C. MacKintosh, MRC Protein Phosphorylation Unit, Department of Biochemistry, University of Dundee, MSI/WTB Complex, Dow Street, Dundee DD1 5EH, UK. Fax: + 44 1382 223778, E-mail: email@example.com
An activity that inhibited both glutamine synthetase (GS) and nitrate reductase (NR) was highly purified from cauliflower (Brassica oleracea var. botrytis) extracts. The final preparation contained an acyl-CoA oxidase and a second protein of the plant nucleotide pyrophosphatase family. This preparation hydrolysed NADH, ATP and FAD to generate AMP and was inhibited by fluoride, Cu2+, Zn2+ and Ni2+. The purified fraction had no effect on the activity of NR when reduced methylviologen was used as electron donor instead of NADH; and inhibited the oxidation of NADH by both spinach NR and an Escherichia coli extract in a time-dependent manner. The apparent inhibition of GS and NR and the ability of ATP and AMP to relieve the inhibition of NR can therefore be explained by hydrolysis of nucleotide substrates by the nucleotide pyrophosphatase. We have no evidence that the nucleotide pyrophosphatase is a specific physiological regulator of NR and GS, but suggest that nucleotide pyrophosphatase activity may underlie some confusion in the literature about the effects of nucleotides and protein factors on NR and GS in vitro.
In response to water stress or when photosynthesis is blocked, the cytosolic enzyme nitrate reductase (NR) is inhibited by a two-step mechanism; firstly, a serine residue (Ser543 in the spinach enzyme) is phosphorylated [1,2]. The phosphorylation alone has no effect on enzyme activity. The addition of a phosphate to this serine, within this amino acid context, generates a phosphopeptide motif (Arg-Ser-X-phosphoSer- X-Pro) that is recognized by and binds to NIP (nitrate reductase inhibitor protein) [3,4], which comprises isoforms of 14-3-3 proteins [5,6]. Binding of 14-3-3 proteins inhibits the phosphorylated NR. The inhibition of phosphorylated NR by 14-3-3 proteins also requires millimolar Mg2+ or Ca2+. The inhibited, 14-3-3-bound NR can be activated by dephosphorylation [4,5], dissociation of 14-3-3 proteins by a competitor 14-3-3-binding phosphopeptide , or chelation of metal ions .
During purification of 14-3-3 proteins, a second protein factor was found to ‘interfere’ with the inactivation of NR by phosphorylation and 14-3-3 proteins . The inhibition of NR by the ‘interfering’ protein was blocked by ATP. This means that the inhibition of NR by Mg-ATP, NR kinase, and 14-3-3 proteins can be counteracted by what seems like Mg-ATP-dependent activation of NR in fractions containing the ‘interfering’ protein (see Results). The effect of Mg-ATP on the ‘interfering’ protein was reminiscent of a reportedly ‘NR-specific inhibitor’ from spinach leaves, termed NRI [8–11]. This ‘factor’ has been discussed in the literature for over a decade and is still described in reviews today .
Recently, forms of glutamine synthetase (GS) from cauliflower  and Chlamydomonas reinhartii were purified by 14-3-3-affinity chromatography. Moreover, cytosolic GS extracted from leaves of Brassica napus L. and plastid GS from tobacco have been found to be activated and/or stabilized by interaction with 14-3-3 proteins [15,16].
Here, we aimed to purify GS for further characterization of its regulation by 14-3-3 proteins. However, during the first purification step, a poly(ethylene glycol) fractionation, we repeatedly noticed an apparent threefold to fourfold increase in total GS activity compared with the crude extract. A similar observation was made  while purifying soybean hypocotyl glutamine synthetase. We report that this apparent activation of GS is due to separation of GS from a protein that has been identified as a nucleotide pyrophosphatase. We demonstrate that the nucleotide pyrophosphatase has identical properties to the NR ‘interfering’ protein , and shares some of the reported properties of the NRI nucleotide pyrophosphatase [8,11]. However, in contrast to Sasaki et al.  and Sonoda et al. , we have no evidence that the nucleotide pyrophosphatase is a physiologically relevant, specific inhibitor of NR that promotes oligomerization of NR. All the apparent inhibitory effects on NR and GS can be explained simply by the enzymatic properties of the nucleotide pyrophosphatase.
Materials and methods
Purification of a nitrate reductase/glutamine synthetase inhibitor protein
All steps were performed at 4 °C. The outer curd (≈ 1000 g) of two cauliflower's (Brassica oleracea var. botrytis; Tesco Supermarket, Dundee) was homogenized in a Waring Blender in 1 vol ice-cold buffer (50 mm Hepes/OH (pH 7.5), 1 mm dithiothreitol, 0.2 mm phenylmethanesulfonyl fluoride, 1 mm benzamidine and 1% (w/v) insoluble polyvinylpolypyrollidone) and clarified by centrifugation at 12 000 g, 4 °C, for 30 min. After filtration through glass wool and two layers of miracloth, the inhibitor was precipitated with 0–9% poly(ethylene glycol) added from a 50% (w/v) solution of poly(ethylene glycol) 8000 dissolved in 25 mm Tris pH 7.5, 1 mm dithiothreitol, 1 mm MgCl2 (buffer A). After stirring for 15 min and centrifugation at 12 000 g, 4 °C, for 20 min, the pellet was resuspended in 25 mm Mes-OH pH 6, 1 mm dithiothreitol, 1 mm MgCl2 (buffer B) and clarified by centrifugation at 100 000 g, 4 °C, for 45 min. The sample was then filtered through a 0.4-µm syringe filter and loaded at 3 mL·min−1 onto a (6 × 1.6 cm) Hiload S-Sepharose column equilibrated in buffer B and eluted with a 0–0.4 m NaCl gradient in buffer B over 200 mL with 5-mL fractions. Peak fractions were pooled and dialysed into 25 mm Tris pH 8.5, 1 mm dithiothreitol and 1 mm MgCl2 (buffer C). The sample was filtered through a 0.2-µm syringe filter and loaded onto a HR (5/5) Mono-Q anion-exchange column equilibrated in buffer C and fractionated over 20 mL with a 0–0.5 m NaCl gradient in buffer C with 1-mL fractions. Peak fractions were pooled and dialysed into buffer A, concentrated to less than 200 µL, and chromatographed on a Superose 12 gel filtration column equilibrated in buffer A plus 100 mm NaCl at 0.4 mL·min−1. Fractions of 0.2 mL were collected. Peak fractions were pooled, dialysed into buffer B and fractionated over 2 mL with 0.1-mL fractions and a 0–0.3 m NaCl gradient in buffer B on a Mono-S (PC 1.6/5) cation-exchange column using a Pharmacia Smart chromatography system. The Superose 12 column was calibrated with the following standards: thyroglobulin (670 kDa), γ-globulin (158 kDa), bovine serum albumin (66 kDa), ovalbumin (44 kDa), myoglobin (17 kDa).
When using concanavalin A (Con-A) Sepharose, the pooled fractions from the Mono-Q step were made 0.25 m NaCl and loaded at 1 mL·min−1 onto a 1-mL Con-A-Sepharose column equilibrated in buffer A plus 0.25 m NaCl. The column was washed with buffer A plus 0.25 m NaCl and protein eluted with 0.25 m methyl α-d-glucopyranoside in buffer A plus 0.25 m NaCl.
GS activity was measured by the formation of γ-glutamyl hydroxamate using the transferase assay . Reaction mixtures contained, in a final volume of 100 µL: 50 mm Hepes/OH pH 7.7, 50 mm monosodium glutamate, 7.5 mm ATP, 75 mm MgCl2, 0.5 mm EDTA, 1 mm dithiothreitol and clarified source of GS. A 9–20% poly(ethylene glycol) fraction from cauliflower curd was the source of GS for routine assays of the GS inhibitor, and each inhibitor assay contained 4 milliunits (mU) of GS activity. One mU of GS activity produced 1 nmol γ-glutamyl hydroxamate per min at 30 °C.
Reactions were started by the addition of hydroxylamine (pH 7.2) to a final concentration of 2.5 mm. After incubation for 10 min at 30 °C, assays were stopped by the addition of 25 µL of a 1 : 1 : 1 mixture containing 10% (w/v) FeCl3·H2O in 0.2 m HCl, 50% (v/v) HCl and 24% (w/v) trichloroacetic acid. The A504 was measured and the γ-glutamyl hydroxamate produced was quantified using commercial γ-glutamyl hydroxamate as standard. Control assays were performed in the absence of ATP to ensure that the reaction was dependent on ATP.
Nitrate reductase (NR) was assayed in a total volume of 100 µL in buffer D (50 mm Hepes pH 7.5, 10 mm MgCl2, 10 µm FAD, 1 mm dithiothreitol). Assays were incubated for 5 min at 30 °C and the reaction initiated by the addition of 50 µL buffer E (buffer D containing 2 mm KNO3 plus 0.5 mm NADH). After 5 min, reactions were stopped with 10 µL of 0.5 m zinc acetate and NADH removed by adding 10 µL 155 µm phenazine methosulfate and incubating for 20 min at room temperature in darkness. Sulfanilamide (50 µL of 1% (w/v) in 3 m (HCl) and 50 µL of 0.02% (w/v) N-(1-naphthyl)ethylenediamine dihydrochloride were added. After 5 min, the mixtures were clarified by centrifugation at 16 000 g, 4 °C, for 2 min and the amount of nitrite was determined by measuring A540 and comparing with a standard curve. For routine assays of NR inhibitor, NR was partially purified by ammonium sulfate fractionation (0–30%) of a spinach leaf crude extract prepared as described in . Each inhibitor assay contained 0.5 mU of NR, where one mU of NR activity is defined as 1 nmol nitrite produced per min at 30 °C.
Assays of NADH hydrolysis by purified NR were followed continuously as the decrease in A340 in the presence of NADH (0.4 mm or as stated in results) and 2 mm KNO3.
Nucleotide pyrophosphatase assays were performed as described in Frick and Bessman  using the coupled enzyme assay. In stage I of the assay, 5 µL of purified NR/GS inhibitor was incubated with the indicated amounts of NADH, FAD, or ATP in 100 mm Tris pH 7.5, 0.1 mm MgCl2 in a total volume of 100 µL for 10 min at 30 °C after which the reaction tube was heated at 95 °C for 5 min. The AMP generated was determined as follows in stage II of the assay. Fifty microlitres of the AMP-containing sample was added to 950 µL of reaction mixture (62 mm Tris pH 7.5, 20 mm KCl, 6 mm ATP, 10 mm MgCl2, 4 mm phosphoenolpyruvate, 0.4 mm NADH, 10 U·mL−1 lactate dehydrogenase (Worthington) and 10 U·mL−1 pyruvate kinase. The reaction was started with the addition of 5 U of adenylate kinase and monitored by A340 and converted to nmol using a molar extinction coefficient of 6.22 m−1·cm−1. The change in A340 was subtracted from a control with no enzyme, which was run parallel to each assay. The assays which contained ATP in stage I were diluted 100-fold for stage II. All assays were performed at least in duplicate.
NADH oxidase in a desalted, cell-free extract of Escherichia coli DH5α was assayed continuously as the decrease in A340 in the presence of NADH (0.4 mm) in 50 mm Mops/NaOH, pH 7.3, 5 mm MgCl2.
Amino acid sequencing
Proteins (≈ 5 µg) were excised from SDS/PAGE gels that had been lightly stained with Coomassie blue. The gel slices were washed in Milli-Q water (5 × 1 mL) for 1 h, brought to near dryness by rotary evaporation, suspended in 250 µL buffer F (50 mm Tris/HCl pH 8.0, 0.01% alkylated Triton X-100) containing 1 µg of alkylated trypsin (Roche) and incubated with shaking for 20 h at 30 °C. The supernatant was removed and a further 250 µL buffer F without trypsin was added for 4 h. The combined supernatants were dried to 50 µL and applied to capillary C18 column (0.5 × 150 mm) from Applied Biosystems (Warrington, UK) equilibrated in 0.1% (v/v) trifluoroacetic acid attached to a Applied Biosystems ABI 173A Microblotter Capillary HPLC system. The column was developed with a linear acetonitrile gradient in 0.09% (v/v) trifluoroacetic acid with an increase in acetonitrile concentration of 0.5% per min. A214 was recorded with an on-line monitor. The flow rate was 7 µL·min−1. Selected peptides were sequenced on an Applied Biosystems 476A protein sequencer. For N-terminal sequencing, proteins were run on SDS/PAGE, transferred to Problott (Applied Biosystems), stained on the membrane as described by the manufacturer, and the relevant bands were excised and sequenced from the membrane.
Properties of a nitrate reductase inhibitor
On reporting the purification and identification of 14-3-3 proteins as the NR inhibitor protein, NIP, we reported “another protein that interfered with the NIP-14-3-3 assay, and that was eluted in the 0.2 m NaCl wash” during anion-exchange chromatography of extracts of spinach leaves , or cauliflower (Fig. 1). In contrast to 14-3-3 proteins, the ‘interfering’ protein inhibited both phosphorylated and dephosphorylated NR (Fig. 1 and data not shown). A protein inhibitor of NR, NRI, that bound to Con-A had been reported previously [9,10]. Similar to NRI, we found that our interfering protein bound to Con-A (Fig. 2). The inhibition of NR by the interfering protein could be prevented by addition of either EDTA, ATP, or AMP directly to NR activity assays (Fig. 1 and data not shown). A similar Con A-binding NR inhibitor was also present in extracts of Chlamydomonas (M. Pozuelo Rubio, MRC Unit, University of Dundee, UK, personal communication).
Identification and purification of a GS inhibitor with identical properties to the NR inhibitor
We partially purified GS from cauliflower curd in preparation for studies aimed at discovering whether GS activity is regulated by its interaction with 14-3-3 proteins. During the first purification step, a 0–9% poly(ethylene) glycol fractionation, we noticed an apparent threefold to fourfold increase in total GS activity compared with the activity in the crude extract (GS activity remains in the 0–9% poly(ethylene glycol) supernatant). The apparent activation of GS did not occur if a 0–20% poly(ethylene glycol) fraction was made. Addition of a 0–9% poly(ethylene glycol) fraction inhibited the GS activity in a 9–20% poly(ethylene glycol) cut (Table 1).
Table 1. Purification of GS inhibitor from cauliflower curd. One unit of activity is the amount of inhibitor that will decrease the activity of 4 mU of GS by 50% during the assay.
Specific activity (U·mg−1)
0–9% Poly(ethylene glycol)
The protein responsible for the apparent inhibition of GS was purified further (Fig. 2 and Table 1). Similar to the interfering protein, the GS inhibitor was eluted from Q-Sepharose by 0.2 m NaCl (Table 1 and data not shown). Consistent with the possibility that the NR inhibitor and GS inhibitor were identical proteins, the EDTA-sensitive NR inhibitor and GS inhibitor cochromatographed throughout the purification (Fig. 2 and data not shown). The final fractions contained three protein bands with apparent molecular masses of 70, 47 and 45 kDa on SDS/PAGE that were most abundant in the fractions containing highest NR/GS inhibitory activity (Fig. 2). The NR/GS inhibitor behaved on Superose 12 gel filtration as a 55-kDa protein (not shown). The intact proteins and peptides produced from tryptic digests of the protein bands were sequenced. blast searches of sequence databases revealed that the band of 70 kDa belonged to the acyl-CoA oxidase protein family, while all of the peptides derived from the 47 and 45 kDa bands matched most closely with an Arabidopsis thaliana nucleotide pyrophosphatase-like protein (Table 2). Sequenced peptides covered 22% of the nucleotide pyrophosphatase with 71% identity and 79% similarity. The N-terminal sequence obtained began at residue 49 and the predicted mass of the protein from this residue onwards (45.9 kDa) closely matches the observed mass of the sequenced bands on the gel (Fig. 2) indicating the protein was proteolytically cleaved.
Table 2. Sequences, identities, and accession numbers of proteins that were copurified with inhibitory activity towards glutamine synthetase and nitrate reductase. *, N-terminal sequence.
All of the glutamine synthetase and nitrate reductase inhibitory activity bound to Con-A and was eluted with 0.25 m methyl α-d-glucopyranoside, with a recovery of between 20 and 40% activity in different preparations. However, the Con-A step would not have improved the overall purification because both the acyl-CoA oxidase and the nucleotide pyrophosphatase bound to, and were eluted from, the Con-A column, as determined by amino acid sequencing of the ≈ 70 and ≈ 47 kDa proteins seen in Fig. 2C. Similarly, both proteins bound to AMP-Sepharose (not shown).
The fractions from the final Mono-S column containing NR/GS inhibitory activity were found to catalyse the production of AMP from NADH, ATP and FAD (Table 3). Ninety percent of the amount of NADH used in our standard NR assay, and 60% of the amount of ATP in a GS assay, was converted into AMP within 10 min at 30 °C by an amount of a Mono Q fraction that appeared to inhibit NR by 55% (Table 3). These findings suggest that the NR/GS inhibitor functions during both the assay preparation and the assays by converting the cofactors necessary for the NR or GS reactions into AMP. Consistent with this notion, the hydrolysis of NADH to NAD by either purified NR or an extract of E. coli were clearly inhibited by the Mono-S fractions in a time-dependent manner, and transiently restored by adding extra NADH (data not shown). The inhibitor preparation had no effect on the activity of NR when reduced methylviologen was used as electron donor instead of NADH (not shown). In addition, using NADH as substrate, the enzyme displayed a Km of 70 µm and a Vmax of 20 µmol AMP produced per min per mg protein. The enyzme activity was inhibited by >95% using 1 mm Cu2+, Zn2+ and Ni2+ (all as chloride salts) in the assay, in common with other nucleotide pyrophosphatases . The nucleotide pyrophosphatase was inhibited 54% by 10 µm NaF, but was unaffected by 10 µm NaCl, again when employing NADH as substrate. Similarly, the apparent inhibition of NR was blocked 48% by 10 µm NaF, using an amount of Mono S fraction that gave 20% inhibition of NR in the standard assay.
Table 3. Generation of AMP from NADH, ATP and FAD by purified NR/GS inhibitor in a 100-µL incubation at 30 °C for 10 min, using the amounts and concentrations of ATP and NADH used in standard GS and NR assays, respectively. Data are presented as mean ± SEM.
AMP generated (nmol)
% cofactor hydrolysed to AMP
ATP (750 nmol)
662 ± 17
NADH (50 nmol)
33 ± 1.4
FAD (5 nmol)
1.64 ± 0.20
An activity that inhibited both GS and NR was highly purified from cauliflower (B. oleracea var. botrytis) extracts. The final preparation contained an acyl-CoA oxidase and a second protein of the plant nucleotide pyrophosphatase family. Nucleotide pyrophosphatases belong to a family of widely distributed hydrolases that are active on a variety of derivatives of nucleoside diphosphates (hence the name nudix hydrolases), and/or non-nucleotide diphosphoinositol polyphosphates, and characterized by the mutT motif (GX5EX7REUXE3GU; where U represents one of the bulky hydrophobic amino acids, usually I, L or V) [21–25]. These enzymes are often extracellular and their physiological substrates may include signalling metabolites, including toxic derivatives.
The purified protein catalysed the hydrolysis of NADH, ATP and FAD (Table 3). Moreover, the apparent inhibition of NR is consistent with the hydrolysis of NADH by the nucleotide pyrophosphatase, and the ability of ATP to relieve the inhibition of NR (Fig. 1) is because ATP protects NADH from hydrolysis, most likely by providing an alternative substrate for the nucleotide pyrophosphatase (Table 3). The apparent inhibition of GS can be explained by hydrolysis of the Mg-ATP substrate by the nucleotide pyrophosphatase, and generation of 5′-AMP, a GS inhibitor . While we were unable to separate the active nucleotide pyrophosphatase from the acyl-CoA oxidase by a number of procedures that maintained the activity of the nucleotide pyrophosphatase (Fig. 1, Table 2 and data not shown) there is no obvious mechanistic reason to implicate the acyl-CoA oxidase in the apparent inhibition of NR and GS.
Sonoda et al. [11,12] reported purification of an irreversible inhibitor of NR, termed NRI, and its identification as a spinach nucleotide pyrophosphatase. In contrast to our findings, the enzyme purified by Sonoda et al. [11,12] did not affect the NADH-dependent activities of glutamate dehydrogenase or lactate dehydrogenase , and was speculated to be NR specific with a possible physiological role in NR inactivation during leaf senescence [11,12]. Moreover, NRI was reported to promote the assembly of NR into oligomeric forms that had a retarded electrophoretic mobility, and oligomerization was suggested to be mediated via action of the nucleotide pyrophosphatase on the FAD cofactor bound to NR [8–12]. In contrast, we have no evidence that the inhibitory activity we have purified here causes NR polymerization. Thus, while the cauliflower nucleotide pyrophosphatase has very similar chromatographic properties, size on SDS/PAGE and inhibition by EDTA to the enzyme purified by Sonoda et al. [11,12], we have no clear evidence to suggest that the nucleotide pyrophosphatase has regulatory effects on NR or GS in planta. The binding to Con-A indicates that the protein is glycosylated and may therefore be extracellular, as are many nucleotide pyrophosphatases .
The nonspecific hydrolysis by nucleotide pyrophosphatases has previously caused confusion in regulatory systems that use ATP and adenine dinucleotides . We suggest that the nucleotide pyrophosphatase may have caused much confusion in studies on NR regulation. For example, we and others have found that the nucleotide pyrophosphatase activity in plant extracts can often be so high that when Mg-ATP is added to an extract the activity of NR appears to go up because Mg-ATP prevents the hydrolysis of NADH, instead of down due to the effect of phosphorylation and binding to NIP-14-3-3 proteins (Fig. 1). In addition, NR inactivating factors found in rice and Neurospora extracts were dependent on NADH and blocked by EDTA [28–30] and a protein inhibitor was reported  that had similar effects on GS to the nucleotide pyrophosphate that we have found here.
5′-AMP has been widely reported to activate NR and GS in cell-free extracts [7,14,26,32,33]. The mechanism of 5′-AMP activation of NR may, in part, involve binding to 14-3-3 . However, the apparent inhibition of NR by the nucleotide pyrophosphatase was largely relieved by millimolar concentrations of 5′-AMP, presumably acting as a product inhibitor, and it seems likely that this mechanism contributes to the reported 5′-AMP activation of NR and GS.
Both the NADH pyrophosphatase activity and the inhibition of NR were inhibited ≈ 50% by 10 µm NaF, which is consistent with the proposal that the apparent inhibition of NR is due to the pyrophosphatase. Other nucleotide pyrophosphatases have been reported to be inhibited by micromolar concentrations of NaF . The inhibitory effect of NaF on the nucleotide pyrophosphatase is useful: NaF is commonly used to inhibit protein serine/threonine phosphatases, including the PP2A that dephosphorylates NR, and we know that at concentrations up to 15 mm, fluoride has no obvious effect on NR kinases  and up to at least 2.5 mm has no effect on NR activity or 14-3-3 binding . We therefore suggest that analysis of the regulation by phosphorylation/14-3-3 proteins of NR or GS in crude fractions be performed in the presence of >1 mm NaF, or after passing through a Con-A column to remove the ‘interference’ from extracts containing high nucleotide pyrophosphatase activity.
This work was supported by funds from the UK Biotechnology and Biological Sciences Research Council (to C. M.).