Kinetic properties and ammonium-dependent regulation of cytosolic isoenzymes of glutamine synthetase in Arabidopsis.

Glutamine synthetase (GS; EC 6.3.1.2) is a key enzyme of nitrogen assimilation, catalyzing the synthesis of glutamine from ammonium and glutamate. In Arabidopsis, cytosolic GS (GS1) was accumulated in roots when plants were excessively supplied with ammonium; however, the GS activity was controlled at a constant level. The discrepancy between the protein content and enzyme activity of GS1 was attributable to the kinetic properties and expression of four distinct isoenzymes encoded by GLN1;1, GLN1;2, GLN1;3 and GLN1;4, genes that function complementary to each other in Arabidopsis roots. GLN1;2 was the only isoenzyme significantly up-regulated by ammonium, which correlated with the rapid increase in total GS1 protein. GLN1;2 was localized in the vasculature and exhibited low affinities to ammonium (Km = 2450 +/- 150 microm) and glutamate (Km = 3.8 +/- 0.2 mm). The expression of the counterpart vascular tissue-localizing low affinity isoenzyme, GLN1;3, was not stimulated by ammonium; however, the enzyme activity of GLN1;3 was significantly inhibited by a high concentration of glutamate. By contrast, the high affinity isoenzyme, GLN1;1 (Km for ammonium < 10 microm; Km for glutamate = 1.1 +/- 0.4 mm) was abundantly accumulated in the surface layers of roots during nitrogen limitation and was down-regulated by ammonium excess. GLN1;4 was another high affinity-type GS1 expressed in nitrogen-starved plants but was 10-fold less abundant than GLN1;1. These results suggested that dynamic regulations of high and low affinity GS1 isoenzymes at the levels of mRNA and enzyme activities are dependent on nitrogen availabilities and may contribute to the homeostatic control of glutamine synthesis in Arabidopsis roots.

the so-called GS/GOGAT cycle. Plants have two types of GS isoenzymes that localize in different compartments: one located in the cytosol (GS1) and the other in the plastid/chloroplasts (GS2) (1)(2)(3)(4). GS1 is the major form of GS in plant roots, and the ammonium taken up from the soil is directly converted to Gln by its reaction. Molecular biological studies have identified a number of genes encoding GS1 from various plant species (5)(6)(7)(8)(9). The presence of multiple GS1 isoenzymes complicates the overall understanding of their physiological functions. The isoenzymes of GS1 show organ-and cell-specific patterns of expression and are developmentally regulated (10 -17). In addition, the expression of GS1 is metabolically regulated by the availability of nitrogen and carbon sources (18 -22, 24).
In the roots of legumes, GS1 is regulated by ammonium supplied from the environment or the symbiotic nitrogen fixation (18 -21). Transgenic studies with the soybean GS1 promoter suggested that ammonium-dependent regulation is specific for nitrogen assimilation in leguminous plants; the soybean-derived GS1 promoter was not able to display the ammonium-induced expression of reporter activity in tobacco (20). This poses questions of whether the stimulation of GS1 expression by ammonium is genuinely associated with symbiosis and whether dicotyledonous plants are furnished or not furnished with the regulatory mechanisms that may perceive the ammonium signal. The availability of ammonium also caused changes in the expression of some of the GS1 genes in non-leguminous plants. For example, among the five isoenzymes of GS1 (GS1a, GS1b, GS1c, GS1d, and GS1e) identified from maize (9,13), the mRNA of GS1c and GS1d accumulated following application of ammonium to roots, whereas those of GS1a and GS1b were suppressed under the same condition (22).
In Arabidopsis, five putative genes for GS1, GLN1;1, GLN1;2, GLN1;3, GLN1;4, and GLN1;5, are encoded in the genome (23). Oliveira and Coruzzi (24) reported that the application of sucrose to the medium causes induction of GLN1;1, GLN1;2, and GLN1;3 mRNAs and that amino acids attenuated the effect of sucrose, suggesting that a metabolic regulation of GS1 is associated with the relative abundance of carbon skeleton versus amino acids accumulated in the root tissue. These results suggest a negative feedback regulation of GS1 by Gln or the downstream nitrogen metabolites as in the case of ammonium transporters. However, the exact functional roles and physiological diversities of the individual GS1 isoenzymes in Arabidopsis have not been well characterized.
In the present study, we determined the kinetic properties and ammonium response of the Arabidopsis GS1 isoenzymes. We demonstrated that GS1 expressed in roots (GLN1;1, GLN1;2, GLN1;3, and GLN1;4) can be classified into two distinct groups, the high affinity and low affinity enzymes localized at the root surface and vasculature, respectively. The characteristics of the GS1 isoenzymes in Arabidopsis were fitted into the negative regulatory pathways of ammonium influx and assimilation, which is apparently different from those operated positively for the active assimilation of externally supplied ammonium in the roots of legumes and maize.
The results presented here clearly demonstrated that the multiplicity of cytosolic GS1 in higher plants does not simply provide functional redundancy but, rather, confers specified roles to the individual isoenzymes bearing distinctive kinetic properties and ammonium responsiveness. This is the first report presenting the complete and precise data of enzymatic properties and cell type-specific localization of GS1 isoenzymes in higher plants. EXPERIMENTAL (25). Arabidopsis cell culture (26) was propagated in Murashige and Skoog liquid medium supplemented with 2.5 mM potassium phosphate and 4.5 M 2,4-dichlorophenoxyacetic acid by continuous shaking under dark conditions.
Quantitative Real Time PCR Analysis-Extraction of RNA and cDNA synthesis was carried out as described above for cDNA cloning. Genespecific primers are presented in Table I. Constitutive expression of ubiquitin (UBQ2; GenBank TM accession number J05508) was confirmed in parallel (Table I). The PCR products were detected and quantified as SYBR Green fluorescence (Applied Biosystems, Foster City, CA) using the Gene Amp 5700 sequence detection system (Applied Biosystems). The mRNA content was quantitatively determined by using a purified cDNA clone as a standard for its calibration.
Preparation of Anti-GS1 Antibody-The cDNA fragment encoding rice GS1, RGS28 (GenBank TM accession number X14245) (7) was inserted into the NcoI site of pTrc99A (Amersham Biosciences). The nucleotide sequence downstream of the translation initiation codon was modified to an AT-rich structure to obtain high expression of the recombinant GS1 protein (22). The first 24-bp portion of the RGS28 coding sequence was deleted by HincII digestion and replaced with the oligonucleotide 5Ј-TCGAGCCATGGCTTCTTTAACTGATCTCGTC-3Ј without changing the encoding amino acid residues. The 3Ј-end of RGS28 was digested with PflMI and ligated with the oligonucleotide 5Ј-CTGGAAGCCCCATCATCATCATCATCATTGACCATCAT-3Ј, adding His 6 codons and a stop codon at the C terminus. The modified RGS28 was cloned in pTrc99A and transformed to Escherichia coli JM109. Culture of E. coli and purification of the recombinant protein were performed as described previously (22). To prepare polyclonal antibody against the rice recombinant GS1 protein, 500 g of the purified antigen emulsified with adjuvant was injected into a rabbit. Additional injections (500 g each) were done twice every 2 weeks. Whole blood was obtained from the carotid artery, and the antiserum was collected by centrifugation.
Western Blot Analysis-Proteins were separated in 12.5% (w/v) polyacrylamide gel and transferred to a polyvinylidene difluoride membrane (Bio-Rad) by electroblotting. The amount of proteins applied to the gel is indicated in the figure legends. The membrane was incubated with the anti-GS1 polyclonal antibody and goat anti-rabbit IgG alkaline phosphatase conjugate (Promega, Madison, WI) as described previously (28). Proteins cross-reacting with the antibodies were visualized using 5-bromo-4-chloro-3-indolylphosphate-p-toluidine and nitroblue tetrazolium chloride (Promega).
GS Enzyme Assay-The Gln synthetic activity of GS was determined by quantifying L-Gln synthesized from ammonium and L-Glu (28). A 0.1-ml reaction mixture contained 100 mM tricine-HCl (pH 7.8), 80 mM L-Glu, 6 mM ammonium, 8 mM ATP, 20 mM MgSO 4 , and 8 mM 2-mercaptoethanol. The amount of proteins used for the enzyme reaction is indicated in Figs. 2 and 6 and Table II. The reaction mixture was incubated at 30°C for 15 min, terminated by heating at 96°C for 2 min, and filtered through Ultra free-MC 5000 NMWL Filter Unit (Millipore, Bedford, MA). Five microliters of filtrate was incubated with AccQ Fluor reagent (Waters, Milford, MA), and the resulting AccQ-derivative of Gln was measured by Waters 2695 high performance liquid chromatography system on an AccQ Tag column (3.9 mm diameter ϫ 150 mm length; Waters) according to the manufacturer's instructions.
Extraction of Proteins from Arabidopsis Plants-Approximately 0.2 g of frozen root or leaf tissues were homogenized in an equal volume of GS extraction buffer (50 mM Tris-HCl pH 7.6, 10 mM MgCl 2 , 10 mM 2-mercaptoethanol, and 5% (w/v) polyvinylpyrrolidone) (28) using a chilled mortar and pestle or with MM-300 mixer mill (Qiagen). The homogenates were centrifuged at 20,000 ϫ g for 10 min under 4°C. The supernatants were immediately desalted on PD-10 (Amersham Biosciences) or Probe Quant G-50 Micro Columns (Amersham Biosciences) equilibrated with buffer A (25 mM Tris-HCl, pH 7.6, 1 mM MgCl 2 , and 10 mM 2-mercaptoethanol). The crude enzyme fractions obtained were stored in 50% (v/v) glycerol. Soluble protein content was determined by the Coomassie Blue dye binding method (Bio-Rad) using bovine serum albumin as a standard.
Purification of Native GS1 Protein from Arabidopsis Cell Culture-All of the steps of enzyme purification were performed in a cold room (4°C). One hundred fifty grams of cultured cells were homogenized in 500 ml of GS extraction buffer using a chilled warring blender (28). The homogenate was centrifuged at 27,000 ϫ g for 15 min. The supernatant was fractionated by (NH 4 ) 2 SO 4 precipitation between 30 and 60% saturation. The precipitate was suspended in 50 ml of buffer A containing 1.2 M (NH 4 ) 2 SO 4 and then immediately loaded onto a HiTrap butyl-FF column (15 ml bed volume) (Amersham Biosciences) equilibrated with the same buffer. The pooled fraction was saturated to 70% (NH 4 ) 2 SO 4 and centrifuged at 27,000 ϫ g for 15 min. The precipitate was suspended in 2 ml of buffer A and loaded for gel filtration chromatography on a HiLoad 26/60 Superdex 200pg (320 ml bed volume) (Amersham Biosciences) equilibrated with buffer A. The fractions containing GS activities were further purified by anion exchange chromatography using a HiTrap Q HP (1-ml bed volume) (Amersham Biosciences) equilibrated with buffer A. Protein was eluted by a linear gradient of NaCl from 0 to 1 M.
Expression of Recombinant GS1 Proteins in E. coli-The KpnI and SalI sites were created on the 5Ј-and 3Ј-ends of the coding region of GLN1 cDNAs by PCR using KOD-Plus DNA polymerase (Toyobo). The amplified PCR products were cloned into pCR-Blunt II-TOPO (Invitrogen) and fully sequenced. The cDNA inserts were digested by KpnI and SalI and ligated into the KpnI and SalI sites of pQE2 (Qiagen) to produce the His 6 fusion constructs, termed pQE2-GLN1;1, pQE2-GLN1;2, pQE2-GLN1;3, and pQE2-GLN1;4, respectively. E. coli HB101 was transformed with the recombinant construct and grown at 30°C in 300 ml of LB medium containing 1 M sorbitol and 100 g/ml ampicillin. When A 600 reached 0.5, the culture was cooled down to 20°C. One millimolar isopropyl-1-thio-␤-D-galactopyranoside was supplied to induce the expression of recombinant GS1 proteins, and the culture was further incubated at 20°C for 12 h by continuous shaking (180 rpm).
Purification of Recombinant GS1 Enzymes-One volume of E. coli cell pellets collected by centrifugation was re-suspended in 5 volumes of lysis buffer (50 mM NaH 2 PO 4 , pH 8, 300 mM NaCl, and 10 mM imidazole). Cell suspensions were incubated for 30 min at 4°C with lysozyme (Nacalai Tesque, Kyoto, Japan) at the final concentration of 1 mg ml Ϫ1 . After ultrasonication, the insoluble fraction was removed by centrifugation. The His-tagged protein in the soluble fraction was affinitypurified using a nickel-nitrilotriacetic acid agarose column (1-ml bed volume) (Qiagen). Washing was carried out with a buffer containing 50 mM NaH 2 PO 4 , pH 8, 300 mM NaCl, and 20 mM imidazole. Protein was eluted with the same buffer containing 250 mM imidazole. The eluates were desalted and equilibrated in buffer A using PD-10 (Amersham Biosciences). Protein was further purified by anion exchange chromatography with a HiTrap Q HP (1-ml bed volume) (Amersham Biosciences) using the AKTA Prime fast protein liquid chromatography system (Amersham Biosciences). Protein was eluted by 10 ml of a linear NaCl gradient from 0 to 1 M, and the protein peak fractions were analyzed for the presence of a single band of recombinant GS1 protein (40 kDa) on a Coomassie Blue-stained polyacrylamide gel. The Gln biosynthetic activities of the pooled fractions were assayed to determine the kinetic properties.
Construction of GLN1 Promoter-Green Fluorescent Protein (GFP) Plants-The fusion gene constructs of Arabidopsis GLN1 gene promoters and EGFP (Clontech) for plant transformation were constructed as follows. The 2463-bp, 2501-bp, 1213-bp, and 1161-bp 5Ј-promoter regions of GLN1;1, GLN1;2, GLN1;3, and GLN1;4, respectively, were amplified from Col-0 Arabidopsis genomic DNA by PCR using KOD plus DNA polymerase (Toyobo). The BamHI site was created on the 5Ј-end of the forward primer. The reverse primer containing the NcoI restriction site (CCATGG) was created at the translation initiation site of GLN1 gene to make the promoter-GFP fusion. The amplified PCR products were cloned into pCR-Blunt II-TOPO (Invitrogen) and fully sequenced. The GLN1 promoter regions were cut out as BamHI-NcoI fragments and ligated with the NcoI-NotI fragment of EGFP (Clontech) and the NotI-EcoRI fragment of the nopaline synthase terminator (29). The resultant GLN1 promoter:EGFP:nopaline synthase terminator cassette was placed in the position of the ␤-glucuronidase and nopaline synthase terminator in the binary plasmid, pBI101 (Clontech). The binary plasmids were transferred to Agrobacterium tumefaciens GV3101 (pMP90) (30) by the freeze-thaw method (31). Arabidopsis plants were transformed according to the floral dip method (32). Transgenic plants were selected on GM medium (33) containing 50 mg liter Ϫ1 kanamycin sulfate. Fluorescence of GFP in transgenic plants was observed under a BX61 microscope equipped with a FV500 confocal laser scanning system and a 505-525-nm band pass filter (Olympus, Tokyo, Japan), as descried previously (34).

Cytosolic GS1 Is the Major Form in Arabidopsis Roots-
The mRNA and protein contents of GS1 and GS2 in Arabidopsis were determined by real time PCR and Western blot analyses (Fig. 1). Arabidopsis plants were grown on agar medium containing 7 mM nitrate as a sole nitrogen source for 14 days. The total GS1 mRNA was calculated as a sum of GLN1;1, GLN1;2, GLN1;3, GLN1;4, and GLN1;5 mRNAs. The gene-specific primer pairs that specifically amplify the transcripts of individual GS isoenzymes are described in Table I. The GS1 mRNA was ϳ20-fold more abundant than GS2 mRNA in Arabidopsis roots (Fig. 1A). Western blot analysis indicated accumulation of a 40-kDa protein that corresponds to the molecular size of GS1 in roots (Fig. 1B). GS2 (44 kDa) was hardly detectable in roots (Fig. 1B). The results indicate that GS2 mRNA is weakly expressed in roots as a minor GS isoenzyme; however, judging from the Western blots, the contribution of GS2 to the Gln synthesis is not appreciable as compared with GS1. Thus, we concluded that GS1 is the major isoenzyme that carries out the ammonium assimilation in Arabidopsis roots. In leaves, the mRNAs and proteins for both isoenzymes were expressed. On real time PCR, GS2 mRNA was slightly more abundant than GS1 mRNA in leaves (Fig. 1A). Western blot showed two major bands corresponding to the molecular sizes of GS2 and GS1 subunit proteins, respectively (Fig. 1B).
Increased Accumulation of GS1 Protein by Ammonium Treatment Is Not Correlated with the GS Activity-The enzyme activity and the abundance of GS1 polypeptides were simultaneously monitored during a time course of ammonium treatment. Two week-old plants were deprived of nitrogen for 3 days and treated with 10 mM ammonium chloride. Root tissues were harvested for Western blotting and GS enzyme assays (Fig. 2). The amount of GS1 protein was significantly higher in the NH 4 ϩ -treated plants than in the control without nitrogen (Fig.  2, A and B). The abundance of GS1 protein was increased 3-fold within 6 h after transferring the plants to ammonium excess conditions (Fig. 2B). However, under the same condition, the activity of GS was not changed remarkably (Fig. 2C). No significant differences were observed between the GS activities of the ammonium-treated and control plants.
The mRNA contents of individual GS1 isoenzymes were quantified by real-time PCR under nitrogen limitation and ammonium excess (Fig. 3). The gene-specific primer pairs that specifically amplify GLN1 cDNAs are described in Table I. GLN1;1, GLN1;2, GLN1;3, and GLN1;4 were expressed in roots, but the GLN1;5 transcript was not detected. In addition, microarray analysis for Arabidopsis root reported that a signal value for the GLN1;5 mRNA level indicates an "absent" call (35). Thus, we quantified the expression of the four GLN1 genes, GLN1;1, GLN1;2, GLN1;3, and GLN1;4, in roots. Among them, GLN1;1 mRNA was abundantly expressed in roots after 3 days of nitrogen deprivation (Fig. 3A). The amounts of GLN1;1, GLN1;3, and GLN1;4 mRNAs decreased ϳ0.5, 0.7, and 0.3-fold, respectively, after the application of 10 mM ammonium to the medium (Fig. 3B). Conversely, GLN1;2 mRNA was increased ϳ5-fold by ammonium (Fig. 3B) and maintained at high levels during 24 h of ammonium treatment (Fig. 4A). The time course increase in GLN1;2 mRNA ( Fig. 4A) corresponded with the accumulation of GS1 protein detected on the Western blot (Fig. 2A). These results suggest that the accumulation of GS1 protein under ammonium excess conditions is associated with the increased expression of GLN1;2 mRNA.
As shown in Figs. 2-4, the increase in GLN1;2 mRNA and the corresponding GS1 protein contents was not correlated with the total GS activities in Arabidopsis roots. To clear these discrepancies, we determined the enzymatic properties and cell typespecific localization of GS1 isoenzymes in Arabidopsis roots.
Purification of Individual Arabidopsis GS1 Isoenzyme-GLN1;1, GLN1;2, GLN1;3, and GLN1;4 were overexpressed in E. coli as His-tagged enzymes and purified to near homogeneity. A native GS1 protein was purified from Arabidopsis cell culture for reference. Coomassie Blue staining of SDS-PAGE

FIG. 1. GS mRNAs and polypeptides in Arabidopsis roots and leaves.
A, GS mRNA contents quantified by real time PCR. Arabidopsis plants were grown for 2 weeks on agar medium containing 7 mM nitrate. Real time PCR was carried out on first strand cDNA synthesized from total RNA as described under "Experimental Procedures." B, accumulation of GS polypeptides detected by the anti-rice GS1 antibody. Western blot analysis was carried out against 5 g of total soluble proteins extracted from root (R) and leaf (L) tissues.
indicates a single polypeptide band in the purified fraction that corresponds to the molecular size of GS1 (Fig. 5A). The purified proteins equally cross-reacted with the anti-rice GS1 antibody, which shows nearly the same affinities against the four Arabidopsis GS1 isoenzymes (Fig. 5B).
Plant GS proteins are made up of eight subunits and have molecular masses ranging from 320 to 380 kDa (1). To determine whether recombinant GS1 enzymes are properly assembled in an octameric structure, the molecular masses of the purified GS1 proteins were estimated under a native condition by gel filtration chromatography calibrated with protein markers of a known size (Fig. 5C). Both of the GS1 proteins purified from the cell culture and recombinant systems showed a similar molecular size of 340 -350 kDa, which is consistent with the octameric composition of ϳ40 kDa subunits (Fig. 5C). From these results, we concluded that the recombinant enzymes are functionally equivalent to the native GS protein derived from Arabidopsis cell culture.
The Arabidopsis GS1 Isoenzymes Exhibit Different Kinetic Properties-The kinetic properties for ammonium, glutamate, and ATP were determined by assaying the Gln biosynthetic activities of individual GLN1 isoenzymes as described under "Experimental Procedures." As shown in Table II, the four isoenzymes of GS1 expressed in Arabidopsis roots can be classified into two major groups by the affinities to substrates. GLN1;1 exhibited an extremely high affinity to ammonium (K m Ͻ 10 M). GLN1;4 also showed a relatively high affinity to ammonium (K m ϭ 48 Ϯ 6.0 M). By contrast, the affinities of GLN1;2 and GLN1;3 to ammonium were lower than those of GLN1;1 and GLN1;4 by two orders (K m values; 2450 Ϯ 150 M for GLN1;2 and 1210 Ϯ 40 M for GLN1;3, respectively). GLN1;2 and GLN1;3 can be classified as low affinity subtypes of GS1. As for Glu, GLN1;1 and GLN1;4 again showed substantially high substrate affinities. The K m values of GLN1;1 and a The pairs of forward (RF or F) and reverse (RR or R) primers were used to specifically amplify GLN1 and UBQ2 cDNAs. GLN1;4 for Glu were approximately one-fourth of those calculated for GLN1;2 and GLN1;3. Furthermore, the kinetic constants calculated for the cell culture-derived GS1 fairly matched within the range of K m and V max values of GLN1;2 and GLN1;3 (Table II), suggesting the cells propagated with ample ammonium supply from MS medium may contain the low affinity subtypes as major forms of GS1.
As shown in Fig. 6, the kinetics of GLN1;3 activities over Glu concentrations showed distinctive characteristics (Fig. 6C). The GLN1;3 activities followed a Michaelis-Menten saturation curve when Glu was supplied at the concentrations lower than 20 mM. GLN1;3 exhibited substantially high specific activities within this range, as indicated by a relatively high V max value (162.0 Ϯ 42.0 nanokatal/mg protein) compared with those of the other isoenzymes (Table II). However, the addition of a surplus amount of Glu in the assay mixture significantly inhibited the activity of GLN1;3 (Fig. 6C). The activity was decreased ϳ50% of the calculated V max value when 80 mM of Glu was present in the reaction mixture.
Cell Type-specific Expression of GS1 in Arabidopsis Roots-Arabidopsis plants were transformed with fusion gene constructs of GS1 promoters and GFP. At least five independent lines from each construct were analyzed to see the identity of the expression. GFP, under the control of the GLN1;1 promoter, was abundantly accumulated in the root surface. It was detected in root tips, root hairs, and epidermis (Fig. 7, A and B). By contrast, strong promoter activities of GLN1;2 and GLN1;3 were observed in the pericycle in the vasculature (Fig. 7, C, D,   FIG. 4. Fluctuation of GS1 mRNA abundance during ammonium treatment in Arabidopsis roots. Two weekold Arabidopsis plants were subjected to nitrogen starvation for 3 days prior to the treatment and then transferred to the medium without nitrogen (dashed line) or with 10 mM ammonium chloride (solid line). Total RNA was extracted from the roots at 0, 3, 6, 12, and 24 h after the treatments, and the quantitative realtime PCR analysis was performed using gene specific primers for the GS1 isoenzymes (A-D) and ubiquitin (UBQ2) (E) as a control. The PCR products were detected and quantified as SYBR Green fluorescence. Means of independent triplicate samples and S.D. values (n ϭ 3) are indicated.
FIG. 5. Arabidopsis GS1 isoenzymes purified from the E. coli overexpression systems and cell culture. A, purified GS1 proteins from Arabidopsis cell culture (500 ng of protein) and recombinant GS1 isoenzymes (250 ng protein). GS1 proteins were visualized by the Coomassie Blue dye binding method. B, Western blot of Arabidopsis GS1 isoenzymes using the anti-rice GS1 antibody. C, determination of the octameric structures. Five hundred micrograms of the purified GS1 from cell culture (q), GLN1;1 (X), GLN1;2 (f), GLN1;3 (-) and GLN1;4 (OE) were loaded onto HiLoad 26/60 Superdex 200pg (Amersham Biosciences). Calibration standards (MW-GF-1000 kit; Sigma) are indicated by the arrowheads for thyroglobulin (669 kDa), apoferritin (443 kDa), ␤-amylase (200 kDa), and alcohol dehydrogenase (150 kDa). and F). Fluorescence of GFP was observed in the vasculature from the top to the bottom of root tissues in GLN1;2-GFP plants (Figs. 7, C-E), whereas the signal of GLN1;3-GFP was restricted in the mature part of roots (Fig. 7F) and was not detected in the root tip (Fig. 7H). GLN1;4 was the low abundant form of high affinity isoenzyme expressed under nitrogen limitation (Figs. 3 and 4). The signals of GFP in GLN1;4-GFP plants were observed in the pericycle but were restricted to the basal region of lateral root emergence (Fig. 7I). DISCUSSION Ammonium and nitrate are the predominant nitrogen sources for plants. When plants are grown under different concentrations of ammonium supply, they are capable of optimizing the rate of ammonium influx by regulating ammonium transporters (36 -38). This initial regulatory step of ammonium assimilation is suggested to localize at the surface layers of roots, as indicated by the preferential accumulation of nitrogen-responsive ammonium transporters of tomato in the root hairs (39 -40). Furthermore, it is suggested that Arabidopsis may hire a Gln-mediated negative feedback regulation for the repression of ammonium transporters (37). However, under the condition that overrides the regulatory gate of transport systems, plants are obligated to accept excessive amounts of ammonium. Under such circumstances, the activities of the down-stream metabolic enzymes of nitrogen assimilation must be regulated in parallel. In the present study, we demonstrated that ammonium assimilation in Arabidopsis roots is regulated by finely controlled mechanisms relevant to the enzyme capacities and differential expression of the four distinct isoenzymes of GS1 responding to the nitrogen availabilities. It is suggested that the regulation of GS1 is significant in conjunction with the function of ammonium transporters during acclimation to the fluctuation of nitrogen environments.
Among the GLN1 genes in Arabidopsis, GLN1;2 was the only isoenzyme up-regulated by ammonium supply (Fig. 3), indicating that the GS1 protein accumulated there (Fig. 2, A and B) is mainly composed of GLN1;2 protein. Induction of GLN1;2 mRNA and the corresponding protein accumulation by ammonium is likely interrelated with its enzymatic properties and localization. An excess supply of ammonium can shut down the activities of ammonium transport systems (36 -37) predominantly localized in the root epidermis (39 -40); however the ammonium may overflow this regulatory barrier and reach the inner cell layers when a surplus amount is applied to the root tissue. Under such conditions, GLN1;2 in the vasculature (Fig.  7,C and D) apparently participates in Gln synthesis. The low affinity kinetics of GLN1;2 (Table II)   a Purified GS1 protein was incubated in a 0.1 ml assay mixture as described under "Experimental Procedures." One microgram of GLN1;1, GLN1;2, and cell culture-derived GS1 and 0.5 g of GLN1;3 and GLN1;4 were used for the GS enzyme assays. One katal of enzyme activity was defined as 1 mol of Gln synthesized per second at 30°C. nium allows substantial accumulation of corresponding protein (Fig. 2, A and B); however, the low substrate affinities of this isoenzyme can reduce the rate of Gln synthesis and may render a homeostatic control of total GS activities in roots (Fig. 2C).
GLN1;3 was the other low affinity GS1 isoenzyme expressed in Arabidopsis roots (Table II). The purified GLN1;3 enzyme showed a high capacity of Gln synthetic activity as indicated by a relatively high V max value; however, the activities were significantly inhibited when Glu was supplied at the concentrations higher than 20 mM (Fig. 6C). This enzymatic characteristic of GLN1;3 may lower considerably the total GS activities in roots under the ammonium excess conditions. In addition, GLN1;3 was located in the vasculature (Fig. 7F) and was downregulated by ammonium at the mRNA levels ( Fig 3B). These results suggest that the function of GLN1;3 is strictly controlled both at the levels of mRNA and of enzyme activities by nitrogen availabilities in Arabidopsis roots. In addition to the induction of the low capacity isoenzyme, GLN1;2, the negative regulation of GLN1;3 can substantially reduce the rate of Gln synthesis in roots under ammonium excess.
In contrast to these low affinity enzymes, GLN1;1 showed high affinities both to ammonium and glutamate with relatively low V max values (Table II). The GLN1;1 mRNA was abundantly expressed in nitrogen-starved plants and was significantly repressed by the excess supply of ammonium (Fig. 3). The expression of GLN1;1 was found at the root surface, root tips, root hairs, and epidermal cells (Fig. 7A). These results strongly suggest the importance of GLN1;1 under nitrogenlimited conditions, facilitating a rapid conversion of ammonium at the cell layers that are in contact with the soil solution. During nitrogen limitation, high affinity ammonium transporters are regulated at the mRNA levels; i.e. the increase of AMT1;1 and AMT1;3 mRNA in Arabidopsis is accompanied by induction of the ammonium uptake (36 -37). GLN1;1 accumulated in the root hair and epidermis may operate the Gln synthesis coupled with the action of these high affinity ammonium transporters. GLN1;4 was the other high affinity isoenzyme expressed under nitrogen-limited conditions (Fig. 3A). The GLN1;4 mRNA was less abundant than GLN1;1 and spe-cifically localized at the pericycle cells of the basal region of lateral root emergence (Fig. 7I). These cells could have a direct contact to the soil solution due to the breakage of outer cell layers by the emergence of lateral roots. GLN1;4 may partly contribute to the Gln synthesis in these specific cell types when ammonium is undersupplied.
Taken together, both the ammonium influx and Gln synthesis are controlled by negative feedback regulation in Arabidopsis roots. The ammonium uptake was associated with the expression of high affinity ammonium transporters, induced by nitrogen limitation, and suppressed by the nitrogen supply and Gln (36 -37). Unlike the case in rice (41), the induction of ammonium transporter by ammonium supply was absent in Arabidopsis (36). The present study indicated that the high affinity isoenzyme of GS1 in the root epidermis (GLN1;1) is down-regulated by ammonium; however the expression of GLN1;2 that exhibits the lowest affinity is conversely stimulated in Arabidopsis roots. In addition, the other low affinity isoenzyme, GLN1;3, which shows the highest capacity of Gln synthesis, was down-regulated by ammonium excess at the mRNA levels, and, strikingly, its enzyme activity was negatively controlled by Glu. Consequently, the overall Gln synthetic activity in Arabidopsis root is steadily controlled during ammonium treatment, which does not meet with the marked increase in GLN1;2 mRNA and the corresponding GS1 protein contents. These results suggested a unique regulatory system of ammonium assimilation in Arabidopsis, totally different from those that exist in the legumes and monocots (18 -22, 41).