Regulation of the High-Affinity NO₃⁻ Uptake System by NRT1.1-Mediated NO₃⁻ Demand Signaling in Arabidopsis

Abstract

The NRT2.1 gene of Arabidopsis thaliana encodes a major component of the root high-affinity transport system (HATS) that plays a crucial role in uptake by the plant. Although NRT2.1 was known to be induced by and feedback repressed by reduced nitrogen (N) metabolites, NRT2.1 is surprisingly up-regulated when concentration decreases to a low level (<0.5 mm) in media containing a high concentration of or Gln (≥1 mm). The NRT3.1 gene, encoding another key component of the HATS, displays the same response pattern. This revealed that both NRT2.1 and NRT3.1 are coordinately down-regulated by high external availability through a mechanism independent from that involving N metabolites. We show here that repression of both genes by high is specifically mediated by the NRT1.1 transporter. This mechanism warrants that either NRT1.1 or NRT2.1 is active in taking up in the presence of a reduced N source. Under low provision, NRT1.1-mediated repression of NRT2.1/NRT3.1 is relieved, which allows reactivation of the HATS. Analysis of atnrt2.1 mutants showed that this constitutes a crucial adaptive response against toxicity because taken up by the HATS in this situation prevents the detrimental effects of pure nutrition. It is thus hypothesized that NRT1.1-mediated regulation of NRT2.1/NRT3.1 is a mechanism aiming to satisfy a specific demand of the plant in relation to the various specific roles that plays, in addition to being a N source. A new model is proposed for regulation of the HATS, involving both feedback repression by N metabolites and NRT1.1-mediated repression by high .

Higher plants acquire mineral nitrogen (N) from the soil mainly in the form of , through the activity of both high-affinity transport systems (HATS) and low-affinity transport systems (LATS), respectively (Crawford and Glass, 1998; von Wirén et al., 2000). Except in agricultural soils after fertilizer application, where concentration can rise up several millimolar (Crawford and Glass, 1998), it is generally assumed that root uptake is mostly determined by the activity of the HATS (Crawford and Glass, 1998; von Wirén et al., 2000; Malagoli et al., 2004). The current model of the HATS is constituted by at least two genetically separate transport systems: (1) a constitutive HATS present even in the absence of (Wang and Crawford, 1996; Crawford and Glass, 1998); and (2) an inducible HATS (iHATS), quantitatively more important than the constitutive HATS, and activated within hours after provision to the plant (Behl et al., 1988; Mackown and McClure, 1988; Crawford and Glass, 1998; Daniel-Vedele et al., 1998; Forde, 2000).

In Arabidopsis (Arabidopsis thaliana), two gene families have been found to encode transporter proteins involved in root uptake (Forde, 2000; Williams and Miller, 2001; Orsel et al., 2002; Okamoto et al., 2003): the NRT2 family (seven members) and the NRT1 family, belonging to the large PTR family of transporter genes (53 members). Convincing evidence has accumulated that the -inducible NRT2.1 gene encodes a major component of the iHATS (Filleur and Daniel-Vedele, 1999; Lejay et al., 1999; Zhuo et al., 1999). First, atnrt2.1 mutants display a strong reduction of root influx in the low external concentration range (i.e. below 1 mm; Filleur et al., 2001; Orsel et al., 2004). Second, inducibility of high-affinity root uptake is suppressed by NRT2.1 deletion (Cerezo et al., 2001). Recently, another key component of the HATS has been identified in the form of the product of the NAR2-like Arabidopsis gene NRT3.1 (Okamoto et al., 2006). It is suspected that interaction between NRT2.1 and NRT3.1 is required for functionality of the iHATS because NRT2 proteins need to be coexpressed with NAR2-like proteins to generate uptake activity in Xenopus oocytes (Quesada and Fernandez, 1994; Zhou et al., 2000; Tong et al., 2005).

Besides induction by , NRT2.1 expression and HATS activity are also feedback repressed by reduced N metabolites, such as and amino acids (Lejay et al., 1999; Zhuo et al., 1999; Nazoa et al., 2003). This regulation involves systemic signaling (Gansel et al., 2001) and ensures the specific control of root uptake by the N status of the whole plant (Imsande and Touraine, 1994; Forde, 2002). Indeed, N sufficiency triggers the repression exerted by N metabolites and down-regulates the HATS, whereas N deprivation has the opposite effect (Crawford and Glass, 1998; Lejay et al., 1999; Zhuo et al., 1999; Gansel et al., 2001). Again, NRT2.1 plays a key role in these processes because both repression of HATS by N metabolites and its stimulation by N starvation are suppressed in the atnrt2.1-1 mutant (Cerezo et al., 2001). Another important aspect of NRT2.1 function relates to the recent reports showing that it is not only involved in high-affinity uptake, but also plays a -sensing role in the regulation of lateral root initiation (Little et al., 2005; Remans et al., 2006). This latter role is still unclear because, depending on the conditions, NRT2.1 either represses (Little et al., 2005) or, on the contrary, stimulates (Remans et al., 2006) lateral root development.

Recently, investigation of several chl1 mutants of Arabidopsis unexpectedly revealed that disruption of another transporter gene, NRT1.1 (formerly CHL1), results in a major alteration of the regulation of NRT2.1 expression by the N status of the plant (Muños et al., 2004). First, feedback repression of NRT2.1 by either or Gln supply is suppressed or strongly attenuated in chl1 mutants compared with wild type, resulting in its overexpression in normally suppressive conditions (e.g. in NH₄NO₃-fed chl1 plants). Second, expression of NRT2.1 is no more stimulated by N starvation in chl1 mutants. These data suggest that mutation of NRT1.1 blocks both NRT2.1 expression and HATS activity in some kind of derepressed state, making chl1 mutants the only known genotypes affected in the regulation of root uptake in higher plants.

NRT1.1 is an unusual dual-affinity transporter (Tsay et al., 1993; Wang et al., 1998), shifting from low to high affinity in response to phosphorylation of the Thr-101 residue (Liu and Tsay, 2003). Although NRT1.1 also contributes to root uptake (Tsay et al., 1993; Huang et al., 1996; Touraine and Glass, 1997), its precise role in the regulation of NRT2.1 expression remains unclear. A first hypothesis was that NRT1.1 may be directly involved in the signaling pathway responsible for feedback repression of NRT2.1 by N metabolites. However, this interpretation was questioned by the finding that repression of NRT2.1 by could also be alleviated in wild-type plants when the external concentration was decreased down to 0.1 mm in the presence of 1 mm (Muños et al., 2004). This showed that mutation of NRT1.1 is not required to suppress the repressive effect of N metabolites and led to a second alternative hypothesis that NRT2.1 up-regulation in NH₄NO₃-fed chl1 plants could actually be due to the lowered uptake resulting from NRT1.1 mutation. Because NH₄NO₃-fed chl1 plants were not N deficient, this in turn suggested that NRT2.1 expression is not only repressed by reduced N metabolites, but also by itself (Muños et al., 2004).

This work had three aims: (1) to investigate the occurrence of such repression of NRT2.1 by in wild-type plants; (2) to determine whether NRT1.1 plays a direct (i.e. specific) or indirect (through modulation of uptake rate) role in triggering this repression; and (3) to determine the physiological role of this regulation. Concerning this last point, we hypothesized that repression of NRT2.1 by may correspond to a mechanism allowing the HATS to be stimulated by a specific demand of the plant independently of its overall N status (Muños et al., 2004). In particular, we anticipated that this may have a crucial function to avoid that the bulk of N acquisition from mixed NH₄NO₃ medium is made through uptake, thus protecting the plant against the toxicity generally associated with nutrition (Givan, 1979; Hageman, 1984; Salsac et al., 1987; Kronzucker et al., 2001).

RESULTS

Low NO₃⁻ Availability in the Presence of NH₄⁺ Up-Regulates Both Root NRT2.1 Expression and NO₃⁻ HATS Activity

As commonly observed in many experimental conditions, the supply of together with at equimolar concentration (1 mm for both ions) is associated with very faint expression of NRT2.1 in the roots (Fig. 1A, lane T0). However, transfer of plants to fresh nutrient medium with lower concentration (0.1 mm), but unmodified availability (1 mm), rapidly resulted in a marked increase in NRT2.1 mRNA accumulation, which was detectable after 6 h and leveled off between 3 and 4 d (Fig. 1A).

Figure 1.

Changes in NRT2.1 mRNA accumulation in response to various / mixtures in the external medium. Plants were grown hydroponically for 6 weeks on complete nutrient solution containing 1 mm NH₄NO₃ before transfer to the various media indicated in the figure. A, Time-course response in plants transferred to nutrient solution containing 1 mm plus 0.1 mm . B, Effect of transfer for 4 d to media containing various combinations of and concentrations. Transcript accumulation was monitored in roots both by northern-blot analysis (gel-blot images) and real-time RT-PCR (bar graphs) on the same samples. A 25S rRNA signal is presented as a loading control for northern blots. Quantification by real-time RT-PCR was normalized using actin genes as controls. Errors bars represent se (n = 3). Different letters on top of the bars indicate statistically significant differences (P < 0.05; t test).

To identify more precisely which parameter modified by the above treatment ( concentration, total plus concentration, or to ratio) was responsible for the up-regulation of NRT2.1 expression, we explored the effects of various combinations of and concentrations. Figure 1B displays typical results obtained from this series of experiments. They show that the increase of NRT2.1 transcript accumulation is not due to the lowering of the total external N ( + ) concentration because it was also observed after transfer to solutions containing 0.1 mm supplemented with 2 mm, or even 5 mm . On the other hand, a high to ratio in the external medium is also not responsible per se for up-regulation of NRT2.1 expression because, when compared to 1:0.1 mm :, the 5:0.5 mm : mixture (with the same : ratio of 10) resulted in much less pronounced NRT2.1 up-regulation (Fig. 1B). Furthermore, NRT2.1 is also overexpressed by low availability in mixed N medium when Gln (another potent repressor of NRT2.1 expression) is provided as the reduced N source (Supplemental Fig. S1). This shows that NRT2.1 up-regulation is not specifically due to excess supply. Altogether, the data obtained suggest that NRT2.1 expression is responsive to the absolute concentration present in the mixed N solution. Indeed, whatever the nature and concentration of the reduced N source ( or Gln, 1–5 mm), the NRT2.1 transcript level was always high when the external concentration remained low at 0.1 mm, but strongly decreased when the concentration was raised at 0.5 mm.

One interesting finding related to the increased expression of NRT2.1 under low /high availability is that the effect is local and not systemic. Indeed, split-root experiments showed that when only one side of the root system is provided with 0.1 mm + 1 mm , NRT2.1 is up-regulated in this side only and not in the other one fed with 1 mm NH₄NO₃ (Fig. 2, compare lanes b and c). Furthermore, the NRT2.1 transcript level in the side of the split-root system supplied with 0.1 mm + 1 mm (Fig. 2, lanes c and e) was the same, whatever the availability on the other side (1 or 0 mm in Fig. 2, lanes b and f, respectively), and equaled that in plants homogeneously supplied with 0.1 mm + 1 mm on the whole root system (Fig. 2, lane d).

Figure 2.

Effect of local changes in the / external balance on NRT2.1 mRNA accumulation in split-root plants. A, Representation of the experimental protocol. Plants were grown hydroponically for 5 weeks on complete nutrient solution containing 1 mm NH₄NO₃. After separation of their root system into two approximately equal sides, plants were transferred to specific containers and left undisturbed during 3 d in the 1 mm NH₄NO₃ solution before application of localized supplies with different / mixtures for 4 d, as indicated in the figure. B, NRT2.1 mRNA accumulation. Transcript accumulation was monitored in various sides of the split-root systems both by northern-blot analysis (gel-blot images) and real-time RT-PCR (bar graphs) on the same samples. The lane labels relate to the various localized treatments as shown in A. A 25S rRNA signal is presented as a loading control for northern blots. Quantification by real-time RT-PCR was normalized using actin genes as controls. Errors bars represent se (n = 3). Different letters on top of the bars indicate statistically significant differences (P < 0.05; t test).

In wild-type plants, up-regulation of NRT2.1 expression by low availability in the presence of resulted in similar stimulation of HATS activity (as measured by root ¹⁵ influx at 0.2 mm), which was not observed in the atnrt2.1-1 mutant (Fig. 3A). Interestingly, the uptake system did not display this stimulation because root ¹⁵ influx measured either at 1 mm (Fig. 3B) or 0.2 mm (data not shown) was only marginally increased upon lowering of the external concentration. Total N content of roots and shoots was unaffected between the treatments (data not shown), most probably as a result of both increased HATS activity and sustained uptake.

Figure 3.

Effect of various / mixtures on NRT2.1 mRNA accumulation and root N uptake in Ws wild-type and atnrt2.1-1 mutant plants. The plants were grown hydroponically for 6 weeks on complete nutrient solution containing 1 mm NH₄NO₃ before transfer for 4 d to various media containing the same concentration (1 mm), but various concentrations of (0.1, 1, or 10 mm). A, NRT2.1 mRNA accumulation and root ¹⁵ influx. Transcript accumulation was monitored in roots both by northern-blot analysis (gel-blot images) and real-time RT-PCR (bar graphs) on the same samples. A 25S rRNA signal is presented as a loading control for northern blots. Quantification by real-time RT-PCR was normalized using actin genes as controls. Errors bars represent se (n = 3). Root ¹⁵ influx was assayed by 5-min labeling at 0.2 mm external ¹⁵ concentration to specifically determine activity of the HATS. Each value is the mean of eight to 12 replicates ± se. B, Root ¹⁵ influx. Root ¹⁵ influx was assayed by 5-min labeling at 1 mm external ¹⁵ concentration to determine the uptake activity at the same concentration as for growth and treatment of the plants. Each value is the mean of eight to 12 replicates ± se. N.D., Not determined. Different letters on top of the bars indicate statistically significant differences (P < 0.05; t test).

Collectively, the above results suggest that low availability alleviates repression of both NRT2.1 expression and HATS activity triggered by provision of a reduced N source to the plant.

Regulation of NRT2.1 Expression by External NO₃⁻ Availability in the Presence of NH₄⁺ Is Specifically Mediated by NRT1.1

In previous experiments with chl1-10 plants, it was observed that increasing concentration from 0.1 to 10 mm in the presence of 1 mm failed to repress NRT2.1 expression (Muños et al., 2004), unlike what is found in wild type (Fig. 3A). To determine whether this is due to a direct or an indirect (through lowered total uptake) role of NRT1.1 in regulating NRT2.1 expression, we checked whether the mutation of NRT1.2, encoding another important component of the low-affinity transport system (Huang et al., 1999), also resulted in NRT2.1 up-regulation. Therefore, we investigated the quantitative relationship between uptake from NH₄NO₃ medium and NRT2.1 expression in the roots of three genotypes: the Wassilewskija (Ws) wild type, the chl1-10 mutant, and the atnrt1.2-1 mutant, carrying a T-DNA insertion in the NRT1.2 gene.

Increasing the external concentration from 0.1 to 5 or 10 mm in the presence of 1 mm resulted in a strong and almost similar down-regulation of NRT2.1 expression in both Ws and atnr1.2-1 plants, whereas this down-regulation was absent in chl1-10 plants (Fig. 4A). Surprisingly, root ¹⁵ influx (measured at the same concentration as that used for treatment of the plants) was little affected by external concentration in Ws plants (Fig. 4B), where a 50-fold decrease of this concentration (from 5–0.1 mm) only resulted in a 30% slowing down of root ¹⁵ influx. This indicates strong homeostasis of root uptake from NH₄NO₃ medium, which probably explains why root influx was only marginally stimulated by the decrease in external concentration (Fig. 3B). Root ¹⁵ influx was lower in the atnrt1.2-1 mutant than in the wild type, but not specifically in the high external concentration range. Unexpectedly, root ¹⁵ influx was not reduced in chl1-10 plants compared with Ws and was even higher in the middle range of external concentration (1–5 mm). These data unambiguously demonstrate that unrepressed NRT2.1 transcript accumulation in chl1-10 plants at high external availability cannot be explained by reduced root uptake rate as compared to either Ws or atnrt1.2-1 plants (Fig. 4C). Thus, whatever the signal responsible for down-regulation of NRT2.1 expression by high external availability, it requires NRT1.1 to trigger the response.

Figure 4.

Relationship between NRT2.1 mRNA accumulation and root influx as a function of concentration in the presence of in the external medium. Plants of Ws wild-type, atnrt1.2-1 mutant, and chl1-10 mutant were grown for 6 weeks on complete nutrient solution containing 1 mm NH₄NO₃, before transfer for 4 d to the nutrient solutions containing 1 mm and at various concentrations, as indicated in the figure. A, NRT2.1 mRNA accumulation. Transcript accumulation was monitored in roots both by northern-blot analysis (gel-blot images) and real-time RT-PCR (graph) on the same samples. A 25S rRNA signal is presented as a loading control for northern blots. Quantification by real-time RT-PCR was normalized using actin genes as controls. Errors bars represent se (n = 3). Differences between chl1-10 and the other genotypes are statistically significant at **P < 0.01 and ***P < 0.001 (t test). B, Root ¹⁵ influx. Root ¹⁵ influx was assayed by 5-min labeling at the external ¹⁵ concentration corresponding to that supplied to the plants during the 4-d treatment. Each value is the mean of six to 12 replicates ± se. C, Plot of NRT2.1 mRNA accumulation (A) against root ¹⁵ influx (B).

Additional evidence for NRT1.1-dependent control of the HATS by external availability was provided by expression analysis of NRT3.1, which shows the same response pattern as NRT2.1, suggesting coregulation of the two genes (Fig. 5). However, the quantitative changes in NRT3.1 transcript abundance were generally less pronounced than those of NRT2.1.

Figure 5.

NRT2.1 and NRT3.1 coregulation. A, Response of NRT3.1 mRNA accumulation in Ws, atnrt1.2-1, and chl1-10 plants to the variation of concentration in the presence of 1 mm in the external medium. The experiment is the same as in Figure 5. B, Correlation between changes in NRT2.1 and NRT3.1 mRNA accumulation. All RNA samples used to determine changes in NRT2.1 expression presented in Figures 1 to 4 and Supplemental Figure S1 were analyzed for NRT3.1 mRNA accumulation. The figure presents the relative changes in transcript level for both NRT2.1 and NRT3.1 normalized to the control (Ws plants left on 1 mm NO₃NH₄) of each experiment. Gray squares, Experiment of Figure 1; white diamonds, experiment of Figure 2; gray triangles, experiment of Figure 3; white hexagons, experiment of Supplemental Figure S1; black circles, experiment of Figure 4.

Taken together, the above data strongly support the hypothesis of coordinated regulation of NRT2.1 and NRT3.1, mediated by NRT1.1, which down-regulates the HATS when external availability in mixed N medium is above 0.2 to 0.5 mm. Below this threshold, this repression is alleviated, which enables the plant to reactivate the HATS, despite the strongly repressive conditions related to the ample supply of reduced N source.

Up-Regulation of NRT2.1 by Low NO₃⁻ Availability in the Presence of NH₄⁺ Prevents Growth Inhibition Associated with NH₄⁺ Toxicity

To determine the physiological significance of NRT1.1-dependent control of the HATS by external availability, we investigated the hypothesis that up-regulation of NRT2.1 by low concentration in mixed N medium plays a role in preventing toxicity under situations of excess over supply. Therefore, we analyzed growth and uptake of wild-type plants and of two independent NRT2.1 knockout mutants (atnrt2.1-1 and atnrt2.1-2) under pure or mixed NH₄NO₃ nutrition. Seedlings were first grown for 5 weeks on 1 mm NH₄NO₃ before submitting them for 10 to 14 d to the various N nutrition regimes. During these experiments, wild-type and mutant plants were placed in the same container to make sure that both genotypes experienced the same changes in external pH.

As expected, the supply of 1 mm as the sole N source led to the appearance of toxicity symptoms in the shoots of both wild-type and atnrt2.1-1 genotypes (Fig. 6). These symptoms became pronounced between 6 and 10 d after transfer to the nutrient solution. In particular, -fed plants started to bolt very early and their leaves wilted and yellowed. Interestingly, addition of 0.1 mm in the 1 mm nutrient solution fully prevented the appearance of toxicity symptoms in wild-type plants, but not in the mutant, which seemed to remain as sensitive as when is the sole N source supplied (Fig. 6).

Figure 6.

Effect of NRT2.1 mutation on the protective effect of against toxicity. Images show the appearance of toxicity symptoms in shoots of Ws and atnrt2.1-1 mutant plants as a function of the N source supplied in the external medium (either 1 mm alone or in combination with 0.1 mm ).

These observations were confirmed by shoot growth analysis (Fig. 7A). With 1 mm as the sole N source, shoots of atnrt2.1-1 plants grew at the same relative rate as those of wild-type plants, despite slightly lower biomass at the beginning of the experiment. Relative growth rate (RGR) values (determined from the slopes of the linear relationships between ln [fresh weight] and time) were 0.126 g g⁻¹ d⁻¹ (r² = 0.986) and 0.125 g g⁻¹ d⁻¹ (r² = 0.975) for wild-type and atnrt2.1-1 shoots, respectively. Supply of 0.1 mm together with 1 mm markedly stimulated shoot growth in wild-type plants (RGR = 0.165 g g⁻¹ d⁻¹; r² = 0.992), but not in atnrt2.1-1 plants (RGR = 0.137 g g⁻¹ d⁻¹; r² = 0.988). In agreement with the fact that NRT2.1 is up-regulated by low concentration in mixed N medium, and that it encodes a major component of the HATS, the cumulative uptake in wild-type plants transferred on 0.1 mm + 1 mm was much higher than that in atnrt2.1-1 plants (Fig. 7B). All the above observations were confirmed with the atnrt2.1-2 mutant allele (Supplemental Fig. S2). On 1 mm NH₄NO₃, however, a situation where NRT2.1 has a low contribution to total uptake (Cerezo et al., 2001), no significant difference was recorded for both shoot biomass and cumulative uptake between wild-type and atnrt2.1-1 plants (Supplemental Fig. S3). Thus, the decrease in shoot growth observed in atnrt2 mutants as compared to wild types correlated with the reduction of uptake resulting from the NRT2.1 mutation. On the other hand, root biomass was not affected by NRT2.1 mutation in all media investigated (Supplemental Figs. S2 and S3).

Figure 7.

Shoot growth and root uptake of Ws and atnrt2.1-1 mutant plants supplied with 1 mm with or without 0.1 mm . Plants were grown hydroponically for 5 weeks on complete nutrient solution containing 1 mm NH₄NO₃ before transfer for 12 d to media containing either 1 mm alone or 1 mm plus 0.1 mm ¹⁵. A, Shoot fresh weight. Values are the means of 10 to 18 replicates ± se. B, Cumulative ¹⁵ uptake from the 1 mm plus 0.1 mm ¹⁵ solution during the 12-d period determined by total ¹⁵N analysis in both root and shoots at the end of the period. Values are the mean of 10 to 12 replicates ± se. Differences between wild type and mutant are statistically significant at **P < 0.01, ***P < 0.001 (t test). Ns, Not significant (P > 0.05).

Taken together, the above data demonstrate that, even at a low concentration of 0.1 mm, the presence of in mixed N solution is able to alleviate the detrimental effects of pure nutrition on shoot growth in wild-type plants, but not in atnrt2.1 mutants. This indicates that the protective action of 0.1 mm against toxicity is dependent on NRT2.1 expression.

DISCUSSION

A Novel Regulation of NRT2.1 Expression Involving NRT1.1-Mediated Repression by High NO₃⁻

Knowledge concerning the control of NRT2.1 expression by N is that this gene is under two main regulations, namely (1) induction by ; and (2) repression by high N status of the whole plant (Filleur and Daniel-Vedele, 1999; Lejay et al., 1999; Zhuo et al., 1999; Forde, 2000; Gansel et al., 2001; Orsel et al., 2002; Nazoa et al., 2003; Okamoto et al., 2003). The molecular mechanisms underlying these regulations are unknown, but there is some consensus about the nature of the signal molecules involved. In particular, the ion itself is believed to be the inducer (Crawford and Glass, 1998; Forde, 2000), and and Gln are thought to be the main signal molecules involved in the feedback repression exerted by the N status of the plant (Lejay et al., 1999; Zhuo et al., 1999; Nazoa et al., 2003).

This model now appears to be incomplete because NRT2.1 overexpression in wild-type plants under low /high availability (Figs. 1–4) cannot be explained by the above mechanisms (i.e. induction by and repression by ). A striking illustration of this is the observation that transfer of the plants from 1 mm NH₄NO₃ to 5 mm + 0.1 mm led to 4-fold stimulation of NRT2.1 expression (Fig. 1B), although this corresponded to a 10-fold decrease in the concentration of the inducer () and a 5-fold increase in the concentration of the repressor (). Furthermore, as already shown for up-regulation of NRT2.1 in NH₄NO₃-fed chl1 mutants (Muños et al., 2004), up-regulation of NRT2.1 by low in the presence of cannot be mistaken with relief of the feedback repression exerted by the overall N status of the plant: (1) plants subjected to low /high availability have a high total N content of the tissues and thus have high N status; (2) root influx is not up-regulated in these plants (Fig. 3B), whereas it is also under negative feedback control by the N status of the plant (Gazzarrini et al., 1999; Rawat et al., 1999; von Wirén et al., 2000); and (3) stimulation of NRT2.1 expression by low /high availability is under purely local control (Fig. 2), whereas NRT2.1 regulation by the N status of the plant involves systemic signaling (Imsande and Touraine, 1994; Gansel et al., 2001; Forde, 2002).

Altogether, our data provide evidence that NRT2.1 expression is also modulated by a third important regulatory mechanism, triggering repression of this gene by high external availability, which superimposes on repression exerted by or Gln. Indeed, in the presence of either reduced N source, NRT2.1 expression in wild-type roots was consistently found to be primarily determined by external concentration, with strong down-regulation as soon as this concentration exceeded the 0.2 to 0.5 mm range. Although surprising at first glance, the hypothesis that may have opposite regulatory effects (induction and repression) is already well documented for its role in the control of lateral root growth (local stimulation of lateral root elongation and systemic repression of lateral root emergence; Zhang et al., 1999). Furthermore, down-regulation of root uptake by itself has already been postulated, in particular from barley (Hordeum vulgare) experiments where root uptake was found to be negatively correlated with root concentration, but only when root concentration exceeded a certain threshold level (Siddiqi et al., 1989; Crawford and Glass, 1998). Although these physiological studies provided circumstantial evidence for repression of root uptake by high , our results bring insight to these aspects because they highlight NRT2.1 and NRT3.1 as molecular targets of this regulation. Recently, NRT3.1 expression was also shown to be induced by (Okamoto et al., 2006) and repressed by high N status of the plant (Remans et al., 2006). Thus, NRT3.1 appears to be, at least partially, controlled by the same regulatory network as NRT2.1, suggesting coordinated regulation of these two components of the HATS. Most importantly, our data also reveal specific involvement of NRT1.1 in triggering repression by high (Figs. 4 and 5). Indeed, down-regulation of NRT2.1/NRT3.1 expression by high availability in the presence of is fully suppressed in the chl1-10 mutant (and not in a nrt1.2 mutant), whereas root uptake is not reduced in chl1-10 compared to wild type (whereas it is reduced in atnrt1.2-1). This clearly invalidates one of our initial hypotheses that this phenotype of chl1 mutants is simply a compensatory response to a general defect in acquisition and thus an indirect consequence of NRT1.1 mutation. Our proposal for the regulatory role of NRT1.1 in wild-type plants is that the increase in external concentration results in an increasing activity of this transporter, which in turn generates an increasing repressive signal for NRT2.1 expression (Fig. 4A). Whether this indicates a direct signaling function for NRT1.1 (in analogy with the role of the sensor recently proposed for NRT2.1) and calls for the specific involvement of one isoform of NRT1.1 (high or low affinity) are open questions that deserve further investigation.

To account for our observations, we propose a model for N regulation of NRT2.1 expression (Fig. 8). In addition to the positive regulation corresponding to the induction by , this model postulates dual negative regulation involving both feedback repression by reduced N metabolites and NRT1.1-mediated repression by high external . An important point is that the absence of NRT1.1-mediated repression (due to mutation of NRT1.1 or to low external availability) overrides the negative feedback exerted by reduced N metabolites to yield a high NRT2.1 expression level even in the presence of ample supply to the plant (e.g. 5 mm; see Fig. 1B). Conversely, there is also evidence that lack of negative feedback regulation by reduced N metabolites overrides the repressive effect of high external availability. This is shown by the high NRT2.1 expression level in nitrate reductase-deficient plants supplied with as a sole N source (Lejay et al., 1999; Zhuo et al., 1999). Taken together, these observations suggest that NRT2.1 expression is suppressed only when both negative regulations by reduced N metabolites and by high are effective (as illustrated in Fig. 8). Whether this means that the two respective signaling pathways directly interfere at some common crucial node or, alternatively, that they are independently strong enough to overcome each other is not known. However, because NRT1.1 mutation prevents NRT2.1 repression by high , but does not alter its reinduction by after a period of N starvation (Muños et al., 2004), it is concluded that these opposite actions of most probably involve independent signaling pathways.

Figure 8.

Model for N regulation of NRT2.1 expression in Arabidopsis roots. The model postulates that, in addition to the induction by , NRT2.1 is also under dual control by feedback repression by reduced N metabolites and NRT1.1-mediated repression by high external availability, which are both required to suppress NRT2.1 expression.

NRT1.1-Mediated Regulation of NRT2.1 Allows an Adaptive Response of the Plant to NH₄⁺ Toxicity

A key issue concerning NRT1.1-mediated regulation of the HATS by high is to determine what physiological role such a mechanism may play. In analogy with the well-accepted postulate that repression of root (or ) uptake systems by reduced N metabolites corresponds to a regulation by the N demand of the plant (Imsande and Touraine, 1994; vonWirén et al., 2000; Forde, 2002), we hypothesize that repression of NRT2.1 by high corresponds to regulation by a demand of the plant. Accordingly, relief of this NRT1.1-mediated repression due either to decreased availability in the presence of , or to NRT1.1 mutation activated the HATS but not the uptake system (Fig. 3; Muños et al., 2004). This shows that root uptake, and not total root N uptake, is the specific target of this mechanism. Clearly, what the plant perceives under these situations is a lack of and not an overall nutritional N deficiency.

In this context, the significance of the model depicted in Figure 8 is that repression of NRT2.1 by reduced N metabolites in N-sufficient plants is allowed only when NRT1.1 is active in transporting . This warrants that a significant uptake rate is always ensured in any situation, either by NRT1.1 or by NRT2.1, even when accounts for only a minor fraction of the total N available in the external medium (Fig. 4B). One of the most obvious interests of such a mechanism is to protect the plant against toxicity. It is known for decades that pure nutrition is toxic for many plant species (Givan, 1979; Hageman, 1984; Salsac et al., 1987; Kronzucker et al., 2001). In particular, growth of herbaceous dicotyledons such as tomato (Lycopersicon esculentum), French bean (Phaseolus vulgaris), spinach (Spinacia oleracea), and here Arabidopsis, is generally strongly hampered by pure nutrition, with a decrease in yield of up to 60% as compared with supply of as the sole N source (Salsac et al., 1987). A general observation is, however, that toxicity is fully prevented by supply of as a N source together with (Cox and Reisenauer, 1973; Kronzucker et al., 1999; Rahayu et al., 2005). Actually, highest growth rates are generally achieved with mixed + supplies (Cox and Reisenauer, 1973; Heberer and Below, 1989; Adriaanse and Human, 1993; Cao and Tibbitts, 1993).

Despite this firmly established role of in preventing the detrimental effects of nutrition, a strong paradox remained unresolved. On the one hand, uptake by the plant was shown to ensure full protection against toxicity and, on the other hand, uptake systems were shown to be strongly repressed by the supply of high concentration to the plant. To date, no mechanism was known to stimulate uptake in the presence of potentially toxic concentrations of in the external medium. We propose that the NRT1.1-mediated regulation of NRT2.1/NRT3.1 corresponds to such a mechanism because it relieves repression of the HATS under low /high availability (Figs. 3A and 4). Furthermore, the phenotype of both atnrt2.1 mutants demonstrates that up-regulation of the HATS under this condition constitutes an essential adaptive response of the plant to avoid toxicity (actually the only one documented at the molecular level; Figs. 6 and 7; Supplemental Figs. S2 and S3).

A surprising aspect of the HATS repression by high is that it seems to rely on purely local signaling because only the portions of the root system subjected to low /high availability react in up-regulating NRT2.1 expression (Fig. 2). Furthermore, high supply on one portion of the root system does not prevent the adaptive response of NRT2.1 in other portions fed with excess over (see Fig. 3, lanes b and c). This suggests that the demand governing NRT2.1 expression is not sensed at the whole-plant level and that the adaptive response of NRT2.1 aims at stimulating uptake specifically in the root cells experiencing high external availability. This is in full agreement with the results from split-root experiments on maize (Zea mays) and soybean (Glycine max), indicating that plays its protective role against toxicity only when it is locally supplied together with , and not when the two N sources are separately provided to only one half on the root sy stem (Schortemeyer et al., 1993; Saravitz et al., 1994).

Our data thus show that, besides its key role in ensuring the bulk of N acquisition by the plant in many various environmental conditions, NRT2.1 also plays a critical function in maintaining a healthy balance between and uptake. It is highlighted that, in this latter case, NRT2.1 activity is not required to supply an N source for amino acid synthesis, but to allow the plant to benefit from a specific role of that cannot fulfill. This illustrates very well why NRT2.1 cannot be regulated only by feedback repression by N metabolites because this regulation aims at adjusting N uptake to amino acid utilization and is not specific for uptake systems (vonWirén et al., 2000). Interestingly, up-regulation of NRT2.1 expression by low availability was also observed in the presence of Gln (Supplemental Fig. S1), suggesting that demand signaling may be operative under other circumstances than those associated with toxicity. It is thus tempting to postulate a more general role for this signaling in regulating acquisition by the plant. Nitrate is not only a nutrient, but also a key signaling compound governing crucial aspects of plant metabolism and development (Crawford, 1995; Stitt, 1999). In particular, regulates many genes related to N or C metabolism (Crawford, 1995; Stitt, 1999), triggers several adaptive responses of root and shoot growth (Forde, 2002; Walch-Liu et al., 2005), and modulates cytokinin signaling (Sakakibara, 2003). Thus, plants may have evolved specific regulatory mechanisms to tightly control these important signaling effects of . At the uptake level, this would require a regulatory mechanism specific for transporters and independent from the feedback regulation by reduced N metabolites, which aims at ensuring efficient use of this ion (as well as of ) as a nutrient. In this context, the model of Figure 8 corresponds to an elegant mechanism for integrating both requirements for as a nutrient and as a signal in the regulation of root uptake. Therefore, the question of whether NRT1.1-mediated regulation of NRT2.1 reported in this work has additional functions other than just protecting the plant from toxicity deserves further investigation. In particular, given the role of NRT1.1 in regulating NRT2.1 expression, and the role of NRT2.1 in controlling lateral root initiation, it will be of interest to investigate whether putative signaling mediated by NRT1.1 is also involved in modulating root branching as a function of external availability.

MATERIALS AND METHODS

Plant Material and Treatments

The Arabidopsis (Arabidopsis thaliana) genotypes used in this study were the wild-type Ws and Columbia-0 ecotypes; the atnrt2.1-1 mutant in the Ws background (formerly atnrt2a), obtained from the collection of Institut National de la Recherche Agronomique, Versailles, and deleted for the NRT2.1 (At1g08090) and NRT2.2 (At1g08100) genes (Filleur et al., 2001); the atnrt2.1-2 mutant in the Columbia-0 background, obtained from the Salk Institute (SALK_035429), and carrying a T-DNA insertion in the first intron of NRT2.1 (these two mutants were renamed according to the nomenclature proposed by Little et al., 2005); and the atnrt1.2-1 mutant in the Ws background, obtained from the collection of the Institut National de la Recherche Agronomique, and carrying a T-DNA insertion in the third intron of NRT1.2. NRT1.2 mRNA was not detected by reverse transcription (RT)-PCR in the roots of this mutant (data not shown).

Plants were grown for 6 weeks in hydroponics under nonsterile conditions, as previously described by Lejay et al. (1999). The growth chamber was set with the following environmental conditions: 8-h light/16-h dark 22°C/20°C temperature, 250 μmol m⁻² s⁻¹ irradiance, and 70% hygrometry. Briefly, seeds were sown directly on sand contained by a cut 1.5-mL Eppendorf tube closed at the bottom by a stainless grid. Tubes were supported by PVC discs (six Eppendorf/disc) placed on a floating polystyrene raft (12 discs/raft). These systems were disposed on top of 10-L tanks filled with tap water for the first week, and then with nutrient solution for 4 to 5 additional weeks (during this period, nutrient solutions were renewed weekly). The basal nutrient solution common to all experiments included 1 mm KH₂PO₄, 1 mm MgSO₄, 0.25 mm K₂SO₄, 0.25 mm CaCl₂, 0.1 mm FeNa-EDTA, 50 μm KCl, 30 μm H₃BO₃, 5 μm MnSO₄, 1 μm ZnSO4, 1 μm CuSO₄, and 0.1 μm (NH₄)₆Mo₇O₂₄. For growth of the plants, 1 mm NH₄NO₃ was added to the basal medium as the N source. Depending on the experiments, 1 mm NH₄NO₃ was replaced as a N source by either KNO₃ or NH₄Cl, or various mixtures of these salts, as indicated in the text and figures. The pH of all solutions was adjusted to 5.8, and the solutions were renewed every other day during the experiments to prevent nutrient depletion. For experiments with media at low concentration (0.1 mm), nutrient solutions were renewed daily, which allowed maintenance of the external concentration above 0.06 to 0.07 mm. For treatments with Gln, 25 mg L⁻¹ chloramphenicol and 50 mg L⁻¹ penicillin were added to solutions to prevent microbial development. For time-course studies, various treatments were initiated at different times to allow harvest of all plants at the same time of the day (7–8 h into the light period) to prevent any diurnal effect on NRT2.1 expression (Lejay et al., 2003).

For split-root experiments, the protocol was reported previously (Gansel et al., 2001). At the age of 2 weeks, seedlings were cleared to leave only one plant per tube. After gentle separation of the root system into two approximately equal portions, 5-week-old plants were transferred to specific containers and allowed to adapt 3 d to split-root conditions, with the two parts of the root system supplied with the 1 mm NH₄NO₃ solution. The various treatments were then initiated at the end of this period.

For growth analysis, roots and shoot were separated after harvest and their fresh weight determined. Fresh weight data are the means of 10 to 18 replicates.

RNA Extraction and RNA Gel-Blot Analysis

RNA extraction was performed as previously described (Lobreaux et al., 1992) from eight to 12 plants per treatment (except for six plants in the split-root experiment). Ten micrograms of total RNA were then separated by electrophoresis on MOPS-formaldehyde agarose gel and blotted on nylon membrane (Hybond N⁺; Amersham-Pharmacia Biotech). Membranes were prehybridized for 2 h at 60°C in church buffer: 0.5 m NaHPO₄, 1% bovine serum albumin, and 7% SDS (pH 7.2 with H₃PO₄). Hybridization was performed overnight at 60°C after addition of a randomly primed ³²P-labeled cDNA probe in the hybridization buffer. Membranes were washed twice at root temperature for 2 min and twice at 60°C with 0.5× SSC, 0.1% SDS. The probe used in this study corresponds to the full-length of AtNRT2.1 cDNA (Filleur and Daniel-Vedele, 1999). A 25S rRNA probe was used to normalize quantifications achieved using a phosphor imager (BAS-5000; Fujifilm).

Quantitative RT-PCR

Ten to 15 μg of total RNA were digested by RQ-DNase (Promega). After phenol-chloroform purification and isopropanol precipitation, RNA was reverse transcribed to one-strand cDNA using Moloney murine leukemia virus reverse transcriptase (Promega) and dT(18) V primers, according to the manufacturer's protocol. Gene expression was determined by quantitative real-time PCR (LightCycler; Roche Diagnostics) using gene-specific primers: NRT2.1 (forward, 5′-aacaagggctaacgtggatg-3′ and reverse, 5′-ctgcttctcctgctcattcc-3′); NRT3.1 (forward, 5′-ggccatgaagttgcctatg-3′ and reverse, 5′-tcttggccttcctcttctca3-′); ACT2/8 (forward, 5′-ggtaacattgtgctcagrggtgg-3′ and reverse, 5′-aacgaccttaatcttcatgctgc-3′), and LightCycler FastStart DNA Master SYBR Green (Roche Diagnostics). Expression levels of tested genes were normalized to expression levels of the ACT2/8 genes (Charrier et al., 2002).

¹⁵NO₃⁻ and ¹⁵NH₄⁺ Uptake

Influx of either ¹⁵ or ¹⁵ into the roots was assayed as described previously (Lejay et al., 1999) by 5-min labeling in basal nutrient medium (pH 5.8) supplemented with appropriate concentrations of K¹⁵NO₃ or ¹⁵NH₄Cl (atom % ¹⁵N excess: 99%). For specific determination of the activity of the HATS, ¹⁵ or ¹⁵ were at 0.2 mm in the labeling solution. Cumulative uptake during long-term growth studies (10–12 d) was assayed by supplying the plants with nutrient solution containing ¹⁵ (atom % ¹⁵N excess: 1%) for the whole experimental period and by measuring total ¹⁵N accumulation in roots and shoots at the end of this period. Each influx or cumulative uptake value is the mean of eight to 12 replicates.

The total N content and atomic percentage ¹⁵N abundance of the samples were determined by continuous-flow mass spectrometry, as described previously (Clarkson et al., 1996), using a Euro-EA Eurovector elemental analyzer coupled with an IsoPrime mass spectrometer (GV Instruments).

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Up-regulation of NRT2.1 in the presence of Gln.

• Supplemental Figure S2. toxicity in the atnrt2.1-2 mutant.

• Supplemental Figure S3. toxicity in the atnrt2.1-1 mutant.