce nitrate-responsive gene expression. In contrast to the NR gene promoters, all of the analyzed NIR gene promoters from various higher plants have been clearly shown to direct the nitrate-responsive expression of reporter genes, suggesting that NREs are located in these promoter sequences (Neininger et al., 1994; Sander et al., 1995; Rastogi et al., 1997; Dorbe et al., 1998; Sivasankar et al., 1998; Warning and Hachtel, 2000). However, the precise cis-elements had not been defined in these reports, probably because progressive deletions from the 5′ ends of the NIR gene promoters causes a gradual decrease in both nitrate induction and basal promoter activity, thereby making it difficult to identify these elements.
Information regarding the cis-regulatory sequence that is sufficient for nitrate-responsive transcription is absolutely necessary to advance our understanding of the mechanisms underlying nitrate-responsive gene expression. Surprisingly, however, such information is currently scarce. In our current study, we analyzed the promoter of an Arabidopsis NIR gene (NIR1) and successfully identified a cis-element sufficient to confer nitrate-inducible transcription. Because glutamine, another possible regulator of N metabolism (Miller et al., 2007), has been shown to negatively regulate the expression of NIR1 (Vincentz et al., 1993; Sivasankar et al., 1997), we also investigated the relationship between our identified cis-element and glutamine-mediated feedback repression, and found that the cis-element is specifically involved in nitrate signaling but not in feedback repression. These results indicate that multiple cis-regulatory elements for the N response are present in Arabidopsis.
Nitrate-inducible expression of the GUS reporter gene when fused to the 5′ upstream and 3′ downstream sequences of NIR1
We generated a reporter construct, NIR1-GUS, by fusing both the 5′ upstream and 3′ downstream sequences flanking the NIR1 coding region to the GUS gene (Figure 1a). We then generated transgenic Arabidopsis plants harboring this construct. Histochemical staining of these plants subsequently showed that nitrate treatment of nitrogen-starved seedlings strongly induced GUS activity in the shoots (Figure 1b). Unexpectedly, strong GUS staining was also observed in the roots, even in the absence of nitrate (Figure 1b). Quantitative analysis of GUS activity in the shoots of six independent lines further revealed that nitrate treatment, but not control treatment with potassium chloride, drastically induced GUS activity (Table 1). However, these treatments did not exert any effect on GUS activity in control transgenic lines harboring the GUS gene under the control of the 35S promoter (35S-GUS; Table 1). Furthermore, quantitative RT-PCR analysis using whole seedlings revealed that nitrate strongly elevated the levels of both GUS and endogenous NIR1 mRNAs to a similar extent in the nitrogen-starved seedlings of the transgenic lines (Figure 1c).
Table 1. GUS activity in the shoots of seedlings harboring the NIR1-GUS reporter constructs
GUS activity (pmole MU min−1 mg protein−1)a
aGUS activity in seedlings that were grown on nitrogen-free half-strength MS plates for 3.5 days, and then treated with 10 mm KCl or KNO3 for 4 h. Values are means of biological replicates ± SE (n = 2).
162.0 ± 38.1
2567.1 ± 340.4
20.3 ± 0.1
928.5 ± 10.3
24.2 ± 15.9
610.8 ± 27.0
17.4 ± 4.3
887.8 ± 130.0
199.4 ± 12.0
2398.6 ± 58.3
20.0 ± 5.5
592.8 ± 72.8
6356.0 ± 139.4
5352.3 ± 351.9
672823.5 ± 40864.7
638821.3 ± 14106.3
Specific activation of the NIR1-GUS reporter gene by nitrate
We next examined the effects of various nitrogen sources on the expression of the GUS reporter gene in two independent lines carrying the NIR1-GUS construct (lines a3 and b2). We found that both ammonium nitrate and potassium nitrate markedly increased GUS activity (Figure 1d), whereas ammonium chloride and glutamine did not cause any increase. These results suggest that the NIR1 promoter is induced by nitrate, but not through the general promotion of protein synthesis that could be caused by a nitrogen source. A nitrate concentration of only 0.3 mm was effective in this induction in our assay system (Figure 1e). Our findings are also very consistent with previous reports that have indicated that low concentrations of nitrate (<250 μm) can induce NIR gene expression (Kramer et al., 1989; Wang et al., 2000).
To further validate the nitrate-responsive expression of our reporter gene, we introduced the NIR1-GUS construct into nitrate transporter mutants by crossing these plants with the transgenic line a3. It has been demonstrated previously that the uptake of nitrate at a low concentration is carried out by high-affinity nitrate transporters, including NRT2.1 (Cerezo et al., 2001; Filleur et al., 2001), and also by a dual-affinity transporter NRT1.1 (Liu et al., 1999; Liu and Tsay, 2003). When nitrate is present at high concentrations, however, NRT1.1 is thought to play the main role in its uptake into the plant cell (Tsay et al., 1993). We grew these mutant lines carrying the NIR1-GUS reporter gene on nitrogen-free medium and treated them with different concentrations of potassium nitrate. In the nrt2.1-nrt2.2 Arabidopsis mutant, in which the high-affinity nitrate uptake system is deficient (Li et al., 2007), only high concentrations of nitrate could increase GUS activity (Figure 1f). This suggests that this reporter induction is, at least in part, dependent on the activity of NRT2.1 and/or NRT2.2. On the other hand, the knock-out of the NRT1.1 gene in the chl1-5 null mutant (Tsay et al., 1993; Munos et al., 2004) did not affect the nitrate induction of GUS activity (Figure 1f), indicating that NRT1.1 activity is not critical for nitrate-responsive increases in NIR1 expression in these growth conditions.
It has been reported that the effects of the mutations in NRT1.1 could only be seen when plants were grown in medium containing ammonium prior to nitrate treatment (Touraine and Glass, 1997; Wang et al., 2009). Therefore, we also performed experiments with shoots of seedlings that were grown on nitrogen-free half-strength MS medium supplemented with 2.5 mm ammonium succinate, and then treated with different concentrations of nitrate. Unexpectedly, GUS activity was much higher in the chl1-5 mutant than in the wild-type Arabidopsis when seedlings were treated with low concentrations of nitrate, whereas the activity was comparable in chl1-5 and wild-type Arabidopsis when 10 mm nitrate was used for treatment (Figure S1). We repeatedly obtained similar results in three independent experiments in which different periods of ammonium pre-culture and different batches of seeds were employed, confirming the effect of the chl1-5 mutation on the expression of our reporter gene. On the other hand, the reduced induction of GUS activity in the nrt2.1-nrt2.2 mutant was observed again in this experiment, further supporting a critical role of NRT2.1 and/or NRT2.2 activity in the expression of the NIR1-GUS reporter gene (Figure S1).
Since the histochemical staining of seedlings did not allow the visualization of nitrate-inducible expression of GUS reporter gene in the roots (Figure 1b), we re-examined this issue by measuring GUS activity. Although GUS activity was high even in the absence of exogenous nitrate, a significant increase was detected in this activity in response to nitrate treatment (Figure 1g). However, nitrate treatment only marginally induced this reporter activity in the roots of the nrt2.1-nrt2.2 mutant (Figure 1g). These data suggest that the nitrate-induced expression of NIR1 in roots is also dependent on NRT2.1 and/or NRT2.2 activity. Because GUS activity in roots was found to be lower in the nrt2.1–nrt2.2 mutant compared with wild-type Arabidopsis in the absence of exogenous nitrate, the high basal level of GUS activity in the wild-type plants may be caused by nitrate carried over from the seeds.
Deletion analysis of the 5′ flanking sequence of NIR1
Because a stronger induction of the GUS reporter gene was detectable in shoots, we utilized these organs in our further analyses. To delineate the sequence required for the nitrate response, we generated additional reporter constructs; 5′ flanking sequences of 3120, 856 and 334 bp in length were fused to the GUS gene to generate NIR1pro-GUS, NdeI-GUS and RsaI-GUS constructs, respectively (Figure 2a). Using these constructs, we generated more than four independent transgenic lines for each vector. In all transgenic lines, nitrate treatment appeared to induce GUS activity (Figure 2a). However, a gradual decrease in this reporter activity at both basal and induced levels was observed when the shorter fragments were used as promoters, and no activity could be detected at all for the RsaI-GUS construct without nitrate treatment. We then performed quantitative RT-PCR analysis to evaluate the nitrate-responsive expression of the GUS gene in these transgenic lines, and found the activity levels for all of the deleted promoters increased in response to nitrate (Figure 2b–d). This indicated that an NRE resides in the region between positions −314 and −1 of the NIR1 promoter.
Because previous analyses of the NIR gene promoters from spinach and tobacco have shown that the sequences close to their transcription start sites are sufficient to confer nitrate-responsive increases in reporter enzyme activity (Neininger et al., 1994; Dorbe et al., 1998), we examined the possibility that an NRE of the Arabidopsis NIR1 promoter is also located proximally to the transcription start site. Because sequences in the neighborhood of transcription start sites seed the formation of transcriptional initiation complexes and harbor cis-sequences for basal promoter activity, bulk deletions or mutations within such sequences may cause drastic reductions in gene expression activity. In the case of our reporter system, this would cause difficulties in precisely identifying the NRE. As we speculated that further truncation of the RsaI-GUS reporter gene would result in a loss of detection of GUS activity and mRNA, we decided to make an internal deletion construct, DD-min-GUS, in which the sequence from −188 to −1 of the NIR1 promoter was replaced with the 35S minimal promoter (Figure 3). This chimeric promoter could not confer transcription in response to nitrate treatment (Figure 3, DD-min-GUS), whereas GUS activity in response to nitrate treatment was observed when the sequence from −688 to −1 was used as a promoter (Figure 3, DraI-GUS construct). Activity of the 35S minimal promoter was unaffected by nitrate treatment (Figure 3, min-GUS construct). The result suggested that the sequence between −188 and −1 is indeed required for the nitrate response.
Identification of a sequence motif conserved in the NIR promoters
Because our data indicated that at least one NRE is located in the proximal region of the Arabidopsis NIR1 promoter, we analyzed conserved sequence(s) in the region ranging from −188 to −1 of the NIR1 promoter, making the assumption that the NRE is conserved in the NIR gene family in higher plants. An alignment of the proximal sequences of the Arabidopsis NIR1 and spinach NIR promoters revealed significant homology (data not shown). By further comparison with the NIR gene promoters from bean (Phaseolus vulgaris), birch and tobacco, a pseudo-palindromic sequence, tGACcCTTN10AAGagtcc (completely conserved nucleotides and less conserved nucleotides are shown in uppercase and lowercase, respectively) emerged as a candidate NRE sequence (Figure 4a). This sequence is located between positions −104 and −62 relative to the translation start site of Arabidopsis NIR1. Although the longest full-length transcript of NIR1 has a 5′ untranslated region (5′-UTR) of 76 bp, full-length transcripts with a 5′-UTR of 20–25 bp appear to be most abundant (Figure S2). We thus speculated that the conserved sequence is located about 40-bp upstream of the transcription start site of NIR1.
The role of the 43-bp sequence containing the conserved NRE sequence motif
To test our hypothesis that the identified conserved sequence motif acts as an NRE, we generated a synthetic promoter in which four copies of the 43-bp sequence containing the conserved sequence from the NIR1 promoter were fused to the 35S minimal promoter (Figure 4b). We then produced transgenic plants expressing the GUS gene under the control of this synthetic promoter (4 × NRE-min-GUS lines). By histochemical staining and GUS activity measurements, the reporter activity in nitrogen-starved seedlings of these lines was shown to increase drastically in response to nitrate treatment (Figure 4c; Table 2). Consistent results were also obtained by quantitative RT-PCR (Figure 4d). In addition, the synthetic promoter was not activated by exogenously supplied ammonium or glutamine (Figure 4e). Furthermore, a nitrate concentration of as low as 0.3 mm was sufficient to activate both the synthetic and the native NIR1 promoters (Figure 4f). Hence, the 43-bp sequence was verified as an NRE that was alone sufficient to respond to a nitrate signal.
Table 2. GUS activity in the shoots of transgenic lines harboring NIR1-GUS, 4×NRE-min-GUS or min-GUS
GUS activity (pmole MU min−1 mg protein−1)a
aGUS activity in seedlings that were grown on nitrogen-free half-strength MS plates for 3.5 days, and then treated with 10 mm KCl or KNO3 for 4 h. Values are means of biological replicates ± SE (n = 2).
252.1 ± 12.5
5533.9 ± 873.1
23.6 ± 3.0
1093.7 ± 6.9
366.8 ± 15.9
2988.8 ± 427.9
133.7 ± 37.3
2447.8 ± 222.3
1.6 ± 0.2
64.0 ± 2.4
83.6 ± 39.6
2111.1 ± 248.0
24.5 ± 6.1
416.9 ± 25.1
0.23 ± 0.10
0.20 ± 0.08
0.20 ± 0.09
0.27 ± 0.02
A critical role of the conserved sequence motif in NRE activity
To clarify the role of the conserved sequence motif in nitrate-inducible NIR1 expression, the effects of disrupting this region were examined. Because the NRE is pseudo-palindromic, we introduced mutations on the distal half-site (M1 and M2) and the proximal half-site (M3) independently, and both half-sites (M4) in the native 3.1-kb promoter (Figure 5). Both M2 and M3 mutations largely reduced the nitrate-dependent activation. Furthermore, the M2 mutation also decreased the basal promoter activity. Although the M1 and M4 mutations drastically lowered GUS activity in nitrate-treated seedlings, it was impossible to calculate the magnitude of nitrate induction because GUS activity was not detected in the absence of exogenous nitrate. Collectively, these results suggested that both half-sites of the pseudo-palindromic sequence are involved in nitrate-induced expression of NIR1. Because these mutations drastically reduced the reporter activity in the absence of exogenously applied nitrate, we speculate that the identified NRE might also be the essential element that confers basal promoter activity itself. The higher basal activity of the wild-type NIR1 promoter might result from an endogenous nitrate response.
The function of the NRE is independent of glutamine repression
Nitrogen promotes the biosynthesis of glutamine and generally enlarges the pools of available amino acids. Glutamine is also regarded as another possible regulator of N metabolism (Miller et al., 2007). Glutamine is known to reduce both nitrate uptake and levels of NR and NIR mRNAs (Oaks et al., 1977; Langendorfer et al., 1988; Henriksen and Spanswick, 1993; Vincentz et al., 1993; Sivasankar et al., 1997). However, glutamine induces a much more severe reduction in the mRNA levels, compared with nitrate uptake. In fact, exogenous supply of 10 and 50 mm glutamine induced an about 20 and 45% reduction in nitrate accumulation, respectively (Oaks et al., 1977; Langendorfer et al., 1988), whereas the supply of 1 mm glutamine was sufficient to decrease both NR and NIR mRNA levels severely without affecting the nitrate concentration in tissues. Thus, the glutamine repression of NR and NIR gene expression is supposed to be a direct effect rather than a secondary effect of reduced nitrate uptake (Sivasankar et al., 1997).
To investigate whether this negative regulation is mediated by our identified NRE or other cis-regulatory element(s), we compared GUS activity in the shoots of transgenic lines harboring NIR1-GUS and 4xNRE-min-GUS constructs. Exogenously supplied glutamine was found to reduce GUS activity in the NIR1-GUS lines by approximately 50%, whereas such a decrease was not observed in the 4×NRE-min-GUS lines (Figure 6a). This indicates that the repression of NIR1 by glutamine is mediated by cis-elements other than NRE.
We further found that the GUS activity levels were an order of magnitude higher in the 4×NRE-min-GUS lines than in the NIR1-GUS plants when they were grown on half-strength MS medium containing an abundant N source (Figure 6a). In stark contrast, all of these lines showed comparable GUS activities when transiently treated with nitrate (Figure 4f; Table 2). We thus speculated that a constitutive supply of nitrogen represses the NIR1 promoter but not the 4×NRE-min synthetic promoter. To test this possibility, we monitored fluctuations in GUS activity in these lines after the transfer of nitrogen-starved seedlings to half-strength MS medium. As shown in Figure 6b, reporter activity in the NIR1-GUS transgenic lines reached a peak at 6 h after transfer, and then decreased (Figure 6b). In contrast, GUS activity in the 4×NRE-min-GUS lines continued to increase for the 96-h period, supporting our hypothesis that element(s) other than the NRE mediate this negative feedback regulation of NIR1 expression.
Identification of the cis-regulatory elements that are sufficient for nitrate-responsive transcription is critical to our further understanding of the mechanisms underlying plant growth. This is primarily because the nitrate content in the soil is one of most important effectors of plant growth. In our current study, we analyzed the NIR1 promoter and identified a 43-bp sequence containing a conserved sequence, 5′-tGACcCTTN10AAGagtcc-3′, as a cis-element that is both necessary and sufficient for nitrate-responsive transcription. By crossing transgenic plants harboring reporter genes and nitrate transporter mutants of Arabidopsis, we show in our assay system that the nitrate-responsive expression of the GUS reporter gene is dependent on NRT2.1 and/or NRT2.2 activity. Furthermore, we demonstrate that the NRE is specifically involved in nitrate induction, but not glutamine repression, indicating that multiple cis-elements are involved in the nitrogen-regulated expression of NIR1. Our results therefore provide a platform for further study on the growth control mechanisms that are mediated by responses to plant nutrients.
Nitrate-inducible NIR1 expression
Our results indicate that the high-affinity nitrate transporters NRT2.1 and/or NRT2.2 play a critical role in nitrate-inducible NIR1 expression, particularly when the concentration of environmental nitrate is low. NRT2.1 and/or NRT2.2 might affect NIR1 expression, because NRT2.1-dependent nitrate uptake might modulate the nitrate concentration inside plant cells where nitrate sensing might take place. It is also possible to speculate that NRT2.1 might act as a nitrate sensor of which activity regulates nitrate-inducible NIR1 expression, because NRT2.1 has already been proposed to be a nitrate sensor or signal transducer in regulating the development of the root system (Malamy and Ryan, 2001; Little et al., 2005). Investigation of these possibilities would be of importance for a deeper understanding of the mechanism underlying nitrate-inducible NIR1 expression. Although NRT1.1 has recently been shown as a nitrate sensor (Ho et al., 2009), effects of its mutations were only evident when plants were grown with medium containing ammonium (Touraine and Glass, 1997; Wang et al., 2009). We also observed the unexpected and complicated effect of the nrt1.1 mutation on the expression of the NIR1-GUS reporter gene, when medium containing ammonium was used (Figure S1). Although NRT1.1 activity might be involved in the regulation of NIR1 expression under particular growth conditions, nitrate induction of NIR1 expression is independent of growth in medium containing ammonium. Therefore, sensors other than NRT1.1 may well play a critical role in the nitrate response of the NIR1. Identification of nitrate sensor(s) that transmits the nitrate signal to the NRE identified in this study would be an important challenge in the future.
The 43-bp sequence containing a conserved sequence, 5′-tGACcCTTN10AAGagtcc-3′, functions as an authentic NRE
In general, promoter activity is determined by the combined effects of many cis-elements. Some of these sites mediate particular intercellular or intracellular stimuli, or developmental signals, and others function simply as an enhancer that is independent of a particular signaling pathway. Hence, even if a reduction or loss of reporter enzyme activity can be detected, it does not always mean a loss of cis-elements mediating particular signals. To identify cis-elements mediating particular signals, evaluation with not only deletion analysis but also with complementary analysis, namely gain-of-function experiments, is generally necessary. Herein we have revealed that the 43-bp sequence containing a conserved sequence, 5′-tGACcCTTN10AAGagtcc-3′, is sufficient to mediate a nitrate response, and is necessary for the full activation of the native NIR1 promoter by nitrate. Accordingly, this 43-bp sequence satisfies the criteria for an authentic NRE.
Mutational analysis indicated that both the half-sites of the pseudo-palindromic sequence in the 43-bp sequence are important. This implies that a trans-acting factor might bind the palindromic sequence as a dimer, of which each subunit binds to a half-site separated by a 10-bp spacer. An example of this type of DNA–protein interaction is the binding of yeast GAL4 protein to its recognition sequence, 5′-CGGN11CCG-3′ (Liu et al., 2002). Another example is Auxin Response Factor 1 (ARF1) from Arabidopsis, which forms a dimer and binds, most efficiently, palindromic sequences in which the 6-bp-long half-sites are separated by a spacer of seven or eight bases (Ulmasov et al., 1997). However, we cannot exclude another possibility that the distal and proximal half-sites in the pseudo-palindromic sequence are bound by distinct transcription factors, both of which are important for the NRE function.
A number of previous promoter studies have described nitrate-inducible genes, but none had identified any sequences that are sufficient to confer nitrate-responsive expression, with the exception of a 150-bp region from the Arabidopsis NRT2.1 promoter. In an earlier study of the spinach NIR gene promoter, Rastogi et al. (1997) proposed that a sequence from −220 to −200 (relative to the transcription start site) of the promoter contains a cis-element for nitrate induction, and is bound by a GATA-type factor. However, because the 35S minimal promoter fused to this region (from −195 to −230) failed to produce reporter enzyme activity at detectable levels (Rastogi et al., 1997), it is currently unknown whether this sequence is sufficient to confer nitrate-responsive transcription. Another sequence, 5′-(A/T)7A(C/G)TCA-3′ from the Arabidopsis NR genes, NIA1 and NIA2, has also been proposed to be an NRE (Hwang et al., 1997). However, the activity of this sequence has not yet been evaluated using gain-of-function analysis. Although sequences matching this consensus motif have also been found in the NIR1 promoter, between positions in the ranges from −240 to −229 and from −257 to −268 (Hwang et al., 1997), the relationship between these sequences and the NRE identified in this earlier report is still unclear.
The 150-bp region in the NRT2.1 promoter has been shown to be sufficient for both metabolite repression and stimulation by nitrate (Girin et al., 2007), although no sequence similar to our NRE is present in this region. It is noteworthy that this 150-bp sequence was identified by analysis with seedlings grown for 7 days on media containing different concentrations of N source, whereas we chose in our present study to use nitrate treatments over a short period (1 h) to identify cis-elements involved in an immediate response to nitrate. Thus, the 150-bp region and our NRE might mediate a different type of N response.
Nitrate only induced expression of the RsaI-GUS reporter gene at a low level (Figure 2), suggesting that an element involved in nitrate induction might exist in a region between −334 and −3120 bp. However, the result of mutational analysis suggested that the identified NRE mostly conferred the nitrate responsiveness of the NIR1 promoter in the native promoter context (Figure 5). Therefore, the element in the upstream region alone may partially contribute to the nitrate response of the NIR1 promoter, or may function cooperatively with the identified NRE. In fact, no sequence that matches the identified NRE was found in this region. However, we cannot rule out the possibility that the NRE sequence might have a higher degree of redundancy than we expected from the alignment of the proximal sequences, and that a variant of the NRE might be present in the upstream region. In this regard, it is noteworthy that some transcription factors bind multiple recognition sequences (Vasil et al., 1995; Ogo et al., 2008).
Candidate NRE-binding proteins
Although no trans-acting factors that directly regulate nitrate-responsive transcription in higher plants have been identified to date, such trans-acting factors in fungi and Chlamydomonas have been described. In Aspergillus nidulance and Neurospora crassa, the combination of the GATA-type transcription factors, AreA and NIT2, and GAL4-type transcription factors, NirA and NIT4, has been shown to regulate nitrate-inducible gene expression (Marzluf, 1997). Although higher plants do not possess any GAL4-type transcription factor (Riechmann et al., 2000), GATA-type transcription factors are present in higher plants, one of which has now been implicated in the regulation of nitrogen utilization (Bi et al., 2005). However, the core recognition sequence of the N. crassa GATA factor, NIT2, is 5′-GATA-3′ (Marzluf, 1997), which differs from our identified NRE. Furthermore, given that the GATA-type transcription factors are involved in global nitrogen repression rather than nitrate induction itself in A. nidulans and N. crassa, they are unlikely to be candidate NRE-binding proteins in Arabidopsis. In Chlamydomonas, the nitrate response of the NR gene is dependent on the short regions in the promoter (Loppes and Radoux, 2002). An RWP-RK protein, NIT2 (unrelated to the N. crassa GATA-factor, NIT2), has also been shown to bind this sequence (Camargo et al., 2007). Recently, it has been reported that an RWP-RK protein, NLP7, is involved in nitrate signaling in Arabidopsis. Indeed, nitrate induction of the NIA1, NRT2.1 and NRT2.2. genes was found to be reduced in nlp7 mutants (Castaings et al., 2009). However, because the sequence for binding of NIT2 completely differs from the NRE identified in our current study, RWP-RK proteins are also not candidate NRE-binding proteins.
Among the transcription factors in higher plants for which the consensus binding sequence has been established, the WRKY transcription factors are candidate NRE-binding proteins, as the 43-bp NRE sequence we have identified contains the binding sequence for these proteins, i.e. 5′-TGAC-3′ (Eugelm et al., 2000). Dof transcription factors are also candidates, because 5′-CTT-3′ and 5′-AAG-3′, conserved motifs in the NRE, are partially similar to the Dof binding sequences 5′-AAAG-3′ and 5′-CTTT-3′ (Yanagisawa, 2002). However, it must be noted that the requirement for a complete 5′-AAAG-3′ sequence for the binding of Dof factors is strict (Yanagisawa and Schmidt, 1999). It is also unlikely that the NRE is bound by the ANR1 MADS-box transcription factor that regulates lateral root growth in the nitrate-rich zone (Zhang and Forde, 1998), because the binding sequence of the MADS-box proteins is CC(A/T)6GG (the CArG box; de Folter and Angenent, 2006). It is more likely that a novel-type of transcription factor binds to the NRE, and our identification of this sequence will facilitate a search for this putative factor in future studies.
Feedback regulation of NIR1 expression
Feedback regulation by N metabolites is an important component of the N response in plants. In terms of the identity of the mediator that is directly involved in this feedback regulation, glutamine is regarded as a strong candidate (Miller et al., 2007). Indeed, our current results indicate that exogenously supplied glutamine represses expression from the NIR1-GUS construct, but not from the 4xNRE-min-GUS construct. This observation reveals that the NRE is not susceptible to metabolite repression, and thus suggests that repression by N metabolites and induction by nitrate may be regulated through different molecular mechanisms in Arabidopsis. Our conclusion also supports a previous proposal by Sivasankar et al. (1997) that the feedback repression of NIR1 expression by glutamine is a direct effect on gene expression, rather than a secondary effect of reduced nitrate uptake.
The repression of nitrate-inducible gene expression by downstream metabolites is also known to occur in fungi and Chlamydomonas, although the underlying molecular mechanisms are very different between these organisms. In A. nidulans, the GATA-type transcription factor AreA is responsible for the release from metabolite repression (Marzluf, 1997), whereas the nuclear localization of another GAL4-type transcription factor NirA is regulated by nitrate (Bernreiter et al., 2007). Metabolite repression and nitrate induction are therefore mediated through different transcription factors. In Chlamydomonas, a different mechanism operates during nitrate-inducible gene expression in response to nitrate and ammonia, in which an RWP-RK protein, NIT2, mediates nitrate induction (Camargo et al., 2007). The removal of ammonium from the medium increases the NIT2 transcript levels, but this is not sufficient for the nitrate induction of NR gene expression, and the supply of nitrate in addition to the removal of ammonium is required to activate NIT2 post-transcriptionally in Chlamydomonas (Camargo et al., 2007). In this case, both repression by ammonium and induction by nitrate are thought to converge on NIT2, thereby acting on the same cis-element. Hence, with respect to the fact that different cis-elements mediate nitrate induction and metabolite repression, our current results suggest that the regulation of NIR1 expression by nitrogen in Arabidopsis is more similar to the mechanism in fungi than that in Chlamydomonas.
The Arabidopsis thaliana ecotype Columbia was used as the wild-type strain. The Arabidopsis deletion mutant of NRT1.1, chl1-5 (Tsay et al., 1993; Munos et al., 2004), and the nrt2.1–nrt2.2 mutant (SALK_035429) that has a T-DNA insertion within the NRT2.1-NRT2.2 gene cluster (Li et al., 2007), were obtained from the Arabidopsis Biological Resource Center (http://abrc.osu.edu/). The background of the chl1-5 and nrt2.1-nrt2.2 mutants is the ecotype Columbia. A transgenic line that had the NIR1-GUS reporter gene (line a3) inserted into the At3g04210 locus (data not shown) was crossed with these mutants, and the resulting F2 plants homozygous for both NIR1-GUS reporter insertion allele and the chl1-5 or the nrt2.1–nrt2.2 allele were selected by PCR genotyping.
Construction of reporter plasmids
To generate a binary vector containing the GUS gene fused to both side sequences of the coding region of the NIR1 gene, a 3.1-kb sequence for the upstream sequence and a 1.9-kb sequence for the downstream sequence were amplified by PCR using Arabidopsis genomic DNA as the template. The primers used were as follows: NIR1p-FW-H3, 5′-GTGAAGCTTGTCACCGGGGACCAATGTTAAGAC-3′; NIR1p-RV-NcoI, 5′-GAGCCATGGCGATGATGGCGGAAGAAGG-AGTTGGG-3′; NIR1d-FW-SacI, 5′-GTGGAGCTCTTCAAAAGCTAT-TGGATTCTTAATAAGTC-3′; NIR1d-RV-EcoRI, 5′-GTCGAATTCTGT-GAAACGTCTCATGGCTCAGG-3′. The obtained DNA fragment for the NIR1 promoter was digested with HindIII and NcoI, whereas the PCR product for the downstream sequence was digested with SacI and EcoRI. These fragments were used to replace the 35S-Ω promoter and nopaline synthase terminator in the pCB302-35SΩ-GUS vector (Konishi and Yanagisawa, 2007). A binary vector, pNIR1pro-GUS, was similarly generated by replacing only the 35S-Ω promoter in pCB302-35SΩ-GUS with the 3.1-kb NIR1 promoter fragment. To produce several binary vectors containing NIR1 promoters that were differently truncated at the 5′ end, the 35S-Ω promoter of pCB302-35SΩ-GUS was replaced with appropriate DNA fragments for the truncated NIR1 promoters.
To generate a reporter construct containing four copies of NRE (4xNRE) fused to the 35S minimal promoter, we first cloned a single copy of the NRE into the multicloning region of pUC19, and then produced the 4×NRE construct by ligation of NRE fragments recovered from the resultant plasmid. We replaced the 4×EBS sequence and LUC gene in the EBS-LUC plasmid (Yanagisawa et al., 2003) with the 4xNRE sequence and GUS gene, respectively.
Reporter constructs with mutated NIR1 promoters were generated by PCR using a plasmid containing the 1.3-kb PstI-NcoI fragment of the NIR1 promoter and the primers 5′-ATTGCTCA-AGAGCTCATCTCTTCC-3′ and 5′-GTAAAGGcagtAGTTGTTTCTCTT-TGATCGTATG-3′ for M1, 5′-ATTGCTCAAGAGCTCATCTCTTCC-3′ and 5′-GTAttccGTCAAGTTGTTTCTCTTTGATCGT-3′ for M2, 5′-ttctcCTCATCTCTTCCCTCTACA-3′ and 5′-GAGCAATGTAAAGGGTCAAGTTG-3′ for M3, 5′-ttctcCTCATCTCTTCCCTCTACA-3′ and 5′-GAGCAATGTAttcccagtAGTTGTTTCTCTTTGATCGT-3′ for M4. The mutated sequence is indicated in lowercase. Both ends of the PCR product were phosphorylated using T4 kinase, and this was followed by self-ligation. The cloned fragments were sequenced, and a PstI-NcoI fragment containing the intended mutation was recovered from the resultant plasmid, and used to replace the PstI-NcoI fragment in the NIR1pro-GUS construct.
Plant growth conditions
Seeds were sterilized, sown on half-strength MS plates that contained half-strength Murashige and Skoog salts, Gamborg’s vitamin, 0.8% agar, 0.5 g L−1 2-(N-morpholine)-ethanesulphonic acid (MES), pH 5.7, and subjected to cold treatment for 3 or 4 days. The plates were then transferred to a growth chamber set at 23°C with continuous light (60 μE).
For RT-PCR analysis, seeds were germinated in nitrogen-free half-strength MS liquid medium containing half-strength MS salts (from which ammonium nitrate and potassium nitrate were omitted), Gamborg’s vitamin, and 0.5 g L−1 MES, pH 5.7, and then grown under continuous light at 23°C. After 3.5 days, the medium was changed to fresh nitrogen-free half-strength MS liquid medium supplemented with 10 mm KNO3 or 10 mm KCl. After 1 h of this treatment, the seedlings were collected. One replicate sample was approximately 200–300 seedlings grown in a plastic dish with 20 ml of liquid medium.
For the analysis of the nitrate induction of GUS activity, seeds were sown on nitrogen-free half-strength MS plates containing 0.5% sucrose. Each plate contained 60 ml of medium. After cold treatment, plates were placed vertically at 23°C under continuous light for 3.5 days, and 1 ml of the appropriate concentration of KCl or KNO3 solution was then applied. The plates were placed horizontally and incubated for 4 h prior to sampling. Medium containing ammonium was nitrogen-free half-strength MS liquid medium, pH adjusted to 6.5, and supplemented with 2.5 mm ammonium succinate and 1% sucrose. One replicate sample was shoots or roots of 10–15 seedlings grown on an agar plate, and replicates were prepared from different plates.
For the glutamine repression assay, the seedlings were grown on half-strength MS agar plates containing 1% sucrose with or without 10 mm glutamine for 4 days. For transfer experiments, the seedlings were grown first on nitrogen-free half-strength MS plates containing 0.5% sucrose for 3.5 days and then transferred to half-strength MS plates containing 1% sucrose. Replicate samples were prepared from different plates.
Arabidopsis transformations were carried out using the floral-dip method, as described previously (Konishi and Yanagisawa, 2007). Using the T2 generation, the Arabidopsis lines harboring a T-DNA insertion at a single locus were selected based on the segregation of seedlings that were resistant or sensitive to glufosinate ammonium. T3 lines homozygous for the T-DNA insertion were used in the experiments.
Histochemical staining and quantitative measurement of GUS activity
Histochemical GUS staining was performed as described previously (Konishi and Yanagisawa, 2007), and measurement of GUS activity using fluorescent substrate was carried out, according to the method described by Jefferson et al. (1987). Briefly, the shoots or roots of about 10 seedlings were homogenized in 50 μl of extraction buffer (50 mm NaPO4, pH 7.4, 10 mm EDTA, 0.1% Triton X-100, 0.1% sodium lauryl sarcosine, 10 mmβ-mercaptethanol). Enzyme reactions were carried out in 50 μl of extraction buffer containing 1 mm 4-methylumbelliferyl-β-d-glucuronide at 37°C, and were stopped by the addition of 450 μl of 0.2 m Na2CO3. The 4-methylumbelliferone fluorescence was measured using a DyNA Quant fluorometer (GE healthcare, http://www.gehealthcare.com) or Infinite M1000 microplate reader (TECAN Group Ltd, http://www.tecan.com). Protein concentration was quantified using the Protein Assay reagent (Bio-Rad Laboratories, http://www.bio-rad.com).
RNA extraction and quantitative RT-PCR
RNA extractions and reverse-transcription were performed as described by Konishi and Yanagisawa (2007). PCR reactions were performed with a StepOne PlusTM Real Time PCR System (Applied Biosystems, http://www.appliedbiosystems.com), using the KAPA SYBRR Fast qPCR Kit (KAPA Biosystems, http://www.kapabiosystems.com). The primers used were as follows: 5′-CATGGGATGCTTAACACGAG-3′ and 5′-AATGGAACCAACTCCGTGAC-3′ for NIR1, 5′-ACGGCAGAGAAGGTACTGGA-3′ and 5′-AACGTATCCACGCCGTATTC-3′ for GUS. An Arabidopsis ubiquitin gene, UBQ10, was selected as an internal standard. The primers for UBQ10 have been described in Czechowski et al. (2005).
We thank the Arabidopsis Biological Resource Center for providing seeds for the SALK_035429C and chl1-5 lines. This research was supported by a Grant in Aid for JSPS fellows, the PROBRAIN, the CREST program from JST and KAKENHI (21114004 and 22380043) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.