AtAMT1 gene expression and NH4+ uptake in roots of Arabidopsis thaliana: evidence for regulation by root glutamine levels


  • Suman R. Rawat,

    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, British Columbia V6T 1Z4, Canada, and
    Search for more papers by this author
  • Salim N. Silim,

    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, British Columbia V6T 1Z4, Canada, and
    2. Department of Plant Sciences, University of Western Ontario, London, Ontario N6A 5B7, Canada
    Search for more papers by this author
  • Herbert J. Kronzucker,

    1. Department of Plant Sciences, University of Western Ontario, London, Ontario N6A 5B7, Canada
    Search for more papers by this author
  • M. Yaeesh Siddiqi,

    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, British Columbia V6T 1Z4, Canada, and
    Search for more papers by this author
  • Anthony D. M. Glass

    Corresponding author
    1. Department of Botany, University of British Columbia, 6270 University Boulevard, Vancouver, British Columbia V6T 1Z4, Canada, and
    Search for more papers by this author

*For correspondence (fax +604 822 6089; e-mail


The mechanisms involved in regulating high-affinity ammonium (NH4+) uptake and the expression of the AtAMT1 gene encoding a putative high-affinity NH4+ transporter were investigated in the roots of Arabidopsis thaliana. Under conditions of steady-state nitrogen (N) supply, transcript levels of the AtAMT1 gene and Vmax values for high-affinity 13NH4+ influx were inversely correlated with levels of N provision. Following re-supply of NH4NO3 to N-starved plants, AtAMT1 mRNA levels and 13NH4+ influx declined rapidly but remained high when the conversion of NH4+ to glutamine (Gln) was blocked with methionine sulfoximine (MSX). This result demonstrates that end products of NH4+ assimilation, rather than NH4+ itself, are responsible for regulating AtAMT1 gene expression. Consistent with this hypothesis, AtAMT1 gene expression and NH4+ influx were suppressed by provision of Gln alone, or together with NH4NO3 plus MSX. Furthermore, AtAMT1 transcript levels and 13NH4+ influx were negatively correlated with root Gln concentrations, following re-supply of N to N-starved plants. In addition to this level of control, the data suggest that high cytoplasmic [NH4+] may inhibit NH4+ influx.


Inorganic nitrogen (N) uptake by plant roots is subject to strict regulation according to whole plant demand (Crawford & Glass 1998). There is widespread agreement that rates of NO3 and/or NH4+ uptake are determined by negative feedback from accumulated N (Glass & Siddiqi 1995; Imsande & Touraine 1994; Lee & Rudge 1986; Morgan & Jackson 1988). However, there is a lack of consensus regarding the identity of the cellular N pool(s) responsible for initiating this control. A co-ordinated regulation of all forms of nitrogen transport by feedback from downstream metabolites of inorganic N was proposed by Lee & Rudge (1986). The operation of a carbon/nitrogen sensor in microorganisms, the so-called PII protein, that regulates both glutamine synthetase and NH4+ transport (Arcondeguy et al. 1997; Magasanik 1988) is consistent with this proposal. Recently, Hsieh et al. (1998) have demonstrated the presence of a putative PII homologue in Arabidopsis thaliana.

Cooper & Clarkson (1989) proposed that the cycling of amino acids between shoots and roots was responsible for integrating and regulating N uptake. However, others have proposed that NO3, NH4+ and/or various amino acids may regulate inorganic N influx (see Glass & Siddiqi 1995 for review). Furthermore, details of the underlying molecular mechanisms responsible for these putative feedback loops are rudimentary. Such details will facilitate attempts to improve nutrient uptake efficiency in transgenic crop plants, overexpressing particular ion transporters.

The cloning of genes that encode the NO3 and NH4+ transporters (Ninnemann et al. 1994; Trueman et al. 1996) now makes it possible to evaluate the mechanism(s) responsible for these control systems at the molecular level. Here we focus on a putative high-affinity NH4+ transporter. The importance of NH4+ as a direct source of N for plant growth has been substantially underestimated. In many soils, NO3 is undetectable, and NH4+ and amino acids represent the main sources of N for plant nutrition (Glass & Siddiqi 1995; Kielland 1994; Stark & Hart 1997). Furthermore, in agricultural soils where both NO3 and NH4+ are present, NH4+ strongly inhibits NO3 uptake (Aslam et al. 1996; Lee & Drew 1989). Thus, the uptake of NH4+ and particularly the regulation of this uptake at the physiological and molecular levels warrants much greater attention.

Genes encoding NH4+ transport systems have been isolated and characterized in Saccharomyces cerevisiae and other microorganisms (Arcondeguy et al. 1997; Dubois & Grenson 1979; Marini et al. 1994; Marini et al. 1997). Three such genes, namely MEP1, MEP2 and MEP3, which are thought to encode two high-affinity transport systems and a low affinity-system, respectively, have been cloned from S. cerevisiae (Dubois & Grenson 1979; Marini et al. 1994; Marini et al. 1997). In higher plants, a putative high-affinity NH4+ transporter gene (AtAMT1) was cloned from A. thaliana by heterologous complementation of the mep1-1 mep2-1 NH4+ transport mutant of S. cerevisiae (Ninneman et al. 1994). The predicted sequence of the AtAmt1 transporter was homologous with the Mep1 NH4+ transport protein from S. cerevisiae. Furthermore, when expressed heterologously in the mep1-1 mep2-1 double mutant, it functioned as a saturable high-affinity transporter capable of restoring 14C-methylamine uptake in the yeast mutant.

A root-specific homologue (LeAMT1) has also been isolated from Lycopersicon esculentum (Lauter et al. 1996). In addition to the AMT1 gene homologues, a peribacteroid NH4+ channel encoded by the GmSAT1 gene transports NH4+ across the peribacteroid membrane from the bacteroid to the host plant in soybean (Kaiser et al. 1998). This gene has no homology with the AtAMT1 gene.

The paucity of information concerning the regulation of AtAMT1 gene expression, together with the lack of agreement regarding the control of NH4+ influx from physiological studies, prompted our investigation of the putative regulatory mechanism(s). Preliminary results using heterologous expression of the AtAMT1 gene in S. cerevisiae indicated that NH4+ transport in cells grown in the presence of NH4+ was not subject to down-regulation, and the authors proposed the measurement of AMT1 expression directly in plants subjected to different nutritional conditions (Ninnemann et al. 1994). Here we report the results of such a study of AtAMT1 gene expression integrated with physiological studies of 13NH4+ influx and biochemical analyses of tissue N pools in the roots of A. thaliana. These results demonstrate that AtAMT1 transcript levels are not regulated by tissue NH4+ concentrations, but rather by Gln. Nevertheless, high levels of accumulated NH4+ may exert inhibitory effects, possibly post-translational or allosteric effects, at the level of the high-affinity transport protein.


Transmembrane fluxes and compartmentation of NH4+

Standard compartmental analyses (Kronzucker et al. 1995; Wang et al. 1993a) were undertaken using intact roots to develop protocols for estimating plasma membrane 13NH4+ influx and cellular turnover rates for NH4+. Typically, approximately 98% of 13N effluxing from the cytoplasmic compartment of 13NH4+-labeled Arabidopsis roots consisted of 13NH4+. Semi-log plots of this 13NH4+ efflux against time generated t0.5 values for cell wall and cytoplasmic 13NH4+ exchange of 39 ± 3 sec and 12.6 ± 0.5 min, respectively, in roots of plants provided with 100 μm NH4NO3. Corresponding values for 1 mm NH4NO3-grown plants were not significantly different, as shown in Table 1. These values are similar to those obtained from previous studies (Kronzucker et al. 1995; Presland & McNaughton 1986; Wang et al. 1993a). Flux values (Table 1) calculated from efflux analysis indicated a high value for plasma membrane NH4+ influx (5.2 ± 1.6 μmol g–1 FW h–1) and net uptake (4.5 ± 1.37 μmol g–1 FW h–1) into roots of plants acclimated in 100 μm NH4NO3. Efflux values (0.7 ± 0.2 μmol g–1 FW h–1) were relatively low, while cytoplasmic NH4+ concentrations were estimated to be 22 ± 4 mm. Cytoplasmic values for [NH4+] were in the same range as those estimated by NMR using maize roots (Lee et al. 1992). Steady state growth at 1 mm NH4NO3 increased influx, efflux, cytoplasmic [NH4+] and the ratio of efflux to influx, as observed in previous studies (Kronzucker et al. 1995; Wang et al. 1993a).

Table 1.  NH4+ fluxes (μmol g–1 FW h–1) and cytoplasmic [NH4+] (μmol g–1 FW) derived from compartmental analysis using roots of A. thaliana
Parameter100 μm plants1 mm plants
  1. Plants were grown in 100 μm (n = 6) or 1 mm (n = 11) NH4NO3. φoc: influx of NH4+ across the plasma membrane from cell wall to cytosol, φco: efflux across the plasma membrane from cytosol to cell wall; φnet: net flux across the plasma membrane from cell wall to cytosol, [NH4+]cyt: cytoplasmic [NH4+]. Fluxes were measured under steady state conditions.

φoc5.2 ± 1.69.5 ± 2.1
φco0.7 ± 0.23.3 ± 1.3
φnet4.5 ± 1.46.2 ± 1.3
t0.5 (cell wall)39 ± 3 s36 ± 5 s
t0.5 (cytoplasm)12.6 ± 0.5 min10.8 ± 1.1 min
[NH4+]cyt22 ± 4 mm43 ± 8 mm

The high-affinity NH4+ transport system (HATS): regulation of AtAMT1 mRNA levels and NH4+ influx by N supply

Northern analyses of root AtAMT1 gene expression in plants maintained in 100 μm NH4NO3 revealed an abundant 1.8 kb transcript (Fig. 1a), whose intensity was reduced to approximately 7% of this value by provision of 1 mm or 10 mm NH4NO3. Short-term measures of high-affinity 13NH4+ influx revealed a hyperbolic pattern (Fig. 1b), with Km and Vmax values (derived from direct fits to the Michaelis–Menten equation) of 168 μm and 17 μmol g–1 FW h–1, respectively, for plants previously grown in 100 μm NH4NO3. Vmax values were reduced to 6.1 and 3.1 μmol g–1 FW h–1 in plants maintained in 1 mm and 10 mm NH4NO3. Statistical testing of the differences among Km and Vmax values by F-tests of the slopes of Hoffstee and Hanes plots, respectively, revealed significant differences only among Vmax values.

Figure 1.

Accumulation of AtAMT1 mRNA and short-term 13NH4+ influx (μmol g–1 FW h–1) into roots of A. thaliana as a function of prior N provision.

(a) Northern blot analysis of root total RNA, probed with the AtAMT1 cDNA. RNA molecular size standards are indicated on the left. The same Northern blot was reprobed with At25S rDNA, used as internal standard.

(b) 13NH4+ influx into roots of plants grown in 0.1 mm, 1 mm and 10 mm NH4NO3. 13NH4+ influx values are means ± SEM (n = 4). Each flux experiment was repeated three times.

Time-dependent up-regulation and down-regulation of AtAMT1 mRNAs and high-affinity NH4+ influx on withdrawal or resupply of NH4NO3Transfer of plants from 1 mm to 100 μm NH4NO3 solutions resulted in rapid increases of AtAMT1 mRNA expression levels (Fig. 2a,b), which by 24 h had increased 7.5-fold relative to those at time 0. High-affinity 13NH4+ influx (Fig. 2b) also increased continuously, although lagging behind the changes of AtAMT1 mRNA levels. By 24 h, 13NH4+ influx had increased 12-fold from 0.36 μmol g–1 FW h–1 at the time of transfer to 4.47 μmol g–1 FW h–1. For purposes of comparison, Fig. 2(b) shows the changes of AtAMT1 mRNA levels and 13NH4+ influx during the 24 h treatment as percentages of their values at 24 h.

Figure 2.

Time-dependent up-regulation of AtAMT1 mRNA levels and 13NH4+ influx (μmol g–1 FW h–1) in roots of A. thaliana, following the transfer of 3-week-old plants from 1 mm NH4NO3 to 0.1 mm NH4NO3.

(a) Northern blot analysis of root RNA, probed with the AtAMT1 cDNA. Internal standard was shown in Fig. 1 

(b) AtAMT1 expression levels (○) and 13NH4+ influx (●), expressed as percentages of their maximum values.

Down-regulation of 13NH4+ influx was examined by transferring plants, previously acclimated in 1 mm NH4NO3, to N-free media for 2 days to minimize internal N reserves and increase NH4+ influx. These plants were then exposed to 5 mm NH4NO3 for intervals of up to 24 h. Northern analysis of root AtAMT1 transcript abundance following resupply of N revealed a rapid decline during the first 6 h of down-regulation (Fig. 3a,b). High-affinity 13NH4+ influx, measured at 100 μm NH4NO3, also declined rapidly during this time (Fig. 3b). Thus, by 12 h, 13NH4+ influx had decreased by 93% from 4.82 to 0.34 μmol g–1 FW h–1. Expressing AtAMT1 transcript abundance and NH4+ influx as percentages of their maximum values (Fig. 3b) reveals that NH4+ influx declined more rapidly than the decline of root AtAMT1 transcript abundance. This pattern was consistently observed during down-regulation.

Figure 3.

Time-dependent down-regulation of AtAMT1 mRNA levels and 13NH4+ influx in roots of A. thaliana, following re-supply of NH4NO3.

Plants were supplied with 1 mm NH4NO3 for 19 days, then deprived of N for 2 days, prior to transfer to 5 mm NH4NO3.

(a) Northern blot analysis of root RNA, probed with the AtAMT1 cDNA. Internal standard was as in Fig. 1 

(b) AtAMT1 mRNA levels (○) and 13NH4+ influx (●) during down-regulation, expressed as percentages of their maximum values.

Changes of cellular N pools during down-regulation of AtAMT1 transcript levels and NH4+ influx

In order to identify the N pools (NH4+ or assimilation products) responsible for the observed changes of NH4+ influx and AtAMT1 expression during down-regulation, we measured root and shoot concentrations of NH4+ and individual amino acids, patterns of 13NH4+ influx, and AtAMT1 expression levels, in the presence and absence of 1 mm methionine sulfoximine (MSX) following re-supply of 5 mm NH4NO3. The inhibitor MSX blocks the action of the enzyme glutamine synthetase (GS) resulting in elevated [NH4+] and reduced [Gln] (King et al. 1993; Lee et al. 1992). As shown in Fig. 3(a,b), the AtAMT1 transcript level was reduced to about 20% of its initial (time 0) level, 9 h after the transfer to 5 mm NH4NO3. Therefore, to minimize the duration of exposures to MSX in subsequent experiments, 9 h treatments were employed as the standard for experiments using MSX.

The most prominent change in root N pools during down-regulation was that of Gln (Fig. 4), which increased ninefold, from 2.58 μmol g–1 FW at time 0 to 23.2 μmol g–1 FW at 24 h. No other amino acid showed such a substantial change in concentration, while root [NH4+] increased 3.8-fold. Table 2 provides r2 values for the regressions of AtAMT1 transcript levels (expressing transcript level at time 0 as 100%), against root or shoot amino acid concentrations during the 24 h following resupply of NH4NO3. Only Gln, aspartate and the sum of all amino acids gave statistically significant r2 values (0.92, 0.85 and 0.86, respectively). Figure 5 reveals the strong negative correlations between AtAMT1 levels and root [Gln], and 13NH4+ influx and root [Gln], throughout the 24 h of exposure to 5 mm NH4NO3. In leaf tissue, like roots, the largest absolute and relative change was in Gln concentrations.

Figure 4.

Root NH4+, glutamine and total amino acid concentrations of A. thaliana plants (μmol g–1 FW), at 0 and 9 h after resupply of 5 mm NH4NO3 to N-limited plants, in the presence and absence of 1 mm MSX.

Table 2.  Coefficients of determination (r2) for the relationships between AtAMT1 transcript levels and tissue amino acid concentrations ([A.A.]), and changes in tissue [A.A.] (μmol g–1 FW) 24 h after supplying 5 mm NH4NO3
r2Changes in [A.A.]
Amino acidRootShootRootShoot
  • In parentheses: [A.A.]t = 24 h/[A.A.]t = 0 h ratio.

  • Regressions were based upon measurements at six time intervals (four replicates each) within a 24 h period.

  • Changes in [A.A.] are shown only for the longest interval (0–24 h).

  • *

    P < 0.05,

  • **

    P < 0.01, nc: no change.

Ammonium0.500.460.98 (3.8)0.80 (3.6)
Alanine0.030.72*0.63 (1.4)0.82 (2.7)
Arginine0.230.490.15 (1.9)0.97 (5.7)
Aspartate0.84**0.540.57 (6.5)0.47 (3.2)
Asparagine0.600.390.54 (2.1)1.83 (2.9)
Gaba0.010.52nc0.63 (1.9)
Glutamate0.040.28nc0.54 (3.0)
Glutamine0.92**0.79*20.6 (9.0)14.3 (6.5)
Glycine0.580.220.18 (1.7)nc
Histidine0.240.53nc0.05 (4.6)
Isoleucine0.360.350.05 (0.5)0.02 (1.9)
Leucine0.230.35nc0.07 (4.4)
Lysine0.180.54nc0.09 (3.8)
Phenylalanine0.010.37nc0.03 (2.0)
Proline0.400.8*–0.98 (0.6)1.00 (1.9)
Serine0.130.52nc0.55 (1.8)
Threonine0.080.78*nc0.28 (3.1)
Valine0.070.35nc0.08 (2.2)
Total A.A.0.86**0.74*21.8 (2.8)21.7 (3.8)
Total A.A. – Gln0.240.59nc7.41 (2.4)
Figure 5.

AtAMT1 expression levels (main figure) expressed as a percentage of maximum expression levels at time 0, and 13NH4+ influx in μmol g–1 FW h–1 (inset) during a 24 h period of down-regulation (data from Fig. 3) as functions of root glutamine concentrations in μmol g–1 FW (from Fig. 4).

Regression lines for AtAMT1 expression levels and 13NH4+ influx are given by a linear regression (r2 = 0.92), and a polynomial regression Y = 5.55–0.46 X + 0.01 X2 (r2 = 0.89), respectively.

When MSX was provided together with 5 mm NH4NO3, the AtAMT1 transcript level at 9 h remained at 88% of its original value (Fig. 6a,b). During this time, root [NH4+] increased 27-fold from 0.35 μmol g–1 FW at time 0 to 9.7 μmol g–1 FW at 9 h (Fig. 4), while root Gln concentration remained relatively unchanged (declining from 2.58 to 2.41 μmol g–1 FW). When plants were treated for 9 h with 1 mm MSX, together with 5 mm NH4NO3 and 5 mm Gln, AtAMT1 transcript expression was reduced by 76% (Fig. 6a), confirming the capacity of Gln to circumvent the MSX blockage of NH4NO3 assimilation. Glutamine, either alone or together with 5 mm NH4NO3, also strongly reduced AtAMT1 mRNA expression levels (Fig. 6a). Corresponding measurements of NH4+ influx (Fig. 6b) revealed that the presence of MSX, together with NH4NO3, had reduced influx by <10% by 3 h, compared to 90% in the absence of MSX. By 9 h, however, influx was reduced by approximately 35%. When N-deprived plants (control treatment) were supplied with 5 mm Gln in place of 5 mm NH4NO3, 13NH4+ influx was reduced from 8.67 ± 1.42 μmol g–1 FW h–1 (control) to 1.33 ± 0.23 μmol g–1 FW h–1 by 9 h.

Figure 6.

Effects of NH4NO3, MSX and Gln on AtAMT1 mRNA levels and 13NH4+ influx in roots of A. thaliana.

(a) Down-regulation of AtAMT1 mRNA levels by 5 mm NH4NO3 (left panel) and effects of MSX and Gln (right panel). Internal standard as shown in Fig. 1 

(b) AtAMT1 mRNA levels (○) and 13NH4+ influx (●) in the presence and absence of MSX, expressed as percentages of their maximum values.

The low-affinity NH4+ transport system (LATS): NH4+ influx is independent of N supply

As shown in Fig. 7, at NH4+ concentrations from 1 to 10 mm, a low-affinity (LATS) transport system was evident. In plants grown in 100 μm NH4NO3, this transport system showed no indication of saturation and, in contrast to the down-regulation of high-affinity NH4+ influx, low-affinity influx was not reduced by growth at 1 mm or 10 mm NH4NO3. Since measured 13NH4+ fluxes at high external NH4+ concentration are the resultant of the two transport systems (HATS plus LATS), the Vmax values for 13NH4+ were subtracted from measured 13NH4+ influxes to generate the LATS contribution to measured 13NH4+ fluxes. Figure 7 shows the subtracted values for LATS fluxes in plants grown at 100 μm, 1 mm and 10 mm when 13NH4+ fluxes were measured at 1–10 mm external [NH4+]. The regression lines for 13NH4+ influx against external [NH4+] were so similar that a single regression line provided a good fit of the data (r2 = 0.94) at all values except perhaps at 10 mm external [NH4+].

Figure 7.

13NH4+ influx (μmol g–1 FW h–1) due to the LATS activity (calculated by subtracting the HATS fluxes from measured fluxes) at 1–10 mm external [NH4+].

Plants were grown for 2 weeks in 1 mm NH4NO3 and then transferred for 1 week to 0.1, 1.0 or 10 mm NH4NO3 prior to influx measurements. A single regression line (r2 = 0.98) summarizes the three data sets.


Regulation of NH4+ influx in response to N supply

The results of the present compartmental analyses confirm that roots of Arabidopsis plants behave essentially like those of rice and white spruce, with respect to half-lives of 13NH4+ exchange and accumulation of NH4+ in cytoplasmic and cell wall compartments (Kronzucker et al. 1995; Wang et al. 1993a). We therefore settled on influx and desorption times of 10 and 3 min, respectively, to estimate plasma membrane NH4+ influx. Varying the ambient N supply to A. thaliana plants prior to influx determinations resulted in a strong down-regulation of the high-affinity NH4+ transport system (HATS). This is consistent with several earlier studies using several crop species (Becking 1956; Causin & Barneix 1993; Morgan & Jackson 1988; Wang et al. 1993a; Wang et al. 1993b). The results of these studies established that NH4+ uptake was negatively correlated with prior N supply but failed to provide consensus regarding the underlying mechanism(s) of this feedback, particularly which N pool, NH4+ or products of its assimilation were responsible for the feedback.

High Vmax values and a saturable concentration response for 13NH4+ influx in roots of A thaliana validate the designation of this transport system as a high-affinity (HATS) NH4+ transport system. Furthermore, while Vmax values responded to the NH4NO3 supply under steady-state conditions, as shown in previous reports (Kronzucker et al. 1995; Wang et al. 1993b), Km values were not correlated with N supply. Absolute values for Km varied from 85 to 168 μm, but failed to correlate with N supply as was reported by Wang et al. (1993a) for rice. Literature reports for Km values vary considerably (see Glass & Siddiqi 1995), ranging from approximately 20–167 μm. Lycklama (1963) reported values from 40 to 200 μm according to plant age, using 14NH4+ depletion to determine Km values for net NH4+ uptake by roots of ryegrass.

AtAMT1 gene expression and N supply

The high levels of AtAMT1 gene expression in roots of low-N plants, which showed high HATS activity, provide correlative evidence that this gene encodes a HATS for NH4+ influx in A. thaliana. Likewise, the demonstrated changes of Vmax for 13NH4+ influx, and the short-term changes of influx resulting from manipulating the N supply, were correlated with changes of AtAMT1 gene expression. These results provide evidence that this feedback loop operates, at a minimum, through effects at the level of AtAMT1 mRNA. However, while changes of AtAMT1 expression consistently preceded changes of 13NH4+ influx during up-regulation (Fig. 2), 13NH4+ influx was reduced more rapidly than AtAMT1 transcript levels during down-regulation (Figs 3 and 6). We interpret this anomaly as the result of direct inhibitory effects of cytosolic NH4+ on membrane transport activity (as discussed below).

Regulation of AtAMT1 gene expression depends upon tissue Gln concentrations

The results of the short-term experiments in which N supply was withheld or resupplied provide the context in which to identify which cellular N pool(s) are responsible for regulating AtAMT1 gene expression. The outcome of the experiments using MSX, with or without added Gln, generate a convincing argument that tissue NH4+ is not responsible for the observed effects on AtAMT1 gene expression. Rather, the correlations between AtAMT1 gene expression and root [Gln], and between 13NH4+ influx and root [Gln] (Fig. 5 and Table 2), provide strong evidence that Gln is the controlling agent. Although aspartate and total amino acid concentrations were significantly correlated with AtAMT1 gene expression (Table 2), root aspartate concentrations were typically <4% of Gln concentrations at all time points and, therefore, represented only a minor component of the increased tissue N associated with N resupply. The data from Fig. 4 and Table 2 establish that the major part of absorbed NH4+ was accounted for by conversion to Gln within the roots. Moreover, the high r2 for the sum of all amino acids (Table 2) was reduced from 0.86 to 0.24 when Gln was removed from the regression. Thus the evidence points to root Gln concentrations as responsible for the down-regulation of AtAMT1 transcript levels. These conclusions are consistent with earlier physiological studies using wheat, barley and squash which demonstrated that exogenous application of Gln reduced NH4+ uptake (Causin & Barneix 1993; Lee et al. 1992; Wieneke & Roeb 1998).

Although our analyses of root and shoot Gln concentrations were average tissue concentrations, it has been demonstrated in earlier studies (Winter et al. 1992) that cellular amino acids are mainly localized within the cytosol. The rapidity of the observed changes of 13NH+ influx and AtAMT1 gene expression, resulting from perturbation of the N supply, argue for control via cytoplasmic N pools. While we have no data on rates of amino acid turnover in the cytoplasm, the present compartmental analyses indicate that the t0.5 for cytoplasmic NH4+ exchange is about 12 min. Thus, changes of gene expression and NH4+ transport might be evident within hours of perturbing the N supply.

Effects of cellular NH4+ on the NH4+ transporter

The application of MSX plus 5 mm NH4NO3 caused virtually no down-regulation of AtAMT1 gene expression, which remained at 88% of the control (0 time) value after 9 h (Fig. 6). By contrast, 13NH4+ influx was consistently reduced by 30–40% under the same conditions (Fig. 6). Given that the [NH4+] of MSX-treated roots had increased 27-fold by 9 h (Fig. 4), and that Gln levels were essentially unchanged, increased efflux of NH4+ during influx measurement might have reduced the specific activities of 13NH4+ in the influx media, resulting in an artificial lowering of the calculated influx values. By measuring the [NH4+] of influx media after the 10 min influx period, we demonstrated that there were no changes to the 13NH4+ specific activity. Thus, the reduced values for 13NH4+ influx associated with high cytosolic [NH4+] in the presence of MSX are real and may result from inhibitory effects of cytosolic NH4+ on the activity of the HATS protein. Previous experiments have demonstrated that NH4+ is able to inhibit plasma membrane influx and efflux of NO3 within minutes of its application (Aslam et al. 1996; Glass & Siddiqi 1995; Lee & Drew 1989).

Differential response of the HATS and LATS to N status

Ammonium transport in A. thaliana showed biphasic uptake kinetics as previously reported for roots of Lemna, rice and white spruce (Kronzucker et al. 1996; Ullrich et al. 1984; Wang et al. 1993b). At low external [NH4+], the saturable high-affinity transporter is capable of active NH4+ absorption, while at higher concentrations (> 1 mm NH4+) transport is passive (Ullrich et al. 1984; Wang et al. 1993b). In contrast to the observed down-regulation of HATS activity and AtAMT1 transcript levels, expression of the low-affinity transport system in roots of A. thaliana failed to demonstrate down-regulation when plants were grown at 1 mm and 10 mm external [NH4+] (Fig. 7). This phenomenon whereby internal ion concentrations regulate the activity of high-affinity transporters but not the corresponding low-affinity systems, was first observed for K+(86Rb+) influx in barley and ryegrass (Glass & Dunlop 1978) and subsequently in corn (Kochian & Lucas 1982). It was also evident in studies of NH4+ influx in rice (Wang et al. 1993b). It is perplexing that only the high-affinity systems should be regulated in this way. However, the apparent absence of LATS regulation may account for the toxic effects associated with elevated NH4+ provision in many species (Bloom 1988; Gill & Reisenauer 1993; Magalhaes & Wilcox 1991). It might be argued that the elevated concentrations employed to measure low-affinity transport systems are representative only of agricultural soils where high levels of fertilizer are applied. Hence, the conditions required to measure the operation of the low-affinity transporters could be considered somewhat unnatural. Nevertheless, the existence of biphasic (high- and low-affinity) transporters for a wide range of inorganic and organic nutrients appears to be universal, suggesting functional operation outside the agricultural context.

In summary, both high- and low-affinity NH4+ transporters are expressed in roots of A. thaliana. Expression levels of the AtAMT1 gene were strongly correlated with high-affinity NH4+ influx, providing further evidence for the physiological function of this gene in high-affinity NH4+ transport. Transcript levels of the AtAMT1 gene appear to be regulated by root Gln concentrations, but not by NH4+. Since Gln is the first product of NH4+ assimilation in roots (and in shoots), this compound probably serves as the primary signal of cellular N-status. Nevertheless, high levels of cytoplasmic NH4+ act upon the transport protein, either through direct (e.g. allosteric effects) or via post-translational events. A model summarizing the findings of the present study is shown in Fig. 8.

Figure 8.

A model describing the regulation of high-affinity NH4+ transport, at the level of AtAMT1 transcription and at the level of the NH4+ transporter.

AtAMT1 gene transcript level is controlled through effects of cellular Gln (solid lines), while NH4+ may act upon the pAtAMT1 transport protein (dotted lines) through direct (e.g. allosteric or post-translational) effects.

Experimental procedures

Growth of plants

A. thaliana (L.) plants (ecotype Columbia) were grown in a controlled environment chamber at 20 ± 2°C, in 16 h light/8 h dark cycles, 70% RH and 250 μmol photons m–2 s–1 at plant level (Vitalite fluorescent tubes, Durotest, North Bergen, NJ, USA). Twenty seeds were germinated in 5 ml of sterile culture solution: 29 mm sucrose, 2.6 mm MES, 2 mm KH2PO4, 1 mm MgSO4, 1 mm NH4NO3, 1 mm CaCl2, 20 μm NaFeEDTA, 25 μm H3BO3, 12 μm MnSO4, 1 μm ZnCl2, 1 μm CuCl2, and 0.2 μm Na2MoO4 at pH 5.7. The choice of NH4NO3 as the N source for plant growth was based upon the relatively poor growth of A. thaliana on NH4Cl or (NH4)2SO4 (S,R. Rawat et al. unpublished observations).

After 3 days, seedlings were transferred to sterile culture vessels (Magenta Corporation, Chicago, IL, USA), containing 60 ml of fresh culture solution and were gently agitated on a gyrotory shaker. Nutrient solutions were replaced frequently to maintain steady state with respect to plant N status. In order to generate low-, intermediate-and high-N plants, respectively, 2-week-old plants were transferred from media containing 1 mm NH4NO3 to fresh media containing 100 μm NH4NO3, 1 mm NH4NO3, or 10 mm NH4NO3 which were replaced regularly. In short-term perturbation experiments, low-N plants were transferred to solutions containing 5 mm NH4NO3, and high-N plants were transferred to solutions containing 100 μm NH4NO3. During experiments lasting up to 24 h, NH4+ influx values, tissue N levels and AtAMT1 expression levels were determined.

Efflux analysis

In order to develop an appropriate protocol for influx measurements, efflux analyses were undertaken using 21-day-old plants maintained at 100 μm or 1 mm NH4NO3. Efflux analysis was undertaken according to previously published methods (Kronzucker et al. 1995). Roots were equilibrated for 5 min in fresh media and then exposed to 13NH4+-labeled culture medium containing 100 μm or 1 mm NH4NO3 for 45 min 13N eluting from the superficial solution, the cell wall and the cytoplasmic compartments was counted in a Packard gamma-counter (Minaxi γ, Auto-gamma 5000 series, Downer’s Grove, IL, USA), and the half-lives for 13NH4+ exchange of the above compartments, NH4+ flux values between compartments, and cytoplasmic [NH4+] were estimated (see Kronzucker et al. 1995; Lee & Clarkson 1986). To ensure that the label effluxing from roots consisted of 13NH4+ rather than products of 13NH4+ assimilation, the difference between counts before and after alkalization and boiling was determined. Typically, about 98% of 13N contained in the original eluates were volatilized by 6 min of boiling.

13NH4+ influx determinations

13NH4+ influx was determined from efflux analysis (see above) and by direct measurement of the accumulation of 13N from nutrient solutions during a 10 min influx period followed by 3 min desorption in identical non-labeled solution to remove tracer from the cell wall. These times were based upon the calculated half-lives for tracer exchange of the cell wall (0.56–0.66 min) and cytoplasmic compartments (10.7–12.6 min) as a compromise which optimized removal of tracer from the cell wall while minimizing loss of tracer from the cytoplasm. Total 13N accumulated in roots and shoots was determined by gamma-counting.

Time-dependent analysis of 13NH4+ influx, AtAMT1 mRNA levels and root [NH4+] and [Gln]

For up-regulation, plants previously acclimated in 1 mm NH4NO3 for 21 days were transferred to 100 μm NH4NO3. AtAMT1 mRNA levels and 13NH4+ influx (from 100 μm NH4NO3) were subsequently measured at intervals up to 24 h. To investigate the pattern of down-regulation, plants were acclimated in 1 mm NH4NO3 for 19 days and then transferred to solutions without N for 2 days to reduce tissue N and generate plants with high initial rates of NH4+ influx. These were then transferred to 5 mm NH4NO3, and AtAMT1 mRNA levels and 13NH4+influx (from 100 μm NH4NO3) were measured for up to 24 h. All transfers were staggered so that AtAMT1 mRNA levels and 13NH4+ fluxes were determined at the same time so as to eliminate diurnal effects. This was also essential because of the short half-life (9.98 min) of the 13N tracer. To investigate the down-regulation of NH4+ influx, 1 mm methionine sulfoximine (MSX), an inhibitor of the enzyme glutamine synthetase, was used to block the conversion of NH4+ to Gln. This treatment increased tissue [NH4+] and reduced Gln and amino acid levels (King et al. 1993; Lee et al. 1992). In separate experiments, MSX with or without 5 mm Gln, was added to the pre-treatment solutions containing 5 mm NH4NO3. AtAMT1 expression and 13NH4+ influx were then measured at intervals up to 9 h. All flux data presented are the means of at least three, typically four replicates. All experiments were repeated at least three times. Analysis of root [NH4+] and individual [amino acid] were analyzed by HPLC. Plant tissues from 30 plants were ground to powder in liquid nitrogen and extracted in cold 10 mm sodium acetate buffer (pH 6.5). The extracts were centrifuged at 16 000 g, filtered through a 0.45 μm, filter and then derivatized using the AccQ.Fluor reagent (Waters Chromatography, Milford, MA, USA). Separations were carried out using a 3.9 × 150 mm AccQ.Tag Column (Waters Chromatography, Milford, MA, USA), according to published methods (Van Wandelen & Cohen 1997) on a Waters 600 LC system and detected with a Waters 474 scanning fluorescence detector (Waters Chromatography, Milford, MA, USA).

RNA extraction and Northern blot analysis

For expression studies, each treatment unit consisted of a single culture vessel containing 30 Arabidopsis seedlings. Each treatment was replicated and all experiments were repeated. Total RNAs were isolated from root and shoot tissues using TRIzol Reagent (GIBCO BRL) according to the manufacturer’s instructions. Total RNA (15 μg) were fractionated on a 1.2% MOPS-formaldehyde agarose gel and transferred to nylon membrane (Amersham, Oakville, Canada). Northern hybridizations were carried out according to standard protocols (Sambrook et al. 1989). A 1.75 kb Not1 fragment of pAtAMT1 representing a full-length cDNA was purified, 32P-labeled using random priming, and used as a probe. Final washes of Northern blots were done once at 65°C in 0.1× SSPE and 0.1% SDS and subjected to autoradiography using X-OMAT film at –70°C overnight. All EtBr stained gels (for loading corrections) and autoradiographs were normalized by reprobing the same Northern blots with an At25S rDNA probe, and analyzed by densitometric scanning using the Molecular Analyst computer software program (Bio-Rad, Hercules, CA, USA).


We thank Wolf. B. Frommer (Institut für Genbiologische Forchung, Germany) for providing the AtAMT1 cDNA and the ABRC DNA stock centre (The Ohio State University, OH, USA) for AtrDNA clone. This work was supported by the Natural Sciences and Engineering Research Council of Canada (Grant STR0167384) to A.D.M.G.