The PII Signal Transduction Protein of Arabidopsis thaliana Forms an Arginine-regulated Complex with Plastid N-Acetyl Glutamate Kinase*

The PII proteins are key mediators of the cellular response to carbon and nitrogen status and are found in all domains of life. In eukaryotes, PII has only been identified in red algae and plants, and in these organisms, PII localizes to the plastid. PII proteins perform their role by assessing cellular carbon, nitrogen, and energy status and conferring this information to other proteins through proteinprotein interaction. We have used affinity chromatography and mass spectrometry to identify the PII-binding proteins of Arabidopsis thaliana. The major PII-interacting protein is the chloroplast-localized enzyme N-acetyl glutamate kinase, which catalyzes the key regulatory step in the pathway to arginine biosynthesis. The interaction of PII with N-acetyl glutamate kinase was confirmed through pull-down, gel filtration, and isothermal titration calorimetry experiments, and binding was shown to be enhanced in the presence of the downstream product, arginine. Enzyme kinetic analysis showed that PII increases N-acetyl glutamate kinase activity slightly, but the primary function of binding is to relieve inhibition of enzyme activity by the pathway product, arginine. Knowing the identity of PII-binding proteins across a spectrum of photosynthetic and non-photosynthetic organisms provides a framework for a more complete understanding of the function of this highly conserved signaling protein.

In prokaryotic organisms, the PII protein is recognized as the key mediator of energy, carbon, and nitrogen interactions and is referred to as the central processing unit of carbon:nitrogen metabolism (1)(2)(3)(4). Escherichia coli PII is a 112-amino acid protein that as a homotrimer senses the cellular status of both ATP and the carbon skeleton 2-oxoglutarate (2KG) 3 via allosteric means. Nitrogen status is assessed through glutamine levels by covalent modification (uridylylation) of PII. This metabolic information is signaled to other proteins by proteinprotein interaction and produces an appropriate response that alters gene expression and the activity of glutamine synthetase (3,5). In terms of metabolic sensing, cyanobacterial PII plays a similar role, but in this case, covalent modification is by phosphorylation (6). To date, the processes known to be regulated by PII in cyanobacteria are: ammoniumdependent nitrate/nitrite uptake (7), high affinity bicarbonate transport (8), regulation of the global transcriptional activation by NtcA (9,10), and arginine biosynthesis (11).
In eukaryotes, PII has only been identified in plants and red algae (12), and its sequence is highly conserved when compared with prokaryotic PIIs, with Arabidopsis thaliana PII being 50 and 55% identical to E. coli and Synechococcus elongatus PII, respectively. Plant PII proteins have a conserved N-terminal extension that functions as a chloroplast transit peptide, which is consistent with biochemical data indicating that PII resides in this compartment. We have previously shown that the plant PII protein is not regulated by phosphorylation (13). Like the bacterial protein, plant PII binds 2-oxoglutarate, but only after binding ATP first, and thus likely functions to sense plastid energy and 2KG status (14). PII transcripts appear to be present in all plant organs, and PII protein levels do not change during day/night cycles or N-nutrition (14,15). Although PII T-DNA knock-out lines show increased sensitivity to nitrite and slight alternations in carbon metabolite and amino acid levels during altered N-nutrition (15), no molecular targets have been firmly established for plant PII. Two preliminary yeast two-hybrid studies have indicated that N-acetyl glutamate kinase (NAGK) is a PII interactor in plants (16,17). We searched for PII-interacting proteins by performing affinity chromatography with plant PII and have identified the major PII receptor of A. thaliana as the plastid enzyme N-acetyl glutamate kinase. Biochemical studies showed that PII alters the kinetic properties of NAGK and that the downstream end product of this metabolic pathway (arginine) promotes the interaction of PII and NAGK. Interestingly, arginine levels are barely detectable in plants during the light period and are high during the dark (18), suggesting arginine inhibition of NAGK during the light and relief by PII in the dark.

EXPERIMENTAL PROCEDURES
Affinity Chromatography-A. thaliana PII minus the chloroplast transit peptide was expressed and purified as described (19). After dialysis into PBS, 2 mg of PII and 2 mg of Fraction V BSA (Sigma) were separately coupled to 1 ml of CH-Sepharose (Amersham Biosciences) by following the manufacturer's instructions (coupling was Ͼ95%). A. thaliana suspension cells were grown in culture, harvested, and lysed as described in Smith et al. (13). Typically, ϳ300 g of cells were lysed in a French press with 1 volume of 25 mM Tris/Cl, pH 7.5, 1 mM EGTA, 1 mM EDTA, 1 mM benzamidine, 0.1 mM PMSF, 0.1% (v/v) 2-mercaptoethanol, and 1% (w/v) polyvinylpyrrolidone and clarified by centrifugation at 35,000 rpm for 30 min in a Ti-45 rotor. Following filtration through Miracloth (Calbiochem), the crude extract (ϳ1900 mg of protein) was then split in half and mixed end-over-end at 4°C with 1 ml of PII or BSA affinity matrix that was previously equilibrated in buffer A (25 mM Tris/Cl, pH 7.5, 1 mM benzamidine and 0.1 mM PMSF). After 4 h, the mixture was poured into a column and washed with 300 ml of buffer A plus 150 mM NaCl, and protein was eluted with 8 ml of buffer A plus 1.5 M NaCl. Eluted protein was concentrated in a Centriprep 10 (Amicon) and then in a Centricon 10 to 25 l and boiled in SDS-mixture. In all cases, PII-and BSA-Sepharose chromatographies were performed identically to allow direct comparison.
Mass Spectrometry and Edman Sequencing-PII-and BSA-binding proteins were run on SDS-PAGE and Coomassie Blue-stained, and bands were excised, trypsin was digested, and proteins were identified by MALDI-TOF mass spectrometry (20). The only proteins indicated (see Fig. 1A) are those for which a high level of confidence for identification was obtained (see supplementary Table 1). In all cases, the identified protein was the best match to the peptide mass data, the identified protein predicted mass was very near the observed mass on the gel, and the percentage of peptide coverage was high, ranging from 11 to 62% of the total number of residues using a cut-off of less than 35 ppm. For N-terminal sequencing, the PII-Sepharose-eluted proteins were run on SDS-PAGE, blotted to PVM, stained with Amido Black, and washed with water to visualize the 33-kDa protein. This band was excised and submitted to Edman chemistry as described (21).
Cloning, Expression, and Purification of A. thaliana N-Acetyl Glutamate Kinase and Non-tagged PII-The N-acetyl glutamate kinase gene was PCR-amplified (to start from the predicted chloroplast cleavage site Ala 50 ) from an A. thaliana silique cDNA library template obtained from the A. thaliana Biological Resource Centre, Ohio State University, using the following primers: 5Ј-TATAAGATCTCAGCCACCGTATCAA-CACCAC-3Ј and 3Ј-TATAGAATTCTTATCCAGTAATCATAGT-TCCAGC-5Ј. After digestion with EcoRI and BglII, the PCR product was cloned into EcoRI-BglII-restricted pRSET A plasmid (Invitrogen) to produce a fusion construct that coded for NAGK and a 6-histidine tag on the N terminus with a predicted mass of ϳ36 kDa. This construct was transformed into E. coli DH5␣ cells, and positive transformants were sequenced at the University of Calgary DNA sequencing facility and confirmed to have identical amino acid sequence as A. thaliana NAGK. The resulting plasmid was used for the overexpression of His 6 -NAGK fusion protein in E. coli Rosetta gami pLysRARE cells (Novagen). The transformed bacterial cells were grown at 37°C for 26 h. The cells from 1 liter of culture were harvested by centrifugation and resuspended in 50 ml of bacterial lysis buffer (25 mM Tris/Cl, pH 7.5, 150 mM NaCl, 10 mM imidazole, pH 7.5, 0.5 mM PMSF, 0.5 mM benzamidine, 5 g/ml leupeptin). Cells were lysed by two passes through a French pressure cell at 1000 p.s.i., and cell debris was removed by centrifugation at 35,000 rpm for 45 min at 4°C in a Ti-45 rotor. The soluble recombinant NAGK protein was bound to a 10-ml Ni-NTA agarose (Qiagen) column pre-equilibrated with bacterial lysis buffer. The matrix was washed with 25 mM Tris/Cl, pH 7.5, 1 M NaCl, 30 mM imidazole, pH 7.5, 0.1% (v/v) Tween 20, 0.5 mM PMSF, and 0.5 mM benzamidine. The His 6 -NAGK protein was eluted with 50 mM Tris/Cl, pH 7.5, 150 mM NaCl, 300 mM imidazole, pH 7.5, 0.5 mM PMSF, and 0.5 mM benzamidine. Peak fractions were pooled and dialyzed against PBS plus 50% (v/v) glycerol. The purified expressed protein was confirmed to be NAGK by MALDI-TOF mass spectrometry as described above. One liter of culture produced ϳ45 mg of pure NAGK.
A non-tagged PII protein was made by PCR amplification from the plasmid described in Ref. 19, cloned into the pET3a vector, and after sequence verification, used to transform E. coli Rosetta gami pLys cells. Protein was expressed by growing cells in Terrific broth at 25°C for 16 h in 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside. After harvesting cells and French press lysis as above, the protein was purified by sequential chromatography on Macro-Prep High S matrix (Bio-Rad) and phenyl-Sepharose (Amersham Biosciences). One liter of cells yielded 10 mg of pure PII.
Reverse Affinity Chromatography-The A. thaliana PII cloned without a tag or transit peptide (32 g) was mixed with either His 6 -NAGK (65 g) or His 6 -PR65␣(65 g) for 10 min in a total volume of 100 l with 25 mM Tris/Cl, pH 7.5, 150 mM NaCl, 5% (v/v) glycerol and then mixed with 25 l of Ni-NTA-agarose end-over-end for 1 h at 4°C. After pelleting, beads were washed two times with 500 l of PBS, two times with 25 mM Tris/Cl, pH 7.5, 0.05% (v/v) Nonidet P-40, and 20 mM imidazole, and two times again with PBS, and then proteins were released from the matrix by the addition of 0.3 M imidazole, pH 7.5. The following metabolites were included in the incubation mixture to test for their ability to disrupt or enhance the PII-NAGK interaction: Glu, Gln, Asp, ATP, ADP, carbamoyl phosphate, citrulline, and Lys (all at 5 mM), 2KG, AMP, NaNO 2 (all at 1 mM), NH 4 Cl (20 mM), 2KG ϩ ATP (1 and 5 mM, respectively), MgCl 2 (10 mM), MgCl 2 ϩ ATP (10 and 5 mM, respectively), MgCl 2 ϩ ATP (10 and 5 mM, respectively), N-acetyl glutamate (NAG) (20 mM), and Arg (1, 5, 10, and 20 mM). In one series of experiments in which 5 mM arginine was included in the incubation buffer, the same concentration of arginine was included in all wash steps.
Antibody Production and Localization of NAGK to Chloroplasts-Recombinant NAGK purified by Ni-NTA-agarose (see Fig. 1B) was used to raise antibodies in a rabbit using standard procedures (19). The antibodies were used in Western blots and immunofluorescence experiments as 5000-and 200-fold diluted crude serum, respectively. For immunological staining of fixed cells, A. thaliana suspension cells were grown as described previously (13), harvested, and allowed to settle by gravity. Once settled, the supernatant was removed, and the cells were fixed by adding a freshly prepared 4% (w/v) formaldehyde solution made in PBS. Cells were fixed for 15 min before being washed with PBS (3ϫ 1 ml). The cell wall was partially digested with 500 l of a 0.1% (w/v) pectolyase Y-23 (Seishen Pharmaceutical Co.) solution made in PBS for 15 min at 30°C. The cell membrane was permeabilized by incubation with 1% (v/v) Triton X-100 for 5 min at room temperature. The cells were then washed with PBS (3ϫ 1 ml), blocked in 2% (w/v) BSA in PBS for 10 min, and then incubated overnight at 4°C in 200-fold diluted crude serum specific for NAGK or PII (14). The following day, the cells were washed with PBS (3ϫ 1 ml) and incubated with anti-mouse secondary antibody conjugated to the fluorophore Alexa Fluor 488 (Molecular Probes) and in blocking buffer for 1 h at room temperature. The samples were washed again in PBS (3ϫ 1 ml) before being analyzed with a fluorescence microscope (Leica DMR) using the fluorescein isothiocyanate filter set. For the observation of chloroplasts, chlorophyll fluorescence was achieved using the UV filter set. Images were captured using a cooled CCD camera (Retiga 1350 EX, Qimaging), and image enhancement and deconvolution confocal algorithm manipulations were performed using the Openlab software package (Version 3.0, Improvision). Pseudocoloration and image manipulation was performed in PhotoShop.
Isothermal Titration Calorimetry-ITC experiments were performed as described in Smith et al. (14) using His 6 -NAGK and His 6 -PII without transit peptides, in the following: 25 mM Tris/Cl, pH 7.5, 5% (v/v) glycerol, and 150 mM NaCl. Each run was corrected for heat of dilution.
Enzyme Kinetics-For enzyme kinetics, all assays were done in duplicate in three or more separate experiments, except in the assays where K m and V max values were examined for arginine-inhibited NAGK, plus and minus PII. Here, each assay was done in triplicate. NAGK activity assays with purified proteins were determined as described by Heinrich et al. (11). Unless stated otherwise, the reaction mixture consisted of 300 mM NH 2 OH⅐HCl, pH 7.5, 20 mM Tris/Cl, pH 7.5, 20 mM MgCl 2 , 40 mM NAG, and 10 mM ATP and NAGK diluted into 25 mM Tris/Cl, pH 7.5, and 150 mM NaCl such that each assay contained 0.81 g of enzyme.
Reactions were initiated by the addition of reaction mixture, and the incubation was carried out at 37°C in a volume of 0.10 ml and terminated after 20 min by the addition of 0.1 ml of stop mixture (1:1:1 of 5% (w/v) FeCl 3 ⅐ 6 H 2 O in 0.2 M HCl, 8% (w/v) trichloroacetic acid, and 0.3 M HCl), and after standing for 5 min at room temperature, the tubes were centrifuged for 1 min at 14,000 rpm. Under these reaction conditions, the assay was linear up to at least 30 min. Blank reactions were performed by omitting N-acetyl glutamate from the assay. One unit of enzyme activity is defined as one mmol of product produced in 1 min calculated with a molar absorption coefficient of 456 M Ϫ1 cm Ϫ1 at 540 nm for the N-acetylglutamylhydroxamate-Fe 3ϩ complex. The non-linear regression program GraphPadPrism 4.0 (GraphPad Software, San Diego, CA) was used for estimating the K m and V max values for all data except the data obtained from arginine inhibition. The catalytic efficiency and k cat of NAGK were calculated using K cat /K m and V max / [NAGK] total , respectively. The V max and K m values for NAGK were determined by holding the concentration of the second substrate at a saturating level (NAG at 40 mM or ATP 10 mM (MgCl 2 at 20 mM)). The varied substrate was omitted when carrying out blank assays.
To study the effect of PII on the activity of NAGK, various amounts of A. thaliana PII (14) were added into the diluted NAGK fraction, and the mixture was placed at 4°C for 10 min to allow complex formation. The reaction was initiated by the addition of the reaction mixture into the NAGK-PII complex mixture. For determining the effect of PII on the K m and V max for NAG and ATP, PII was added into the diluted NAGK such that the molar ratio of PII-polypeptide to NAGK-polypeptide was 2:1 (based on a Bradford assay with BSA as standard) and allowed to interact at 4°C for 10 min. For studying the NAGK activity in the presence of PII and its effector 2KG, PII was incubated with 5 mM ATP and 1 mM 2KG for 3 min at 4°C followed by the addition of NAGK. For studying the K m and V max of NAGK for the substrate ATP, PII was preincubated with 0.05 mM ATP and 1 mM 2KG.
The effect of arginine on NAGK was studied under saturating substrate conditions (10 mM ATP and 40 mM NAG). The effect of arginine was also studied at K m conditions in which the concentration of NAG was set to 40 mM and the concentration of ATP was set to the K m value (1.74 mM). For either condition, various amounts of arginine were added into the NAGK fraction, and the mixture was allowed to stand for 5 min were performed for each set of data. The V max values were estimated from Eadie-Hofstee plots, and the K m values were estimated from plots of velocity versus substrate concentration. To test the effect of arginine on the PII-NAGK complex, PII was first added into the diluted NAGK solution, and the two proteins were placed at 4°C for 10 min to allow complex formation followed by the addition of arginine. The enzyme mixture was allowed to stand for 5 min before the initiation of assays.

Identification of A. thaliana PII-binding
Proteins-To identify PIIbinding proteins, an extract from A. thaliana was incubated with PII coupled to Sepharose beads. After extensive washing, bound proteins were eluted with 1.5 M NaCl and run on SDS-PAGE (Fig. 1A). No proteins were eluted from the control matrix (BSA-Sepharose), whereas several proteins were consistently enriched on the PII matrix. The eluted proteins were identified by MALDI-TOF mass spectrometry and are indicated on the gel (Fig. 1A, and see also Supplemental Table 1). The major PII-binding protein of 33 kDa was identified as a putative plastid N-acetyl glutamate kinase (gi 15230338). Other minor bound proteins were the cytosolic ␣, ␤, ␤Ј, and ␥ subunits of the coatomer protein complex I vesicle coat protein complex and several of the smaller cytosolic 60 S ribosomal proteins. ChloroP analysis (ChloroP 1.1 Server) of the NAGK sequence predicts that this gene product is chloroplast-localized, and cleavage of the chloroplast transit peptide is likely between Lys 49 and Ala 50 , yielding a protein of 31.1 kDa. Edman sequencing of the PII-purified NAGK yielded the sequence 51 TVSTPPSI 58 . This confirmed that mature NAGK has its putative chloroplast transit peptide cleaved, but between Ala 50 and Thr-51 , and thus likely localizes to the chloroplast. This is consistent with the expected localization of a PII-interacting protein. Being the most abundant PII-binding protein and likely residing in the same compartment, we further characterized the interaction of PII and NAGK. A. thaliana NAGK was cloned minus the chloroplast transit peptide, expressed in E. coli as a 6-histidine fusion protein, and purified (Fig. 1B).
NAGK and PII Are Chloroplast-localized Proteins-The chloroplast localization of NAGK and PII was confirmed by immunolocalization experiments using antibodies raised against each protein. The NAGK antibody was first tested by probing Western blots of pure enzyme and an A. thaliana crude fraction. The diluted crude serum could easily detect less than 1 ng of recombinant NAGK and one major band of 33 kDa in a crude extract (data not shown). Both PII and NAGK stained in small structures that resemble chloroplasts and were subsequently colocalized with chlorophyll fluorescence (Fig. 2). The immunolocalization of PII and NAGK to the chloroplast is consistent with previous biochemical studies of A. thaliana PII (22) and the predicted chloroplast transit peptides of both PII and NAGK.
Reverse Affinity and Gel Filtration Chromatography Confirms an Interaction between PII and NAGK-Knowing that PII and NAGK localize to the same compartment supports the idea that they are true interacting partners. To further confirm an interaction between PII and NAGK, we performed the reverse experiment in which His-tagged NAGK was incubated with non-tagged PII and the complex was pulled out with Ni-NTA-beads. Fig. 3A shows that NAGK binds PII and that PII in the presence of an unrelated His-tagged protein (PR65) is not retained on the beads, nor is non-tagged PII retained on the beads alone (data not shown).
The mass of the A. thaliana purified NAGK was determined by gel filtration chromatography after elution from the PII matrix. Peak fractions demonstrated a mass of ϳ215 kDa (Fig. 3B). This 33-kDa band was confirmed to be NAGK by Western blot analysis (data not shown) and suggests that the native protein exists as a multimer of six or seven 33-kDa subunits. Fig. 3, C-F , show the elution profiles for recombinant PII alone, NAGK alone, and a mixture of PII plus excess NAGK. Tagged PII displayed a mass of ϳ45 Ϯ 0.5 kDa (14), which is consistent with a mass of a trimer of 16-kDa subunits, whereas recombinant NAGK eluted at ϳ276 Ϯ 2 kDa, which is consistent with a multimer of six or seven 33-kDa subunits (the recombinant protein contains a tag of ϳ5 kDa in addition to the native protein). If the recombinant protein exists as a hexamer, subtracting six tags gives a mass similar to the native protein, giving us confidence that the bacterial produced protein is behaving like the purified plastid enzyme. When the two proteins are mixed (Fig. 3, E and F), the PII elution profile shifts dramatically, suggesting that PII and NAGK form a complex that displays a mass of ϳ300 Ϯ 1.5 kDa. PII and NAGK chromatographed in the presence of 0.4 mM MgCl 2 and 0.2 mM ADP displayed the same mass (data not shown).
Isothermal Titration Calorimetry-ITC allowed us to further address the stoichiometry of binding for PII and NAGK. Titration of PII into a solution of NAGK (Fig. 4) gives an n value of 0.43, suggesting that each PII trimer binds ϳ7 NAGK subunits. In the reverse experiment, in which NAGK was titrated into a solution of PII, n ϭ 1.77, suggesting that about five NAGK molecules interact with each PII trimer. In combination with the gel filtration data, this supports the idea that one PII trimer interacts with one NAGK hexamer and that native NAGK is indeed a hexamer. Fig. 1). The reaction catalyzed by NAGK involves two substrates, ATP and NAG, in which NAGK transfers the ␥-phosphate group from ATP to NAG to form N-acetyl-␥-glutamyl phosphate. For free NAGK (not complexed with PII), the apparent V max and K m obtained when NAG was the variable substrate As shown in Fig. 5B, the V max of NAGK increased by a maximum of 30% when it was in complex with PII. When NAGK forms a complex with PII, the V max for the NAG substrate increased to 13.2 (Ϯ0.3 S.E.) units/mg, and its K m increased slightly to 7.55 (Ϯ0.56 S.E.) mM. The resulting catalytic efficiency of 1020 s Ϫ1 M Ϫ1 corresponds to a modest increase in catalytic efficiency for PII-complexed NAGK, in comparison with free NAGK. When NAGK forms a complex with PII, the V max and K m for ATP both increased to 11.6 (Ϯ0.25 S.E.) units/mg and to 2.03 (Ϯ0.15 S.E.) mM, respectively, giving an overall catalytic efficiency of 3320 s Ϫ1 M Ϫ1 that was indifferent from the one obtained for free NAGK.

Altered Kinetic Behavior of NAGK in the Presence of PII and Its Effector Molecule 2KG-NAGK catalyzes the second step in the arginine biosynthetic pathway (see Supplemental
2-Ketoglutarate is an effector molecule that binds to PII and signals carbon status in the chloroplast (2,14). It binds PII only when the ATP binding site in PII has ATP bound (14). In the presence of PII-ATP-2KG complex, the V max of NAGK increased ϳ2-fold (  the presence of PII and 2KG in comparison with PII alone. PII with 2KG bound increased the catalytic efficiency of NAGK by 38% overall. Feedback Inhibition of NAGK-NAGK catalyzes the committed step in the biosynthesis of arginine, and in cyanobacteria, it has been found to be feedback-regulated by the end product of this biosynthetic pathway (arginine) (23). Under V max conditions, free A. thaliana NAGK exhibited feedback inhibition by arginine with a half-maximum inhibition concentration (IC 50 ) of 0.73 mM, and in the presence of PII, less inhibition of NAGK was observed, with the IC 50 value increasing ϳ3.4-fold to 2.5 mM (Fig. 6A).
To investigate arginine inhibition of NAGK under K m conditions, the concentration of ATP was set at its K m value (1.74 mM), whereas the concentration of NAG was set at a saturating level. The IC 50 for arginine obtained for free NAGK was 0.33 mM, and similar to arginine inhibition at V max conditions, PII relieved the inhibition of NAGK activity by argi-nine, yielding an IC 50 of 0.83 mM (Fig. 6B). The effect of 2KG on the feedback inhibition of NAGK by arginine was studied only at V max conditions, and the resulting inhibition curve gave a half-maximal inhibition value of 2.4 mM, which was indifferent from the IC 50 obtained for PII-complexed NAGK in the absence of 2KG (data not shown).
Next we chose to study how the K m and V max values change for NAGK for three different concentrations of arginine (0.16, 0.32, and 0.64 mM). This was accomplished by examining the V max and K m of NAGK where NAG was the variable substrate and the concentration of ATP was held fixed at a saturating level. For both free NAGK and PIIcomplexed NAGK, V max decreased and K m increased as the concentration of arginine increased (Fig. 7A). Overall, the catalytic efficiency of PII-complexed NAGK was higher than free NAGK, and the difference between the two increased as arginine concentration increased, reaching to a 2.3-fold activation of NAGK by PII at 0.64 mM arginine (435 versus 186 s Ϫ1 M Ϫ1 ). Another interesting feature observed in this experiment was the change in the shape of the plot for free and PII-complexed NAGK when arginine was present. The velocity versus substrate concentration plot for free NAGK inhibited by 0.32 mM arginine showed a sigmoidal curve with a Hill coefficient of 2.3, and a similar plot for NAGK in complex with PII showed a Michaelis-Menten curve (Fig. 7B). The corresponding Hill plot of the data for PII-complexed NAGK gave a Hill coefficient of 1. In general, the Hill coefficient increased as the amount of arginine increased in the sample. For free NAGK, the Hill slope coefficient increased to ϳ3 when the concentration of arginine was 0.64 mM.
The following compounds were tested for their ability to inhibit or activate NAGK and the NAGK-PII complex and shown to have little or no effect on activity. The concentration tested in the assay was 5 mM for Glu, Gln, Asp, Asn, Ser, Thr, His, Lys, Gly, Pro, and 5Ј-AMP or the  1-4) or His 6 -PR65␣ and PII (lanes 5-8). Lanes 1 and 5, the supernatant after incubation with the beads (any protein that did not bind). Lanes 2 and 6, supernatant after the initial wash with column buffer. Lanes 3 and 7, supernatant after the final wash with column buffer. Lanes 4 and 8, supernatant after eluting from the Ni-NTA beads for 10 min with elution buffer (300 mM imidazole). Molecular mass standards are indicated. B, Superdex 200 gel filtration chromatography of NAGK affinity-purified from an A. thaliana extract (100-l fractions). A silver-stained 12% SDS-PAGE analysis of 30 l of each fraction (Fr. 10 -16) shows the major protein of ϳ33 kDa. Panels C-F show complex formation between PII and NAGK. Protein samples were injected onto the Superdex 200 gel filtration column, and 100-l fractions were collected. Fractions 10 -19 were loaded onto 14% SDS-PAGE gels and Coomassie Blue-stained. In C, 12 g of His 6 -PII alone was injected, and 35 l of each fraction was run on the gel. In D, 74 g of His 6 -NAGK alone was injected, and 30 l of each fraction run on the gel. In E, 5 g of His 6 -PII and 135 g of His 6 -NAGK were preincubated together before injection, and 25 l of each fraction were run on the gel. In F, 10 g of PII and 135 g of His 6 -NAGK were preincubated together before injection, and 25 l of each fraction were run on the gel. NAGK and PII Form a Complex That Is Modulated by Arginine-We next tested the ability of several key metabolites to modulate the interaction of PII with NAGK using the pull-down assay described in the legend for Fig. 2. This screen of carbon, nitrogen, and energy status metabolites (see "Experimental Procedures") revealed that only arginine had any effect, promoting binding at this concentration (5 mM) (Fig. 8). Higher arginine concentrations did not increase binding, and lower concentrations produced lower but significantly increased binding (data not shown). Including arginine in the wash steps did not increase binding, indicating that during the incubation, arginine promotes a conformation that allows tighter binding, which is maintained in the wash steps in the absence of arginine.

DISCUSSION
The chloroplast of plants functions not only as the site of photosynthesis and carbon fixation but also as the location of many metabolic biosynthetic pathways, including the assimilation of NO 2 Ϫ into an organic form and the synthesis of nucleotides and amino acids. As such, the chloroplast is the primary site for the interface of carbon and nitrogen metabolism and is also the compartment where plant PII resides.
We have previously shown that like prokaryotic PII proteins, plant PII senses the carbon skeleton 2KG in a manner dependent upon first binding (or sensing) ATP. After sensing metabolic status, PII provides this information to other proteins to allow them to adjust their function or output accordingly. Although PII has been suggested to be critical to carbon/nitrogen sensing and signaling in plants (12,22), our understanding of how it performs this role is still very limited. This work was initiated to define the binding partners of A. thaliana PII as to date, plant PII partners have only been suggested through a yeast two-hybrid screen (16,17). We used an affinity chromatography-proteomics approach to look for PII-interacting proteins and identified NAGK as the primary PII-binding protein of A. thaliana cells, consistent with the yeast two-hybrid result. NAGK has been proposed to reside in the plastid where PII is localized. Sequence analysis tools suggest that NAGK has a chloroplast transit peptide, and we have confirmed by Edman sequencing that the mature PII-associated NAGK is cleaved between Ala 50 and Thr 51 , consistent with cleavage by the transit peptide protease. To further support this observation, we employed immunolocalization to place both PII and NAGK in the chloroplast of A. thaliana cells. Previously, Sugiyama et al. (17) showed that GFP-NAGK was chloroplast-localized. Knowing that the putative interactors reside in the same compartment, we further explored the interaction using pull-down, gel filtration, and ITC experiments. All of these approaches confirmed that PII and NAGK form a complex. In addition, the gel filtration and ITC data support the idea that native NAGK, as well as the recombinant enzyme, functions as a hexamer like the enzyme from Thermotoga maritima and Pseudomomas aeruginosa (24). Recently, S. elongatus PII was shown to interact with the hexameric NAGK of that species (23). One of our goals was to define the function of PII interaction with NAGK. Our enzyme kinetic analysis demonstrated that PII modestly activated the plant enzyme. This is in contrast to the S. elongatus NAGK, where the V max increases 4-fold by binding PII and the K m for NAG displays a 10-fold decrease (23). It is notable that the A. thaliana enzyme alone has a V max near that displayed for PII-activated S. elongatus NAGK, and therefore, we are not surprised that binding PII only increased activity slightly. We found that in the plant system, the primary function of PII binding to NAGK is to relieve feedback inhibition by the downstream pathway product, arginine, shifting the substrate saturation curve to hyperbolic from sigmoidal in the presence of arginine. It is interesting to note that unlike the dimeric E. coli NAGK, which is not regulated by arginine, all hexameric NAGKs appear to be  The activity is plotted as the percentage relative to the activity without arginine present for free (f) NAGK and PII-complexed (⌬) NAGK. A, arginine inhibition of NAGK under V max conditions at which both ATP and NAG were set at saturating levels. In B, arginine inhibition done under K m conditions in which the concentration of ATP was set at K m value. The sigmoidal dose response curve generated in GraphPad Prism 4.0 was used to estimate the IC 50 for each curve, which is indicated. Each value represents a mean of three replicas, and error bars represent standard errors. IC 50 values are indicated for each curve.
inhibited by this amino acid. We performed a wide search of other possible regulators of enzyme activity and were not able to find any (see "Experimental Procedures"). It may be that other, as yet unidentified conditions, reveal further effects on NAGK activity. We then screened a large pool of metabolites for their ability to enhance or disrupt the PII-NAGK complex. Consistent with our kinetic studies, only arginine had an effect and functioned to enhance the interaction of PII with NAGK. Again this differs slightly from the S. elongatus PII-NAGK interaction as in this system, it appears that binding is disrupted by the adenylates ADP or ATP alone but not Mg-ATP. The Mg-ATP effect is again reversed if 2KG is included with Mg-ATP. Maheswaran et al. (23) concluded that this makes an additional tie between PII function and energy status in this organism. We performed similar experiments and were not able to show any effect for these metabolites, suggesting that the interaction of PII with NAGK is an ancient molecular interaction, but during the evolution of oxygenic photosynthetic organisms, the regulatory events surrounding the PII-NAGK relationship have evolved to fit the specific needs of each organism. This is further highlighted when we look at the role of covalent modification in PII function. In E. coli, tyrosine 51 is uridylylated in response to nitrogen status, whereas in S. elongatus, a model is emerging in which the phosphorylation status of PII on serine 49 is in response to nitrogen/carbon status, and this controls the interaction of PII with NAGK. Changing this serine in PII to another amino acid also weakens or disrupts interaction with NAGK, suggesting a key role for this serine in direct binding (11,16). The residue equivalent to serine 49 of S. elongatus PII is highly conserved in plant and other cyanobacterial PII proteins. We have performed an extensive analysis of plant PII for phosphorylation and have concluded that it is not regulated by phosphorylation (13), and this appears to be true for some cyanobacterial PII proteins as well (25,26).
Why would such an ancient molecular interaction exist? It is interesting to note that in plants, the level of arginine is nearly undetectable during the light period (18). This implies that arginine is primarily synthesized during the dark, under the control of PII, and thus may function as a signaling molecule in plants. This type of event is not unprecedented as it has been shown that the amino acid leucine can control the signaling events downstream of the target of rapamycin protein kinase (27), and this signaling cascade is key to the response to energy and amino acid status (28,29). At this point, we must note that the plant PII protein levels are constant during light/dark cycles and various N-regimes (15) and thus could function in a rapid respond to changing cellular conditions as necessary. As discussed above, the phosphorylation state of PII does not appear to control the binding of PII to NAGK in plants (unlike S. elongatus), and the only metabolite found to control the interaction was arginine itself, which enhanced binding and therefore should increase its own synthesis. In a system in which arginine appears to promote its own synthesis, it is as yet unclear what the molecular switch is that turns the pathway on and off. Perhaps NAGK is regulated by covalent modification and the interaction of NAGK with PII may be regulated by this event. Nevertheless, plant PII plays a regulatory role in a pathway that produces the amino acid arginine, and importantly, flux through this pathway results in the consumption of the key amino acids glutamate, aspartate, and the carbon/nitrogen donor molecule carbamoyl phosphate, thus further establishing the link to nitrogen metabolism. Carbamoyl phosphate is also used in the biosynthesis of the pyrimidines, and in plants, this pathway is localized to the plastid compartment (30).
In conclusion, we have characterized the first eukaryotic PII-binding protein interaction after identification using proteomics. Identification of NAGK as the primary PII target in plants makes this the missing and critical link of plant PII to nitrogen metabolism. With plant PII having the same primary target as photosynthetic bacteria, this further establishes the importance of this ancient signaling cascade.