INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

In many natural ecosystems and in high-input monoculture cropping systems, nitrogen availability represents a limiting factor, necessitating the use of large quantities of nitrogen fertilizers (presently approximately 10¹¹ kg per annum globally; http://www.fao.org/) to achieve current high yields of quality crops (Marschner, 1995). However, this high level of nitrogen fertilizer use may generate environmental problems, such as ground water contamination by nitrate, anoxia of rivers and oceanic coastal waters, or even fish kills attributable to leaching of N from agricultural soils (Cohen et al., 1996). In cereal crops, which receive roughly 60% of all N fertilizers, recovery of applied N in grain is typically only 33% of fertilizer applied to soils (Raun and Johnson, 1999). This low N use efficiency is caused by several factors, including volatilization of NH₃, denitrification, and/or leaching of soil nitrate (NO₃). However, it is also the result of inefficient N absorption (K_M values for NO₃ and NH₄⁺ absorption by higher plants may be orders of magnitude higher than those of algae and fungi (Galvan et al., 1996; Unkles et al., 2001). Furthermore, as plants accumulate N, there is a rapid down-regulation of genes encoding high-affinity NO₃ and NH₄⁺ transporters (Rawat et al., 1999; Forde, 2000; Howitt and Udvardi, 2000; Glass et al., 2001), and there are corresponding reductions of high-affinity influx (Glass and Siddiqi, 1995; Gazzarrini et al., 1999; Rawat et al., 1999; Zhuo et al., 1999; von Wiren et al., 2000a). In addition, as external N levels increase, N efflux from roots becomes a significant proportion of net influx, especially in the case of NH₄⁺, where futile NH₄⁺ cycling (leak and pump across the plasma membrane) has been demonstrated to increase respiration rates by 40% in barley (Hordeum vulgare) roots (Britto et al., 2001). It clearly will be essential to better define the mechanisms involved in regulating N acquisition, assimilation, and redistribution in the plant to enable agronomists to improve the management of plant N nutrition and to achieve genetic modifications that might sustain present high rates of crop productivity with reduced N inputs.

A first step in improving plant N use efficiency would be to address the primary acquisition process occurring at the root/soil interface. Genes encoding NO₃ transporters, NRT1 and NRT2 (Tsay et al., 1993; Forde, 2000) and NH₄⁺ transporters, AMT1 and AMT2 (Ninnemann et al., 1994; Lauter et al., 1996; Gazzarrini et al., 1999; Sohlenkamp et al., 2000; von Wiren et al., 2000a) have been identified in plants. These transporters are presumed to be located mainly at the root plasma membrane and, therefore, positioned to take N up from the soil. Nevertheless, in the cases of both NO₃ and NH₄⁺ uptake, many more genes have been identified in Arabidopsis and other organisms than would have been anticipated from physiological studies (Glass and Siddiqi, 1995; Crawford and Glass, 1998; Forde and Clarkson, 1999). An important challenge is to ascribe specific functions to each of these genes.

Although NO₃ is the predominant form of N in well-aerated and pH-balanced soils, NH₄⁺ is an important and commonly underestimated N source for many plant species, in part because in mixtures of NO₃ and NH₄⁺, NH₄⁺ inhibits NO₃ uptake. In Lemna sp., in rice (Oryza sativa), and in Arabidopsis, NH₄⁺ influx consists of both saturable high-affinity influx (HATS) and non-saturable low-affinity influx (LATS) that operate at low and high NH₄⁺ concentrations, respectively (Ullrich et al., 1984; Wang. et al., 1993; Rawat et al., 1999). It is still unknown what molecular mechanisms are involved in either the high- or low-affinity NH₄⁺ transport phenomenon in plant roots.

Some members of the AMT1 family are suggested to be putative high-affinity NH₄⁺ transport proteins in planta based on their functional analysis when expressed in yeast cells (see below). Included in this group of genes are AtAMT1;1, AtAMT1;2, and AtAMT1;3 (Gazzarrini et al., 1999; Rawat et al., 1999) and three AMT1 homologs in tomato (Lycopersicon esculentum; Lauter et al., 1996; von Wiren et al., 2000a). Two other AMT1 homologs, AtAMT1;4 and AtAMT1;5, have been identified with the complete sequencing of the Arabidopsis genome; whether they are involved in NH₄⁺ transport remains to be determined (von Wiren et al., 2000b). A tomato homolog, LeAMT1;1, found in root hairs has recently been expressed in Xenopus sp. oocytes and shown to function as an NH₄⁺ uniporter dependent on membrane potential and NH₄⁺ concentration gradients (Ludewig et al., 2002).

Sequencing of the Arabidopsis genome has also resulted in the identification of a second class of AMT genes designated AMT2 (Sohlenkamp et al., 2000). Functional expression of AtAMT2 in yeast cells revealed a strikingly different phenotype to that of AMT1, where the uptake of NH₄⁺ resembled a low-capacity transport system (Sohlenkamp et al., 2000). Members of the AMT gene families show distinctive patterns of RNA expression in roots and shoots and during the diurnal cycle (Gazzarrini et al., 1999; Rawat et al., 1999; von Wiren et al., 2000a), which suggests that this gene family may serve a number of different functions associated with NH₄⁺ transport across the plasma membrane and within the plant.

AtAMT1;1 has traditionally been considered of prime importance in Arabidopsis NH₄⁺ transport, particularly in the roots where its mRNA is highly abundant. High-affinity NH₄⁺ influx is strongly correlated with transcript abundance, which appears to be regulated by root Gln (but not NH₄⁺) concentration (Gazzarrini et al., 1999; Rawat et al., 1999; Gansel et al., 2001). Furthermore, functional analysis in yeast cells demonstrated that AtAMT1;1 is capable of mediating high-affinity NH₄⁺ transport with a significantly higher affinity for NH₄⁺ (K_M approximately 500 nM) than its counterparts AtAMT1;2 and AtAMT1;3 (K_M approximately 40 µM; Gazzarrini et al., 1999; Shelden et al., 2001). To decipher the role of AtAMT1;1 in Arabidopsis, we have used a reverse genetic approach to identify an amt1;1 transfer-DNA (amt1;1:T-DNA) tagged line. Disruption of AtAMT1;1 activity reduced high-affinity NH₄⁺ influx by only 30%, whereas low-affinity transport was increased. Additional phenotypic effects of this disruption included an altered leaf morphology and a lethal condition in plants grown under sterile conditions in the presence of NH₄⁺ and Suc.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

amt1;1 T-DNA Isolation and Characterization of Inserted Loci

The amt1;1:T-DNA line was identified from the subpool, CD6-4A, of the "Jack enhancer-trap" lines. DNA isolated from individual lines was screened using the PCR with the T-DNA right border primer (P3) and the AtAMT1;1 gene specific primer (P2; Table I). A PCR fragment of the expected size was identified in which the right border of the T-DNA was inserted 6 bp upstream of the first putative Met in the open reading frame of AtAMT1;1 (Fig. 1A). Inverse PCR revealed that the left border was inserted 79 bp upstream of the right border insertion, deleting a 72 bp fragment of the genome (Fig. 1A). A 3-bp insertion of unknown DNA was also inserted in this junction (Fig. 1A). The amt1;1:T-DNA line was selfed and then backcrossed to Col3 gl1. Because no visible phenotype was apparent in the mutant (see below) homozygous lines were selected using both a PCR-based screen and Southern-blot analysis.

View larger version (40K): [in this window] [in a new window]   Figure 1.   Characterization of the plant line with a T-DNA insertion in AtAMT1;1. A, Localization of the T-DNA insert. The diagram illustrates the insertion of the T-DNA 6 bp upstream of the first putative Met in the open reading frame of AtAMT1;1. The bordering nucleotides are presented adjacent to the insertion sites of both the left (LB) and right (RB) borders of the T-DNA insert. In italics, the insertion of three nucleotides of unknown origin is shown at the T-DNA left-border junction as well as the location of primer (P3 P7 P8) binding sites used for selecting tagged and non-tagged AtAMT1;1 alleles. B, Cartoon illustrating the T-DNA insertion into chromosome 4 and the predicted sized fragments after digestion with EcoRI (E) and HindIII (H) not drawn to scale. C, Southern-blot analysis of EcoRI (lane 1) and HindIII (lane 2) digested amt1;1:T-DNA genomic DNA. The blot was probed with a 1.2-kb DIG-labeled -glucuronidase (GUS) gene and demonstrates a single T-DNA insertion in the amt1;1:T-DNA genome. D, Northern-blot analysis of AtAMT1;1 expression in the mutant and wild-type roots and shoots. Plants were grown vegetatively in liquid culture for 5 weeks in 1 mM NH4NO3 and then transferred to nutrient solution without N for 4 d. The blot was probed with DIG-labeled AMT1;1 antisense RNA.

DNA isolated from the backcrossed lines were subjected to two rounds of the PCR. The first round was designed to identify the T-DNA-tagged amt1;1 allele using primers to amplify the 1.7-kb region spanning the right border (P3) of the T-DNA insert and the 3' end of AtAMT1;1 (P7; Fig. 1A). The second PCR used primer pairs (P8 and P7) to identify the wild-type allele, amplifying a 2.3-kb region of genomic DNA, which included the upstream promoter region and the 3' end of AtAMT1;1 (Fig. 1A). A homozygous line that failed to amplify the wild-type allele but did amplify the T-DNA-tagged amt1;1 was selected and selfed (data not shown). To identify the number of T-DNA insertion events, Southern-blot analysis was performed on EcoRI or HindIII digested genomic DNA. Using a 1.2-kb cDNA probe specific to the uidA (GUS) open reading frame, single DNA fragments of 15 and 5.4 kb were identified in the EcoRI- and HindIII-digested DNA, respectively (Fig. 1C). HindIII restriction sites are present close to the right border of the T-DNA and approximately 400 bp upstream from the left border T-DNA insertion site in chromosome 4 (Fig. 1B). As expected, digestion with HindIII liberated a 5.4-kb DNA fragment from the AMT1;1 loci, which contained the majority of the T-DNA insert. Digestion with EcoRI also confirmed the presence of a single-insertion event in amt1;1:T-DNA (Fig. 1C), however, the identified 15-kb fragment was smaller than the predicted fragment size of 20 kb based on a restriction enzyme digest profile of the bacterial artificial chromosome sequence (accession no. AL049656).

Disruption of AtAMT1;1 mRNA Expression

The expression of AtAMT1;1 was examined using northern-blot analysis in both amt1;1:T-DNA and wild-type plants maintained on 1 mM NH₄NO₃ for 5 weeks and then deprived of N for 4 d. This treatment up-regulates AtAMT1;1 mRNA expression in roots (Gazzarrini et al., 1999; Rawat et al., 1999; Gansel et al., 2001). Northern blots containing total RNA from both the T-DNA mutant and wild-type roots and shoots were incubated with a 1.7-kb DIG-labeled AtAMT1;1 anti-sense RNA probe. A 1.7-kb hybridization product was identified in the wild-type root and shoot total RNA extracts but not in the T-DNA mutant (Fig. 1D).

Growth Analysis

No discernible phenotype was evident when T-DNA mutants were grown on a peat-based soil or in an open hydroponic system in nutrient solution with no added C. Plants germinated, bolted, and set flower as did wild-type plants. When grown in open hydroponic systems for 6 weeks on high-N media containing 1 mM NH₄NO₃ or 2 mM KNO₃, or low-N media containing 50 µM (NH₄)₂SO₄ or 100 µM KNO₃, both without added C, no significant differences in fresh weights were detected (Table II). By stark contrast, mutant plants grown in sterile hydroponic conditions with 1% (w/v) Suc and 0.5 mM (NH₄)₂SO₄ grew poorly relative to the wild type and died after 2 to 3 weeks (Fig. 2, A and B). This lethality was not expressed when 0.5 mM (NH₄)₂SO₄ was replaced by 1 mM KNO₃ under otherwise identical conditions and equivalent N levels (Fig. 2C).

View larger version (90K): [in this window] [in a new window]   Figure 2.   Plant growth in magenta boxes containing sterile nutrient solution including Suc. Stratified seeds were sown directly onto nylon mesh sitting on the floating rafts in nutrient solution containing 1% (w/v) Suc and either 0.5 mM (NH4)2SO4 (A and B) or 1 mM KNO3 (C). Closed magenta boxes were placed in growth rooms on orbital shakers and plants grown for 21 (A) and 10 (B and C) d.

Leaves of the amt1;1:T-DNA were found to be distinctly more succulent than those of the wild type. They were characterized by a thickening of the mesophyll tissue along the major vein of the leaf, and a significant decrease in the extent of intercellular air spaces between blade mesophyll cells (Fig. 3).

View larger version (44K): [in this window] [in a new window]   Figure 3.   Cross sections of wild-type and amt1;1:T-DNA leaves. Sections are from leaves of plants grown in a controlled temperature growth chamber in hydroponic tanks (A) or in the greenhouse (B). Plants grown in the greenhouse were on a peat-based potting mix containing a slow-release fertilizer (Osmocote), whereas plants grown in the chamber were identical to those grown on adequate N (1 mM NH4NO3) as previously described for the 13NH4+ uptake experiments. Cross sections from fully expanded leaves of equal age between lines were taken midway along the blade. Arrows indicate air spaces mainly absent in the amt1;1:T-DNA line. Leaf lengths ranged from 5 to 6 cm.

The Root High-Affinity NH₄⁺ Transport System

Rates of high-affinity ¹³NH₄⁺ influx into intact roots of N-replete wild-type and amt1;1:T-DNA plants were similar (Fig. 4A). However, after 4 d of N deprivation, significant differences in ¹³NH₄⁺ influx and V_max values for influx were apparent whereas trends toward higher K_M values were evident in the amt1;1:T-DNA over those of the wild type. K_M and V_max values were 17.2 ± 4.4 µM and 8.4 ± 0.5 µmol ¹³NH₄⁺ g¹ fresh weight h¹ for amt1;1:T-DNA and 11.3 ± 3.2 µM and 11.4 ± 0.6 µmol ¹³NH₄⁺ g¹ fresh weight h¹ for the wild type (Fig. 4B).

View larger version (17K): [in this window] [in a new window]   Figure 4.   13NH4+ influx by the HATS for both amt1:1:T-DNA and wild-type plants. A, Plants were grown in the presence of 1 mM NH4NO3 for 5 weeks and then transferred to nutrient solution with or without 1 mM NH4NO3 for 4 d. 13NH4+ influx was measured over a 10-min period in plants precultured in 12.5 (25 µM NH4+) and 50 µM (100 µM NH4+; NH4)2SO4. Values are means ± SE (n = 5). B, Concentration dependence of 13NH4+ influx into plant roots by the HATS. Plants grown as above were starved of N for 4 d. Data represents the averaged values ± SE of three independent experiments (n = 14-15 plants). The fitted curve was obtained by direct fit to the Michaelis-Menten equation. Estimated KM and Vmax ± SE were 11.3 ± 3.2 and 17.2 ± 4.4 µM (P = 0.2777) and 11.4 ± 0.6 and 8.3 ± 0.5 µmol NH4+ g1 fresh weight h1 (P < 0.05) for the wild-type and amt1;1:T-DNA line, respectively.

The Root Low-Affinity NH₄⁺ Transport System

At higher NH₄⁺ concentrations (up to 10 mM), ¹³NH₄⁺ influx (the sum of LATS and HATS), was higher in the mutant than in the wild type (Fig. 5A). This was even more apparent when the calculated HATS V_max values were subtracted from the combined LATS and HATS activities to give the corrected LATS flux (Fig. 5B).

View larger version (14K): [in this window] [in a new window]   Figure 5.   NH4+ influx by the LATS for both amt1;1:T-DNA and wild-type plants. Both sets of plants were grown on 1 mM NH4NO3 for 5 weeks and then transferred to nutrient solution without N for 4 d. A, Concentration dependence of 13NH4+ influx by both the HATS and the LATS into plant roots over a 10-min period. Data represent the averaged values of three independent experiments (n = 8-12 plants). B, Predicted NH4+ influx as a result of the LATS solely. Vmax values in the high-affinity range (0-250 µM) were subtracted from the individual data points for each concentration and averaged. Vmax values used were 8.34 µmol NH4+ g1 fresh weight h1 for the mutant and 11.4 µmol NH4+ g1 fresh weight h1 for the wild type.

Root and Shoot Tissue N Analysis

Analyses of root and shoot NH₄⁺, NO₃, Gln, and %N and %C revealed only minor differences (that were not statistically significant) between the amt1:1:T-DNA and wild-type lines regardless of N status (Table III). After 4 d of growth without exogenous N, %N decreased resulting in elevated C to N ratios in the shoots and roots of both lines.

View this table: [in this window] [in a new window]   Table III.   Root and shoot nitrogen and carbon pools from wild-type and amt1;1:T-DNA plants grown hydroponically for 5 weeks in the presence of 1 mM NH4NO3 (+N) or after a 4-d period where nitrogen was removed from the nutrient solution (N) Measurements are done on pooled samples of between 17 and 20 plants per treatment which were subdivided into three separate pools and individually extracted and measured. Data represent mean ± SE of the three separate extractions.

Competitive Reverse Transcriptase (RT)-PCR: Profile of mRNA Expression Patterns by the AtAMT1 and AtAMT2 Gene Families

Copy numbers of AtAMT mRNA transcripts in roots of wild-type and mutant plants maintained continuously in 1 mM NH₄NO₃ were obtained using competitive RT-PCR (for a typical reaction, see Fig. 6A). In wild-type plants, gene expression levels were in the order AtAMT1;2> AtAMT1;1> AtAMT1;3 > AtAMT2;1 (Fig. 6C). This hierarchy was relatively unchanged after 4 d of N deprivation, but copy numbers increased by approximately 1.9-, 1.5-, 1.9-, and 1.3-fold for AtAMT1;2, AtAMT1;1, AtAMT1;3, and AtAMT2;1 respectively. These increases were statistically significant (P < 0.05) for AtAMT1;1, AtAMT1;2, and AtAMT1;3. In the amt1;1:T-DNA line, AtAMT1;2 was also the most abundantly expressed AMT followed in order by AtAMT1;3 and AtAMT2;1. N deprivation increased AtAMT1;2, AtAMT1;3, and AtAMT2;1 expression by 1.6, 2.6, and 1.6 times, respectively. Interestingly for AtAMT1;3 and AtAMT2;1, these increases were significantly greater (P < 0.05) than the corresponding increases in wild-type plants (Fig. 6C).

View larger version (26K): [in this window] [in a new window]   Figure 6.   Quantitative AMT gene expression levels estimated using competitive RT-PCR. Plants were grown on 1 mM NH4NO3 for 5 weeks and then transferred to nutrient solution with (T = 0) or without (T = 4) N for 4 d before harvest of root tissues. A, A typical multiplex competitive RT-PCR performed using 25 ng of total root RNA mixed with known amounts of competitor cRNA. B, The quantity of AMT transcript present in each sample was determined from the ratio of endogenous and competitor cDNA products. C, Data represent the combined means and SE of five to six competitive RT-PCR experiments performed separately from two RNA extractions per treatment. Total RNA extracted from each treatment represented six to eight individual complete root systems. N.D., Not detectable. Bars with different letters are significantly different (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

amt1;1:T-DNA was identified from a set of enhancer-trap T-DNA tagged Arabidopsis lines in the Col background (Campisi et al., 1999). Sequence analysis revealed that the right border of the T-DNA tag was inserted 6 bp upstream of the first putative ATG of the open reading frame of AtAMT1;1 (Fig. 1A). Although the T-DNA tag is not inserted within the open reading frame of AtAMT1;1, its presence directly upstream of the coding sequence was sufficient to disrupt the transcription of AtAMT1;1 (Fig. 1D). No AtAMT1;1 mRNA transcripts were found in roots or shoots of amt1;1:T-DNA plants deprived of N for 4 d, a treatment, which resulted in high levels of AtAMT1;1 expression in wild-type plants (Fig. 1D). Failure to identify AtAMT1;1 mRNA supports the initial genome analysis using the PCR that our amt1;1:T-DNA is a homozygous line. However, the observed reduction in expression of AtAMT1;1 could also result from deleterious insertion or deletion by other T-DNA insertions elsewhere in the genome acting in trans to silence AtAMT1;1. Southern-blot analysis using a GUS DNA probe, which would identify other T-DNA insertions, revealed that there was only a single T-DNA insertion in the amt1;1:T-DNA genome (Fig. 1C).

The amt1;1:T-DNA and wild-type plants grown on adequate N (1 mM NH₄NO₃) showed similar rates of ¹³NH₄⁺ flux into roots in the high-affinity range (25 and 100 µM; Fig. 4A). The introduction of a 4-d N starvation period increased ¹³NH₄⁺ influx into roots for both lines (Fig. 4, A and B), however the flux in the amt1;1:T-DNA line was generally only 30% to 40% lower than in roots of wild-type plants. This phenotype was also present with plants grown on 2 mM KNO₃ in the presence of 1% (w/v) Suc followed by a 4-d N starvation period (data not shown). Derepression of root NH₄⁺ influx by N starvation is consistent with previous results in Arabidopsis (Gazzarrini et al., 1999; Rawat et al., 1999) and in rice (Wang et al., 1993). Failure to more substantially reduce NH₄⁺ influx after disrupting AtAMT1;1 suggests that this gene may encode a transporter that normally contributes only 30% to 40% of high-affinity NH₄⁺ influx, despite its suggested higher affinity for NH₄⁺ than other AtAMT1 transporters (Gazzarrini et al., 1999); certainly mRNA levels of AtAMT1;2 were actually higher than those of AtAMT1;1. As an alternative, AtAMT1;1 may quantitatively be a major player in high-affinity NH₄⁺ influx, and other NH₄⁺ transporters partially compensate for its disruption in the mutant. Estimated K_M values for ¹³NH₄⁺ influx were 11.3 ± 3.2 µM and 17.2 ± 4.4 µM for the wild type and amt1;1:T-DNA, respectively. Although not significantly different, there is a trend toward a higher K_M in the amt1;1:T-DNA line than the wild type, which is consistent with the estimated lower K_M values for AtAMT1;1 in yeast (approximately 0.5 µM), compared with values of approximately 40 µM for AtAMT1;2 and AtAMT1;3 (Gazzarrini et al., 1999). However, it is possible that the loss of AtAMT1;1 and its contribution to the plant's combined K_M for NH₄⁺ is being masked by the apparent compensation of other AMT proteins and possibly by other yet unidentified NH₄⁺ transporters. Our estimates of K_M values are substantially lower than were reported in another study in Arabidopsis (Rawat et al., 1999), which also used ¹³NH₄⁺ to measure influxes as a function of external NH₄⁺ concentration. These differences may be attributable to the use of sterile growth conditions within Magenta boxes or to the presence of Suc in the growth media in the study by Rawat et al. (1999). However, it was noted by Wang et al. (1993) that K_M values for NH₄⁺ influx in rice roots varied with N status of the plants.

When ¹³NH₄⁺ influx was measured at higher concentrations of NH₄⁺ (1-10 mM) in plants that had been deprived of N for 4 d, influx of ¹³NH₄⁺ through the low-affinity system had increased in the amt1;1:T-DNA line relative to wild-type plants (Fig. 5). This increased flux by the low-affinity system may be the result of plant compensation for the reduction of high-affinity NH₄⁺ flux caused by the loss of AtAMT1;1. Our earlier studies and those of other groups (see Glass and Siddiqi, 1995) have noted the negative correlations between high-affinity NH₄⁺ influx and accumulated N. In the absence of a functional AtAMT1;1 transporter, the reduced tissue N may result in elevated expression of mRNAs encoding all transporters that are regulated transcriptionally by tissue N and in particular Gln (Rawat et al., 1999). This may account for the increased LATS flux of ¹³NH₄⁺ in the present study and the maintenance of NH₄⁺ pools and %N in the mutant (Table III). However, the well-documented reduction of high-affinity NH₄⁺ influx associated with high-N tissue status failed to apply to low-affinity transport in rice, where ¹³NH₄⁺ influx attributable to LATS was increased in plants previously grown on 1 mM NH₄⁺ compared with plants grown on 2 µM or 100 µM NH₄⁺ (Wang et al., 1993). The apparent failure to down-regulate the LATS after growth on NH₄⁺ was also observed in roots of aspen (Populus spp.), lodgepole pine (Min et al., 2000), and citrus (Cerezo et al., 2001), even though the HATS was significantly down-regulated by this treatment. This anomaly is counterintuitive given that the HATS was down-regulated under these conditions and remains unexplained, but our present observation with the mutant may represent a similar phenomenon, whereby suppression of HATS is associated with increased influx through the LATS. Britto et al. (2001) suggest that at elevated NH₄⁺, which would repress the activity of the HATS, there is a failure to regulate LATS influx, resulting in an energetically costly active efflux of NH₄⁺ and a massive futile cycling across the plasma membrane. This may in part explain the toxic effects of elevated external NH₄⁺ on many plant species (Van der Eerden, 1982; Kronzucker et al., 2001). This observed increase of LATS flux suggests some form of interaction between the high- and low-affinity NH₄⁺ uptake systems, possibly through enhanced expression of NH₄⁺ regulated high-affinity AMT genes including AtAMT1;2, AtAMT1;3, and AtAMT2;1 (Gazzarrini et al., 1999; Shelden et al., 2001; this study). Shelden et al. (2001) have reported that AtAMT1;2 expressed in yeast displays biphasic kinetics, accumulating ¹⁴C-methylammonium at both low (0-100 µM) and high (0.25-10 mM) concentrations. It is possible that the low-affinity NH₄⁺ flux observed in this study may be the result of a switch in protein activity and NH₄⁺ affinity by AtAMT1;2 (Shelden et al., 2001). However, with the loss of AtAMT1;1 activity, AtAMT1;2 gene expression did not increase relative to the wild type as did AtAMT1;3 and AtAMT2;1, and therefore, from a mRNA transcript perspective, it appears less likely that AtAMT1;2 is the primary contributor to the increase in LATS flux in the amt1;1:T-DNA line. However as mentioned above, we cannot rule out a switch in AtAMT1;2 transport activity that may have occurred in the mutant in response to the loss of AtAMT1;1 activity. The protein AtAMT2;1, which is distantly related to members of the AMT1 family in Arabidopsis, has recently been identified and characterized as a putative low-capacity NH₄⁺ transporter (Sohlenkamp et al., 2000). AtAMT2;1 is expressed in both shoots and roots and responds to periods of N starvation (Sohlenkamp et al., 2000). Whether its activity is related to the increased flux of NH₄⁺ in the low-affinity range in the amt1;1:T-DNA line remains to be shown. However, as discussed below, its expression is enhanced in the amt1;1:T-DNA relative to the wild type making it a good candidate for part of this NH₄⁺ transport phenomenon.

To investigate the hypothesis that disruption of AtAMT1;1 was being compensated for by overexpression of other members of the AMT family we made use of competitive RT-PCR to measure total mRNA copy numbers and directly compare expression patterns among members of the AtAMT1 and AtAMT2 families. As expected, wild-type plants subjected to a 4-d period of N starvation increased levels of expression of all genes examined relative to plants grown with sufficient N over that period. This is a result that is consistent with other studies that have demonstrated the strong response of some members of the AMT family (AtAMT1;1 and AtAMT1;3) to N starvation (Gazzarrini et al., 1999; Rawat et al., 1999; Sohlenkamp et al., 2000; Gansel et al., 2001; Shelden et al., 2001). Interestingly, using the sensitive competitive RT-PCR technique we were able to observe consistent increases in AtAMT1;2 mRNA levels that have not been previously observed by other experimenters using total RNA northern blots (Gazzarrini et al., 1999; Shelden et al., 2001).

In wild-type plants, the order in which the AMT genes were expressed in roots after N starvation from most to least abundant was AtAMT1;2 > AtAMT1;1 > AtAMT1;3 > AtAMT2;1 (Fig. 6C). Apart from AtAMT1.1, this same pattern was evident in the amt1;1:T DNA line, however, levels of AtAMT1;3 and AtAMT2;1 were significantly higher than were observed in the wild type. We suggest that this elevated level of gene expression may have increased NH₄⁺ transport capacity through other AMT transporters, which has the effect of masking the loss of the contribution of AtAMT1;1 and which explains the small reduction in NH₄⁺ transport observed in the amt1;1:T-DNA line. This raises an important consideration vis à vis the effectiveness of the strategy of using particular gene mutations to determine the role of those genes. Where there is the capacity to compensate for the loss of activity, the role (particularly at the quantitative level) of disrupted genes may be obscured. The apparent compensation by AtAMT1;3 and AtAMT2;1 also raises the question of the mechanism of this putative compensation. The simplest explanation would be a form of passive compensation. Because the complete absence of AtAMT1;1 transporters should significantly reduce N uptake and tissue N concentration, down-regulation of all genes regulated by tissue N levels should be reduced, and these genes will consequently be overexpressed relative to their expression levels in wild-type plants.

By contrast, in a study of high-affinity NO₃ transport, there was a lack of compensation for disruption of the two high-affinity NO₃ transport systems AtNRT2;1 and AtNRT2;2 (Filleur et al., 2001). A T-DNA tagged mutant disrupted in two genes encoding the high-affinity NO₃ transporters AtNRT2;1 and AtNRT2;2 were shown to have a V_max for ¹⁵NO₃ influx that was only 27% of wild-type rates, with no apparent changes of flux values in the low-affinity range (Filleur et al., 2001). Because two genes have been inactivated in this line, their individual contributions to NO₃ flux may have exceeded any compensation in NO₃ uptake by other NRT2 proteins. As an alternative, it is possible that only AtNRT2;1 and AtNRT2;2 encode transporters involved in high-affinity NO₃ influx into roots, the other AtNRT2 genes possibly participating in internal redistribution of absorbed NO₃ such as NO₃ fluxes to the stele. It is also possible that NO₃ regulatory elements may have been disrupted that could limit subsequent activation of the NRT2 system during the synthesis of the T-DNA mutant, because 25 kb of genomic DNA was deleted with the T-DNA insertion. However, in another study of NO₃ transport in Aspergillus nidulans where there appear to be only two members of the NRT family and no other NO₃ transporters, loss of the high-capacity NRTA (formerly CRNA) gene reduced NO₃ influx to approximately 10% of wild-type levels with no apparent compensation by the second gene NRTB (Unkles et al., 2001).

There was no readily apparent phenotype in the amt1;1:T-DNA line when grown in soil in growth chambers under light conditions of approximately 150 to 200 µmol m² s¹ (B.N. Kaiser, unpublished data). Using an open-top hydroponic system, plants were grown on low (100 µM NH₄⁺ or KNO₃) or adequate (1 mM NH₄NO₃ or 2 mM KNO₃) N without external C under the same conditions used in the ¹³NH₄⁺ influx studies. Regardless of N form (NH₄⁺ or NO₃) or concentration, after 5 weeks there were little differences in fresh weights of roots and shoots between the amt1;1:T-DNA and the wild type with the exception that amt1;1:T-DNA plants grown in 100 µM KNO₃ had lower root and shoot fresh weights than wild-type plants (Table II). Correspondingly, %N, NO₃, NH₄⁺, and Gln levels in roots and shoots of the wild-type and amt1;1:T-DNA lines were similar (Table III).

Although little differences were observed in fresh weights, leaves of the T-DNA mutant were more succulent to the touch. Analysis of embedded cross sections of various aged leaves identified a developmental response with age (B.N. Kaiser, unpublished data) of increased cell density and reduction of intercellular air spaces specifically around the central vein, which was enlarged (Fig. 3). This phenotype was present in both hydroponically grown plants in a growth-chamber and in soil grown greenhouse plants both supplied nutrient solution with adequate levels of N (1 mM NH₄NO₃). Although we saw little difference between the wild-type and the amt1;1:T-DNA lines in ¹³NH₄⁺ influx into roots of plants supplied N, transport properties within the shoots could be entirely different. The loss of AtAMT1;1 activity may influence translocation of NH₄⁺ within leaves and the loading or unloading of NH₄⁺ into the vascular cylinders. The balance between NH₄⁺ and NO₃ in leaves has been suggested to influence leaf growth and cell division through indirect effects on cellular cytokinin levels and osmoticum balance (Walch-Liu et al., 2000). Although NO₃ pool sizes were similar in the shoot and roots of the amt1;1:T-DNA and wild-type lines, there were differences in the balance of NO₃ to NH₄⁺ pools (Table II). The marginally higher ratio of NO₃ to NH₄⁺ in the shoots of the amt1;1:T-DNA line may suggest that cells adjacent to the central vein are experiencing altered N ratios (NO₃>NH₄⁺) and possibly influencing localized cell development. Subsequent studies are planned to investigate the role of AtAMT1;1 on NH₄⁺ transport within leaves and its possible involvement in the regulation of cell division and N remobilization between developing plant organs.

Arabidopsis is often grown under sterile conditions in the presence of Suc (Gazzarrini et al., 1999; Rawat et al., 1999). In the presence of 1% (w/v) Suc and 0.5 mM (NH₄)₂SO₄, the amt1;1:T-DNA would germinate and produce their first set of leaves; but then during the 10 d after germination, plants grew poorly with little new leaf or root production (Fig. 2, A and B). By contrast, when 0.5 mM (NH₄)₂SO₄ was replaced by 1 mM KNO₃, amt1;1:T-DNA grew essentially as the wild-type plants (Fig. 2C). Growth of the amt1;1:T-DNA line with NH₄⁺ and Suc also resulted in reddening of the leaves caused by the accumulation of anthocyanins (Fig. 2A), an Arabidopsis phenotype that has been associated with stress responses including nutrient deficiencies and excess carbohydrate (Dangl, 1991; Martin et al., 2002). Martin et al. (2002) have recently demonstrated that the C to N ratio in Arabidopsis seedlings may regulate growth through possible effects on the mobilization of seed storage reserves and photosynthetic gene expression. Poor root and cotyledon development and accumulation of anthocyanins was significant in Arabidopsis seedlings grown in the presence of high levels of Suc (100 mM) and low N concentrations (100 µM; Martin et al., 2002).

It is interesting to speculate on the negative effects of NH₄⁺ provision in the presence of exogenous Suc in the amt1;1:T-DNA line. AtAMT1;1 mRNA expression has been examined previously under conditions where Suc was present in (Rawat et al., 1999) or absent from (Gazzarrini et al., 1999) the growth media. In the presence of Suc, AtAMT1;1 expression increased (up to 10-fold) and decreased dramatically within very short time periods (<24 h) in response to N starvation and resupply (Rawat et al., 1999). By contrast without Suc, the rate of change in AtAMT1;1 expression was much less rapid. In this study using competitive RT-PCR, we observed a 1.5-fold stimulation of AtAMT1;1 expression after a 4-d starvation period, whereas using traditional northern-blot analysis, Gansel et al. (2001) reported a 2-fold stimulation after a 5-d N starvation period, and Gazzarrini et al. (1999) a 3-fold increase after a 3-d N starvation period. The data of Rawat et al. (1999) suggests that Suc attenuates the activation of AtAMT1;1 expression. The mechanisms involved are unknown, however, it is possible that Suc may be acting as a component of a signaling cascade involved in maintaining C to N balance in plant tissues, with AtAMT1;1 possibly serving as the primary receptor involved in N sensing and NH₄⁺ transport per se. Although, the influx data demonstrated that ¹³NH₄⁺ influx through the LATS system was actually increased in the amt1;1:T-DNA line, and there were also compensatory effects on the expression of other members of the AMT family, the reduced influx of NH₄⁺ at moderate external concentrations (0-1 mM), which was also apparent in Suc grown plants (data not shown), may have been sufficient to disturb the C to N balance in the young developing seedlings resulting in the observed toxic effects of poor seedling development, root growth, and accumulation of anthocyanins.

In summary, we have generated a T-DNA mutant disrupted in the expression of AtAMT1.1. This mutation resulted in a reduction of approximately 30% to 40% in high-affinity root NH₄⁺ transport under N-limiting conditions. Albeit the reduction in high-affinity NH₄⁺ transport is less than initially expected from the strong correlative results of AtAMT1;1 expression and NH₄⁺ transport published by this group (Rawat et al., 1999) and others (Gazzarrini et al., 1999; Gansel et al., 2001), this is the first documented evidence, to our knowledge, that a member of the AMT family is an NH₄⁺ transporter in planta involved in NH₄⁺ uptake into plant roots. The failure to observe a more pronounced reduction in high-affinity NH₄⁺ transport appears to be the result of compensation by overexpression of other members (AtAMT1;3 and AtAMT2;1) of the AMT families of genes. Internal N pools and %N in roots and shoots were similar for the amt1;1:T-DNA and the wild-type lines. The results demonstrate that genetic redundancy in plants, in the case of the AMT family, provides an important ecological plasticity capable of adapting to mutation in nature. Subsequent analysis of other individual and multiple AMT disrupted mutants will help to resolve the question of the functions of each AMT gene and help to explain the compensatory phenomenon observed in this study. There are, nevertheless, as yet unexplained interactions leading to altered leaf morphology in the mutant and the significant interaction between C supply (in the form of Suc) and NH₄⁺ transport in the mutant, resulting in a highly lethal phenotype. These two phenotypes provide an exciting opportunity to examine the dynamic control of cell development by nutritional status controlled by the AMT family of NH₄⁺ transport proteins.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Growth

Arabidopsis Col-3 gl1 (wild type) and amt1-1:T-DNA (mutant) were grown either in a controlled chamber with a 25°C/20°C day/night temperature regimen in a hydroponic system (see below) or in peat-based soil media. Plants were illuminated by Vita-lite fluorescent lamps (Durotest, Fairfield NJ), which generated between 150 and 200 µmol m² s¹ of photosynthetically active radiation at plant level. During T-DNA isolation and subsequent seed multiplication and background crosses, plants were grown under long days (16:8 h light/dark), whereas for all ¹³NH₄⁺ influx experiments, northern analysis and competitive RT-PCR experiments, plants were grown vegetatively (8:16 h light/dark).

¹³N Influx Experiments

¹³NO₃ was produced by proton bombardment of H₂O by the cyclotron facility (TRIUMPH) of the University of British Columbia (Siddiqi et al., 1989). ¹³NO₃ was reduced to ¹³NH₄⁺ using Devarda's alloy, and ¹³NH₃ was distilled into acidified nutrient solution (Wang et al., 1993). Plants used for influx experiments were grown in an open-top liquid culture system or in sterile media in 200-mL Magenta boxes (Magenta Corporation, Chicago) or 500-mL plastic growth containers (Sigma-Aldrich, St. Louis). Seeds were held at 4°C for 4 d in sterile dH₂O and then seeded directly onto discs containing fine river sand (planting density, 1-3 seeds 1.5 cm²) for the open system or onto a porous nylon mesh placed on floating discs (Sigma-Aldrich) for growth under sterile conditions. The discs were supported on polystyrene platforms, which floated in an 8-L basin. In the open system, plants were typically grown in dH₂O for the first 10 d and then transferred to aerated complete nutrient solution (1 mM NH₄NO₃ except where indicated, 1 mM KH₂PO₄, 0.5 mM MgSO₄, 0.25 mM CaSO₄, 50 µM KCl, 25 µM H₃BO₃, 2.0 µM MnSO₄·H2O, 2.0 µM ZnSO₄·7H₂O, 0.5 µM CuSO₄·5H₂O, 0.5 µM Na₂MoO₄, 20 µM Fe-EDTA, and 200 mg of CaCO₃, pH 6.1) for 4 to 6 weeks. Nutrient solutions were replaced weekly (1 mM NH₄NO₃ plants) or daily (100 µM NH₄NO₃ plants). For plants grown under sterile conditions, the nutrient solution consisted of 1 mM KNO₃ except where indicated, 29 mM Suc, 2 mM KH₂PO₄, 1.0 mM MgSO₄, 1 mM CaCl₂, 25 µM H₃BO₃, 2.0 µM MnSO₄·H₂O, 2.0 µM ZnSO₄·7H₂O, 0.5 µM CuSO₄·5H₂O, 0.5 µM Na₂MoO₄, 20 µM Fe-EDTA, and 200 mg L¹ CaCO₃ at pH 6.1.

The influx experiments typically involved two pretreatments whereby 5-week-old plants were maintained for 4 more d on 1 mM NH₄NO₃, (N-sufficient plants) or on media without N (N-deprived plants). On the 4th d the plants were transferred to aerated uptake vessels for a 5-min pre-influx period in nonradioactive nutrient solution and then transferred to identical ¹³N-labeled NH₄⁺ nutrient solution for 10 min followed by a 3-min desorption period in nonradioactive nutrient solution. Plant roots and/or shoots were subjected to a 15-s low speed centrifugation step and counted in a -counter (Minaxi, Auto-gamma 5000 series, Packard, Downer's Grove, IL). For determination of ammonium, nitrate, Gln pools, and %N and C, hydroponically grown plants were immersed in N nutrient solution for 2 min, and then roots and shoots were frozen separately in liquid N₂. Frozen tissues (17-20 plants per treatment) were ground in liquid N₂ and lyophilized in a freeze drier for 4 d at 50°C. Three individual 20-mg fractions of dried tissue were extracted three times with 1 mL of 10 mM sodium acetate (pH 6.42), 4°C. Samples were centrifuged for 5 min at 14,000g (4°C) after each extraction. Pooled supernatant was subjected to three rounds of freeze (liquid N₂) then thaw (iced water bath) periods and then centrifuged for 15 min at 14,000g (4°C). Samples were filtered (0.22 µM) and stored at 20°C. NO₃ pools were estimated following the protocol of Cataldo et al. (1975). NH₄⁺ and Gln pools were determined by derivatization of the extracts with AccQ (Waters, Milford, MA) followed by separation by HPLC (1090 LC, Hewlett-Packard, Palo Alto, CA) using a 4.6- × 25-mm C-18 column followed by fluorescent detection. Total N and C were measured from oven dried lyophilized tissues using an elemental analyzer (EA110 CHN-O, CE Instruments, Milan).

Isolation of amt1;1:T-DNA

The amt1;1:T-DNA line was identified from a pool of approximately 30,000 transfer DNA (T-DNA) lines composed of both the Feldmann CD5 (CD5 1-6; Forsthoefel et al., 1992) and "Jack" CD6 (CD6 1-6) series (Campisi et al., 1999). Extracted DNAs from subpools of these lines were obtained from the Arabidopsis Biological Resource Center (Ohio State University). The subpools of DNA from each set of lines were arranged in two multiplex arrays and screened using a PCR-based reverse genetic approach (Krysan et al., 1996) with forward (P1) and reverse (P2) oligonucleotides specific to AtAMT1;1. These oligonucleotides were used in conjunction with right (P3 and P4) and left (P5 and P6) border oligonucleotides of the T-DNA insertional elements of pD991 (Jack lines) and the 3850:1003 Ti plasmid (Feldman lines), respectively. Each 25-µL PCR contained 0.02 to 0.05 µg of pooled DNA, 25 pmol of each primer, 0.1 mM dNTPs, 2.5 µL of 10× PCR buffer, and 1 unit of Taq DNA polymerase (Invitrogen, Carlsbad, CA). PCR products were size-separated on 1% (w/v) agarose 1× Tris-acetate EDTA gels and blotted in 20× SSC onto nylon membrane (Hybond N⁺, Amersham Biosciences UK, Ltd., Little Chalfont, Buckinghamshire, UK; Sambrook et al., 1989). Blots were probed with a 1.7-kb random-primed [³²P]dCTP labeled AtAMT1;1 cDNA followed by high-stringency washes (Rawat et al., 1999). PCR products that hybridized to the AtAMT1;1 cDNA were cloned into TOPO-TA (Invitrogen) and sequenced using Big Dye terminators (Applied Biosystems, Foster City, CA). A T-DNA insertion in AtAMT1;1 was identified in the Jack lines subpool CD6-4A. Twenty DNA subpools from CD6-4A (3,001-4,000) of 100 lines each (CD6-71 through CD6-90) were subjected to a PCR with primers P3 and P2 and Southern hybridization as described above. A positive signal in CD6-74 and CD6-85 was detected in the multiplex array of DNA pools. On the basis of this analysis, 3,350 was selected as the positive subpool. Approximately 300 seeds from subpool 3,350 were sown individually with a Pasteur pipette into separate pots with fine soil mix. Leaf tissue from 3-week-old seedlings was harvested for genomic DNA extraction according to McKinney et al. (1995). Using primers P2 and P3, a PCR was completed on each individual DNA sample, and amplified products were subjected to Southern hybridization with the AtAMT1;1 cDNA as described earlier. Three lines were identified yielding a positive signal for T-DNA insertion. The plants were grown to complete maturity, and the seeds were harvested. The Arabidopsis plants yielding a positive signal for a T-DNA insertion in AtAMT1;1 were backcrossed once to the Arabidopsis parental line (Col-3 gl1). Homozygous lines were then selected using the PCR with two sets of PCR primers. The first set (primers P3 and P7) were designed to identify the T-DNA insertion by amplifying between the right border of the T-DNA and the 3' end of AtAMT1;1. The second set of primers (P8 and P7) were used to identify the non-tagged AtAMT1;1 allele by amplifying from 670 bp upstream of the first putative ATG through to the 3' end of the of AtAMT1;1 gene. Putative homozygous lines were selected and analyzed further by Southern analysis and inverse PCR.

Southern Analysis and Amplification of Flanking T-DNA Genomic Sequences

Genomic DNA was isolated by cetyl-trimethyl-ammonium bromide extraction (Murray and Thomson, 1980) from both Col-3 gl1 and amt1;1:T-DNA leaves and digested with EcoRI and HindIII. Twenty micrograms of digested DNA was separated on a 0.8% (w/v) agarose 1× Tris-acetate EDTA gel and blotted onto nylon membranes in 20× SSC. The DNA was fixed to the membrane by baking at 120°C for 30 min. Blots were prehybridized in DIG-easy prehybridization mix (Roche Diagnostics, Indianapolis) at 42°C for 1 h and then incubated with PCR-amplified DIG-labeled uidA (GUS; accession no. A00196). The GUS cDNA was amplified from genomic DNA isolated from amt1;1:T-DNA using the PCR with primers P21 and P22. Amplified products were cloned into pGEMTeasy (Promega, Madison, WI) and sequenced. After hybridization, all membranes were washed twice for 15 min in 2× SSC, 1% (w/v) SDS at ambient temperature, twice at 68°C for 30 min in 0.1× SSC, 1% (w/v) SDS, and twice at ambient temperature in 0.1× SCC, 0.1% (w/v) SDS. Digoxygenin was detected using a commercial kit (Roche Diagnostics).

Inverse PCR was used to identify the nucleotide sequence adjacent to left border of the T-DNA insertion in amt1;1:T-DNA. Genomic DNA was isolated from amt1;1:T-DNA and digested with EcoRI. After digestion, the genomic DNA fragments were religated with T4 DNA ligase at 37°C for 2 h and then incubated at 16°C for a further 12 h. The ligated DNA was then subjected to a PCR using primers P23 and P24, which anneal in opposite directions to the left border of the T-DNA insert. Amplified products of the expected size were cloned into pGEMTeasy, and the cDNA insert was sequenced.

Northern Analysis

Total RNA was extracted from frozen plant tissues harvested from 4- to 6-week-old wild-type and amt1;1:T-DNA plants (roots and shoots) using the RNAeasy system (Qiagen USA, Valencia, CA). For northern analysis, 10 µg of total RNA per tissue was size separated on a 1× MOPS 1.2% (w/v) agarose gel containing formaldehyde (Sambrook et al., 1989) and blotted overnight onto Hybond N⁺ nylon membrane in 20× SSC. RNA was fixed to the membrane by baking at 120°C for 30 min. Blots were hybridized with full length DIG-labeled antisense AtAMT1;1 RNA produced using the SP6/T7 RNA DIG-labeling kit (Roche Diagnostics). Blots were hybridized overnight at 68°C in DIG-easy hybridization buffer (Roche Diagnostics). After hybridization, the blots were washed twice for 15 min in 2× SSC, 1% (w/v) SDS at ambient temperature, twice at 68°C for 30 min in 0.1× SSC, 1% (w/v) SDS, and twice for 15 min at ambient temperature in 0.1× SCC, 0.1% (w/v) SDS followed by detection of the digoxygenin label as previously described.

Competitive RT-PCR

Competitor RNA templates were prepared using the RT-PCR competitor construction kit (Ambion, Austin, TX). Each RNA competitor was transcribed from a modified deletion DNA template using T7 RNA polymerase. The AtAMT1;1 DNA template was constructed by cloning AtAMT1;1 (Rawat et al., 1999) into the NotI site of pSPORT2 (Invitrogen) followed by digestion with STYI, which removed 560 bp of the cDNA. The terminal STYI restriction sites of the gel extracted (QiaxII, Qiagen USA) pSPORT2/AtAMT1;1 deletion construct were ligated and the circularized plasmid amplified. The AtAMT2;1 and AtAMT1;3 deletion constructs were both prepared by first amplifying AtAMT2;1 and AtAMT1;3 cDNAs using RT-PCR on total RNA extracted from hydroponically grown Arabidopsis roots using PCR primers P9 and P10 for AtAMT2;1 and P11 and P12 for AtAMT1;3. Both PCR products were cloned into pGEMTeasy and sequenced. The 1.2-kb AtAMT2;1 fragment was inserted into the NOTI site of pSPORT2 and digested with STYI, which liberated a 658-bp fragment. The remaining pSPORT2/AtAMT2;1 construct was gel purified as above, and the terminal STYI sites were ligated. The pGEMTeasy/AtAMT1;3 construct was digested with AVAI, which liberated a 101-bp fragment. The remaining construct was gel extracted, and the terminal AVAI ends were ligated. For each of the above deletion constructs, a T7 RNA polymerase binding site was added to the 5' end of the cDNA using the PCR and primer pairs (P13 and P14 for AtAMT1;1, P15 and P10 for AtAMT2;1, and P16 and P12 for AtAMT1;3). The AtAMT1;2 deletion template was synthesized using the PCR on AtAMT1;2 cDNA (Shelden et al., 2000) with primers P24 and P25, which deleted 435 bp from the 5' end of AtAMT1;2 and added T7 RNA polymerase-binding site. PCR products were gel excised and purified (QiaxII, Qiagen USA). cRNA was generated using T7-RNA polymerase and RNase-resistant modified 2'-CTP (Ambion) in the presence of [³²P]GTP. Products were size fractionated on a 4% (w/v) polyacrylamide 1× TBE gel containing 8 M urea. After autoradiography, the band of expected size was excised, and the RNase-resistant cRNA was eluted. The cRNA was quantified by measuring [³²P]GTP incorporation using a scintillation counter (BeckmanCoulter, Fullerton, CA). Total RNA was extracted using the RNAeasy kit (Qiagen USA) from combined roots of six to eight individual plants per treatment. Total RNA was treated with DNase (DNase later, Ambion) and quantitated by UV spectroscopy. For the competitive RT-PCR experiments, one-step RT-PCR (Qiagen USA) was performed on total root RNA (25 ng) mixed with various concentrations of RNase-resistant competitor cRNA using primers P13 and P17 for AtAMT1;1, P26 and P18 for AtAMT1;2, P11 and P19 for AtAMT1;3, and P9 and P20 for AtAMT2;1. Amplified products were separated on agarose gels followed by staining with SYBR Gold (Molecular Dynamics, Sunnyvale, CA) and fluorescent signal immediately detected using a STORM imager (Amersham Biosciences).

Leaf Ultrastructure

Leaves of different ages were collected and sections sampled from the midway point between the tip of the leaf and the base of the petiole. Leaf samples were taken from plants grown either in the greenhouse in soil media with no supplementary lighting or in hydroponic tanks as described for the ¹³NH₄⁺ influx experiments. Leaf samples that included the major vein were fixed overnight at 4°C in 2.5% (v/v) glutaraldehyde in 50 mM Na₂HPO₄, pH 7.0. Samples were washed and postfixed in 1% (v/v) OsO₄ in 25 mM Na₂HPO₄ (pH 7.0) at room temperature for 1 h and then washed in 50 mM K₂HPO₄ (pH 7.0). Sections were dehydrated in successive ethanol washes (10%, 25%, 50%, 75%, 95%, and 100% [v/v]) and infiltrated and embedded in Spurr's resin at 60°C overnight. Sections (3 µM) were cut on a microtome (Leica Reichert Jung, Wetzlar, Germany) and stained with 0.5% (v/v) toluidine blue dissolved in 1% (w/v) sodium tetraborate (pH 9.5). Final sections were analyzed on a microscope (Zeiss, Welwyn Garden City, UK) and images captured using a digital camera (SPOT: Diagnostic Instruments, Sterling Heights, MI).

Statistical Analysis

Where reference to significant differences were made, analysis of variance of means was performed on the individual data points followed by two-tailed t tests (P < 0.05).