Post-translational regulation of cytosolic glutamine synthetase by reversible phosphorylation and 14-3-3 protein interaction


  • Jørgen Finnemann,

    Corresponding author
      *For correspondence (fax +45 35 28 34 60; e-mail
    Search for more papers by this author
  • Jan K. Schjoerring

    1. Plant Nutrition Laboratory, Department of Agricultural Sciences, The Royal Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Copenhagen, Denmark
    Search for more papers by this author

*For correspondence (fax +45 35 28 34 60; e-mail


Regulation of the cytosolic isozyme of glutamine synthetase (GS1; EC was studied in leaves of Brassica napus L. Expression and immunodetection studies showed that GS1 was the only active GS isozyme in senescing leaves. By use of [γ-32P]ATP followed by immunodetection, it was shown that GS1 is a phospho-protein. GS1 is regulated post-translationally by reversible phosphorylation catalysed by protein kinases and microcystin-sensitive serine/threonine protein phosphatases. Dephosphorylated GS1 is much more susceptible to degradation than the phosphorylated form. The phosphorylation status of GS1 changes during light/dark transitions and depends in vitro on the ATP/AMP ratio. Phosphorylated GS1 interacts with 14-3-3 proteins as verified by two different methods: a His-tag 14-3-3 protein column affinity method combined with immunodetection, and a far-Western method with overlay of 14-3-3–GFP. The degree of interaction with 14-3-3-proteins could be modified in vitro by decreasing or increasing the phosphorylation status of GS1. Thus, the results demonstrate that 14-3-3 protein is an activator molecule of cytosolic GS and provide the first evidence of a protein involved in the activation of plant cytosolic GS. The role of post-translational regulation of cytosolic GS and interactions between phosphorylated cytosolic GS and 14-3-3 proteins in senescing leaves is discussed in relation to nitrogen remobilization.


Glutamine synthetase (GS; EC is the key enzyme involved in ammonia (NH3) assimilation in plants ( Lea et al., 1990 ). GS catalyses the ATP-dependent condensation of NH3 with glutamate to produce glutamine. Plant GS is an octameric isozyme with a native molecular mass of approximately 320 or 380 kDa depending on whether localized in the cytosol (GS1) or in plastids/chloroplasts (GS2; Lea et al., 1990 ). In most plants, GS2 is encoded by a single gene per haploid genome, whereas several genes encoding GS1 isozymes have been sequenced ( Ochs et al., 1999 ; Peterman and Goodman, 1991).

The in vivo function of GS2 has been elucidated using genetically modified barley plants ( Wallsgrove et al., 1987 ). The main role is assimilation of NH3 derived from nitrate reduction and photo-respiration. The in vivo role of GS1 depends on the organ in which it is localized. In roots, GS1 constitutes nearly all GS activity and the main role is assimilation of NH3 for translocation and biosynthesis ( Lea et al., 1990 ). In leaves of C3 plants, GS1 seems mainly to be located in the phloem parenchyma ( Edwards et al., 1990 ; Kamachi et al., 1992 ), where it is involved in the re-assimilation of NH3 derived from amino acid catabolism and phenyl propanoid metabolism ( Lea et al., 1990 ; Vincent et al., 1997 ).

It is generally assumed that GS activity in plants is regulated at the transcriptional level, and many reports have focused on this aspect ( Edwards et al., 1990 ; Lam et al., 1996 ; Oliveira et al., 1997 ; Temple et al., 1995 ). Transcription of GS2 is primarily affected by light and nitrogen ( Edwards and Coruzzi, 1989), whereas the transcription of GS1 in roots depends on the level of external nitrogen supply ( Finnemann and Schjoerring, 1999). Other factors such as leaf age and environmental stresses also seem to affect GS1 expression ( Bauer et al., 1997 ; Buchanan-Wollaston and Ainsworth, 1997).

Recent work suggests that GS1 is not only regulated transcriptionally but also post-translationally. Transgenic alfalfa plants with a considerable reduction in the expression of gs1 isogenes showed no reduction in GS1 activity ( Temple et al., 1998 ). Similarly, a threefold increase in GS1 subunit content of soybean roots was measured without a corresponding increase in transcription ( Ortega et al., 1999 ). Oxidative modifications of GS1 resulted in an inactive enzyme more susceptible to degradation than non-oxidized GS1 ( Ortega et al., 1999 ). However, phosphorylation of the GS1 holo-enzyme following ATP hydrolysis was able to provide protection against the oxidative inactivation. Phosphorylation of GS and interaction with 14-3-3 proteins were observed by Moorhead et al. (1999) , but the regulation and function of these processes were not explored.

The purpose of the present work was to investigate post-translational regulation of GS by reversible protein phosphorylation and further interactions with 14-3-3 proteins. GS was extracted from leaves varying in developmental stage from early to late senescence, the latter dominated by cytosolic GS. After di-ammonium sulphate precipitation, the extracted GS was exposed to various treatments promoting phosphorylation and dephosphorylation. Interactions between phosphorylated GS and 14-3-3 proteins were visualized by two methods; a His-tag column affinity purification procedure with His-tag 14-3-3 proteins, and a far-Western method with overlay of 14-3-3–GFP. Both the in vitro and in vivo post-translational regulation of GS were elucidated.


Expression of gs isogenes, GS subunit composition and activity

Expression of gs isogenes in leaves

Analysis of the leaf-specific gs isogene expression revealed that the chloroplastic isogene (Bngsl1) and one of the two cytosolic isogenes (Bngsr2-1) were expressed. High expression of the Bngsl1 gene encoding GS2 was observed in young leaves, but the expression decreased during ageing ( Figure 1). In contrast, the Bngsr2-1 gene encoding GS1 was expressed at low levels in young leaves and high levels in senescing leaves ( Figure 1).

Figure 1.

Expression of gs isogenes during leaf ageing of Brassica napus.

Aliquots (50 ng) of total RNA were used in an optimized multiplex RT–PCR reaction. (a) Typical gel picture of the gs isogene expression pattern, with the RT–PCR-amplified fragments for the isogenes encoding chloroplastic (Bngsl1) and cytosolic (Bngsr2-1) GS indicated. Actin was used as internal standard. The triangle illustrates the decline in chlorophyll content during ageing of the leaves used for RT–PCR. (b) Quantified isogene expression levels for Bngsl1 and Bngsr2-1 as a function of leaf chlorophyll content. Curves were fitted by minimizing least-squares deviation.

GS subunit composition in leaves

By use of a specific GS antibody, two polypeptides of 44 and 40 kDa, corresponding to GS2 and GS1, respectively, were recognized ( Figure 2). The GS2/GS1 subunit ratio depended on the degree of senescence, with a high amount of the GS2 polypeptide in young leaves and a low content in senescing leaves, opposite to the pattern for the GS1 polypeptide ( Figure 2). At very pronounced senescence (< 0.2 mg chlorophyll/g FW), only the GS1 subunit was present, together with an immunoreactive fragment, most possibly derived from degraded polypeptides of the GS holo-enzyme ( Figure 2).

Figure 2.

Abundance of GS polypeptides in B. napus during leaf ageing.

(a) Typical Western blot of GS subunits with 5 µg total protein subjected to SDS–PAGE followed by immunodetection. The Rubisco large subunit (LSU) is shown together with GS2 and GS1, and a possible degradation fragment (f) of GS. The triangle illustrates the decline in chlorophyll content during ageing of the leaves used for Western blotting. (b) Quantified GS subunit content as a function of leaf chlorophyll content. GS subunit content is related to the maximum value (100%). Curves were fitted by minimizing least-squares deviation.

In green leaves (> 1.5 mg chlorophyll/g FW), the Bngsl1 expression matched the amount of GS2 subunit, whereas a relatively high amount of GS1 subunit was measured despite low Bngsr2-1 expression ( Figures 1 and 2). In senescent leaves (< 0.2 mg chlorophyll/g FW), the GS2 subunit totally disappeared despite expression of the Bngsl1 isogene. By contrast, the Bngsr2-1 isogene encoding GS1 showed highest expression during senescence and the GS1 subunit content paralleled the expression pattern ( Figures 1 and 2).

GS activity in leaves

A nearly constant level of total specific GS activity (GS1 + GS2), corresponding to about 4.40 µmol γ-glutamylhydroxamate mg−1 protein h−1, was found throughout leaf age (data not shown). Combined with the fact that the amount of total GS protein decreased during ageing ( Figure 2), this shows that either an increasing fraction of the GS subunits were assembled into active GS holo-enzyme or that the GS holo-enzyme was regulated post-translationally.

Post-translational regulation of cytosolic GS

Phosphorylation of GS isozymes

To verify that GS is a phospho-protein, a desalted, 40–50% (NH4)2SO4 precipitate (see Experimental procedures) from young and senescing leaves was pre-incubated with [γ-32P]ATP. Subsequently, proteins were separated by SDS–PAGE and transferred to a nitrocellulose membrane ( Figure 3). Strong radioactive bands were visualized both in young and senescing leaves, and subsequent immunodetection of GS revealed that both GS isozymes had become phosphorylated ( Figure 3). The degree of phosphorylation of the GS isozymes followed the subunit content, i.e. highest phosphorylation of GS2 and GS1 in young and senescing leaves, respectively ( Figure 3). Extending the exposure time revealed several radioactive bands (phospho-proteins) of different sizes in young leaves and fewer bands in senescing leaves (data not shown).

Figure 3.

Phosphorylation of GS by pre-incubation with 32P-labelled ATP.

Desalted 40–50% (NH4)2SO4 precipitate from young and senescing B. napus leaves (1.5 and 0.4 mg chlorophyll/g FW; Figure 2) was pre-incubated for 30 min with [γ-32P]ATP prior to SDS–PAGE and membrane transfer. Following autoradiography ([γ-32P]ATP), GS subunits were immunodetected by GS antibody.

Reversible phosphorylation of GS1

To examine the post-translational regulation of GS1, experiments were performed with a desalted, 40–50% (NH4)2SO4 precipitate (see Experimental procedures) from senescing leaves containing GS1 subunits only (< 0.2 mg chlorophyll/g FW; Figure 2). Together with protein kinases and protein phosphatases ( MacKintosh et al., 1995 ), the GS1 enzyme and (as revealed later) proteases are present in the desalted (NH4)2SO4 precipitate. Pre-incubation of the desalted precipitate at 30°C resulted in a > 50% reduction of GS1 activity, but elevated ATP levels (0.1–2.0 m m) completely prevented the loss of activity ( Figure 4). Inclusion of the protein phosphatase inhibitor microcystin (a potent inhibitor of PP1 and PP2A activity) together with ATP and excessive amounts of Mg2+ only resulted in a slight increase in activity ( Figure 4). Pre-incubation with EDTA (which binds free Mg2+ and thereby inhibits nitrate reductase (NR) protein kinases; Glaab and Kaiser, 1996) resulted in a 75–90% loss of activity which could be counteracted by microcystin ( Figure 4). Addition of either ATP, EDTA, excess Mg2+ or microcystin without pre- incubation did not affect GS1 activity (data not shown). The ATP analogue ATP-γ-S; which can provide a non-hydrolysable phosphate group to proteins, was used to study whether the dephosphorylation of GS1 directly involved hydrolysis of the ATP-donated phosphate group. After pre-incubation with either ATP or ATP-γ-S without loss of activity, addition of excess amounts of EDTA immediately reduced the activity ( Figure 5). The decline in activity due to the EDTA-enhanced dephosphorylation was less for the ATP-γ-S-treated GS1, which showed a more than twofold higher activity than the ATP-treated GS1 ( Figure 5).

Figure 4.

Effects of different combinations of ATP, EDTA, microcystin (Mc) and Mg2+ on GS1 activity.

GS1 activity was determined after pre-incubation of a desalted 40–50% (NH4)2SO4 precipitate from senescing B. napus leaves (< 0.2 mg chlorophyll/g FW; Figure 2) with combinations of ATP, EDTA, microcystin and Mg2+ as indicated. Values are means ± SE (n = 4–6).

Figure 5.

Reversal of the ATP- (or ATP-γ-S)-dependent phosphorylation of GS1 by favouring conditions for dephosphorylation.

Pre-incubation of desalted 40–50% (NH4)2SO4 precipitate from senescing leaves (< 0.2 mg chlorophyll/g FW; Figure 2) with ATP or non-hydrolysable ATP-γ-S (2 m m) was followed by addition of EDTA (10 m m). At various pre-incubation times, aliquots were taken to assay GS1 activity. Data are the means of two experiments. The insert shows the increase in GS1 activity after phosphorylation with ATP or the ATP analogue ATP-γ-S prior to addition of 14-3-3 protein to the GS1 extract.

Reactivation of GS1 by phosphorylation and interaction with 14-3-3 proteins

Pre-incubation under deactivating conditions (with EDTA) was followed by a loss of activity ( Figure 6, 1st pre-incubation). After desalting to remove EDTA, addition of EDTA to the desalted extract before the 2nd pre-incubation was followed by a further loss of activity . Reactivation by phosphorylation of GS1 was obtained by addition of ATP to the desalted extract before the 2nd pre-incubation, resulting in a threefold higher GS1 activity relative to that in the EDTA treatment. Addition of both ATP and 14-3-3 proteins to the desalted GS1 extract before the 2nd pre-incubation further increased the activity of the reactivated GS1 enzyme ( Figure 6, 2nd pre-incubation). In the latter case, a more than fivefold increase in GS1 activity was observed relative to the EDTA treatment. Phosphorylation of GS1 with non-hydrolysable phosphate (ATP-γ-S) further enhanced the stimulating effect of 14-3-3 proteins on GS1 activity ( Figure 5, insert).

Figure 6.

Reactivation of GS1 by phosphorylation and interaction with 14-3-3 protein.

Deactivation of GS1 was performed by pre-incubation of a desalted 40–50% (NH4)2SO4 precipitate from senescing leaves (< 0.2 mg chlorophyll/g FW) with 10 m m EDTA (1st pre-incubation). After desalting to remove EDTA, further deactivation, stabilization or reactivation of GS1 activity was promoted by pre-incubation with either 10 m m EDTA, 2 m m ATP, or 2 m m ATP plus 14-3-3 proteins in excess (2nd pre-incubation). At various pre-incubation times, aliquots were taken to assay GS1 activity.

Dephosphorylation of GS1 is an initial step in protein degradation

Loss of activity under conditions favouring dephosphorylation of GS1 may either be due to inactivation or degradation of the holo-enzyme. Western blot of GS extracts from senescing leaves (< 0.2 mg chlorophyll/g FW; Figure 2) revealed that pre-incubation at 30°C resulted in degradation of the GS1 subunit ( Figure 7a). The GS1 subunit was protected against degradation when pre-incubated with ATP or ATP together with microcystin. Under conditions favouring degradation (pre-incubation with excess EDTA), GS1 was completely degraded ( Figure 7a). The loss of activity caused by degradation of GS1 under dephosphorylating conditions occurred during the first 20 min after adding excess EDTA ( Figure 7c). Addition of 14-3-3 proteins under conditions favouring dephosphorylation (excess EDTA) resulted in several-fold higher GS1 activity ( Figure 7c). To test whether the 14-3-3 both protected GS1 against degradation and caused increased activity, pre-incubation experiments of GS extracts with 14-3-3 proteins and increasing amounts of phosphoserine-Raf-259 proteins (Raf1 protein) were performed ( Figure 7b). Phosphorylated Raf1 proteins are known to compete with phospho-proteins for binding of 14-3-3 proteins ( Muslin et al., 1996 ). An increasing concentration of Raf1 protein in the pre-incubation mixture should reduce the interaction between 14-3-3 and GS1, thereby blocking the protective effect of the 14-3-3 interaction and decreasing the amount of GS1 protein. However, the opposite was observed, i.e. the amount of immunodetected GS1 protein increased with increasing addition of Raf1 proteins to the pre-incubation mixture ( Figure 7b). Protection of GS1 against degradation was due to elevated amounts of total protein in the pre-incubation mixture ( Figure 7b). Thus, GS1 is less susceptible to degradation when phosphorylated, and interactions between GS1 and 14-3-3 proteins increase the GS1 activity.

Figure 7.

Phosphorylation of GS1 delays degradation.

(a) Western blot of desalted 40–50% (NH4)2SO4 precipitate from senescing leaves (< 0.2 mg chlorophyll/g FW; Figure 2) pre-incubated with different combinations of ATP, microcystin (Mc) or EDTA. (b) Pre-incubation of desalted 40–50% (NH4)2SO4 precipitate from senescing leaves with 14-3-3 proteins and increasing additions of phosphorylated Raf1 protein (n = 4). (c) Pre-incubation with ATP, EDTA and 14-3-3 proteins. At various pre-incubation times, aliquots were taken to assay GS activity. Values are means ± SE (n = 3–5).

Phosphorylated GS1 and 14-3-3 interactions

To show the interaction between phosphorylated GS1 and 14-3-3, GS extracts from senescing leaves were incubated with His-tag 14-3-3 proteins and subsequently loaded and purified on a Ni-NTA column capable of withholding His-tag proteins (see Experimental procedures). After loading, purification and three successive wash steps, elution of His-tag proteins interacting with GS was performed. To visualize phosphorylated GS1:14-3-3 interactions, successive elution fractions were Western-blotted using a specific GS antibody ( Figure 8a). No GS1 protein could be detected after the 3rd wash step, indicating a purified column from which non-phosphorylated GS1 proteins and/or phosphorylated GS1 proteins interacting with endogenous 14-3-3 proteins ( Figure 8a) had been rinsed. The subsequent elution (lane 4) showed a distinct GS1 band, and thus an interaction between His-tag 14-3-3 proteins and phosphorylated GS1 (GS1-P). The co-elution of His-tag 14-3-3 proteins interacting with GS1 protein was visualized by use of specific anti-His antibody ( Figure 8b). Pre-incubation under conditions favouring phosphorylation (ATP, Mg2+, microcystin) resulted in higher GS1-P:14-3-3 interactions, while conditions favouring dephosphorylation (EDTA) lowered the interaction ( Figure 8c). A complementary method, a far-Western overlay of GS extracts from middle-aged and senescing leaves revealed that both phosphorylated GS1 and GS2 bound to a green fluorescent protein linked to 14-3-3 proteins (14-3-3–GFP; Figure 9). In addition, other phosphorylated proteins, predominantly Rubisco, interacted with the 14-3-3–GFP ( Figure 9).

Figure 8.

Interactions between phosphorylated GS1 (GS1-P) and 14-3-3 proteins.

(a) GS1 extracts from senescing leaves (< 0.2 mg chlorophyll/g FW; Figure 2) were incubated with His-tag 14-3-3 proteins for 1 h (on ice) and the mixture was loaded onto a Ni-NTA column. After loading, purification was performed with three successive wash steps before elution of His-tag proteins binding GS1. After elution, different column fractions were Western-blotted using specific GS antibody: lane 1, flow-through after loading; lanes 2 and 3, flow-throughs from 1st and 3rd wash steps; lanes 4 and 5, 1st and 2nd elution steps. (b) Western blot of Ni-NTA elution fractions using either specific GS antibody (lane 1) or specific anti-His antibody (lane 2); f is a possible degradation fragment of GS. (c) After pre-incubation of GS1 extracts containing 500 µg total protein from senescing leaves (< 0.2 mg chlorophyll/g FW; Figure 2) with different combinations of ATP, EDTA, microcystin (Mc) and extra Mg2+ (see pre-incubation conditions), 50 µg His-tag 14-3-3 proteins were added and the mixtures were incubated 1 h (on ice) and loaded onto a Ni-NTA column. After loading, purification and elution, the elution fraction from each single treatment was Western-blotted using specific GS antibody.

Figure 9.

Far-Western blot overlaid with 14-3-3–GFP.

Western blot of GS extract from leaves (0.7 or 0.2 mg chlorophyll/g FW; Figure 2) overlaid with buffer containing 14-3-3–GFP (left blot). After fluorescence scanning, localization of GS subunits was performed using specific GS antibody (right blot).

Light/dark conditions regulate the reversible phosphorylation

In vivo regulation of GS activity

Incubation of desalted GS extracts with excessive amounts of His-tag 14-3-3 proteins provided evidence for an increased phosphorylation of GS1 in leaves sampled during dark conditions relative to GS1 in leaves sampled in the light ( Figure 10). This was observed both in senescing and young leaves. The difference in the phosphorylation status was most pronounced in senescing leaves (< 0.2 mg chlorophyll/g FW), where a more than fourfold increase during darkness was found ( Figure 10). In young leaves (> 1.5 mg chlorophyll/g FW), GS2 was also phosphorylated and interacted with 14-3-3 proteins, but no variation in the phosphorylation status could be observed ( Figure 10). Diurnals of the reversible GS1 phosphorylation were performed by incubation of desalted GS extracts made from senescing leaves 1 and 14 h into the light period and 1 and 5 h into the dark period with His-tag 14-3-3 proteins (data not shown). The lowest degree of GS1 phosphorylation was observed in leaves exposed to a long light period (14 h of light). The phosphorylation level increased nearly threefold after 1 h in darkness when compared to the phosphorylation level in leaves exposed to a long light period (14 h of light). Extending the dark period only slightly decreased the GS1 phosphorylation levels, whereas the shift from dark to light (1 h in light) strongly decreased the phosphorylation level almost to the level measured in leaves exposed to a long light period (14 h of light).

Figure 10.

In vivo regulation of GS phosphorylation in the leaves of B. napus.

Desalted GS extracts from young (> 1.5 mg chlorophyll/g FW; Figure 2) or senescing leaves (< 0.2 mg chlorophyll/g FW; Figure 2) taken in the light or dark period were incubated with excess amounts of His-tag 14-3-3 proteins (n = 3; see Experimental procedures for further details).

The light/dark regulation of phosphorylation was also measured in vitro by addition of 14-3-3 proteins to GS extracts from senescing leaves (< 0.2 mg chlorophyll/g FW) sampled in light or dark (data not shown). The GS1 activity increased after addition of 14-3-3 proteins; however, the activity increase was approximately 70% higher in GS1 extracts from leaves sampled in darkness compared to light.

Nitrogen content and translocation in leaves

Nitrogen import/export in leaves of B. napus

Senescing leaves (< 0.5 mg chlorophyll/g FW) exporting nitrogen had a high glutamine content and glutamine was by far the most abundant amino acid in the phloem sap ( Table 1). Together with the fact that GS1 was the only GS isozyme at this stage ( Figure 2), this suggests a distinct role for GS1 in nitrogen remobilization. In the transition phase between sink and source (1.0–1.5 mg chlorophyll/g FW), the glutamine content was low, whereas it increased in young leaves (≥ 2.0 mg chlorophyll/g FW) both in the leaf blade and in the phloem sap ( Table 1).

Table 1.  NH4+ and amino acid levels in leaf tissue and phloem sap of B. napus
  Chlorophyll a + b level in leaf blade (µg g−1 fresh weight)
  0–500501–10001001–15001501–2000≥ 2001
  1. Phloem sap extractions were performed from petioles of leaves containing chlorophyll as indicated (for extraction procedure, see Experimental procedures) .aµmol g −1 FW. bµmol amino acid-N µmol −1 total amino acid-N. cµmol NH 4+-N µmol−1 (amino acid-N + NH4+-N). dµmol amino acid-N µmol −1 (amino acid-N + NH4+-N).

Leaves (n = 8–12) NH4+a0.25 ± 0.020.27 ± 0.010.42 ± 0.030.46 ± 0.070.44 ± 0.03
Asn b (%)3.7 ± 0.56.0 ± 1.03.8 ± 0.56.0 ± 1.13.8 ± 0.5
Asp b(%) 3.9 ± 0.75.0 ± 0.66.9 ± 0.55.6 ± 0.67.7 ± 0.4
Gln b (%)15.2 ± 1.49.4 ± 0.69.4 ± 0.410.2 ± 0.514.3 ± 1.2
Glu b (%)21.3 ± 1.223.7 ± 0.523.2 ± 0.720.7 ± 1.721.0 ± 1.3
Translocated inNH4+c (%)3.6 ± 1.29.5 ± 1.610.6 ± 3.39.9 ± 2.712.6 ± 3.1
phloem sapAsn d(%) 3.0 ± 0.53.4 ± 0.53.1 ± 1.59.6 ± 0.710.7 ± 1.3
(n = 4) Asp d (%)1.8 ± 0.93.0 ± 0.93.3 ± 1.72.4 ± 0.32.3 ± 0.6
Gln d (%)44.2 ± 2.941.3 ± 6.523.3 ± 2.238.4 ± 2.436.5 ± 2.5
Glu d (%)4.2 ± 0.65.7 ± 0.78.1 ± 0.23.1 ± 0.22.7 ± 0.3


Glutamine synthetase is a phospho-protein in plants

Pre-incubation of GS extracted from young and senescing leaves with [γ-32P]ATP revealed bands corresponding to the immunodetected bands of chloroplastic and cytosolic GS ( Figure 3). In agreement with the age-dependent changes in gs isogene expression and GS subunit content ( Figures 1 and 2), the chloroplastic isozyme (GS2) was more strongly 32P-labelled than the cytosolic isozyme (GS1) in young leaves, and vice versa in senescing leaves ( Figure 3). Thus, in B. napus, GS isozymes are phospho-proteins.

Reversible phosphorylation of GS1

Post-translational regulation by reversible phosphorylation has been shown for enzymes such as nitrate reductase (NR) and sucrose phosphate synthase (SPS; MacKintosh et al., 1995 ; Toroser et al., 1999 ). The present work shows that GS1 is also regulated post-translationally by reversible phosphorylation catalysed by protein kinases and protein phosphatases which together with GS1 are present in a desalted 40–50% (NH4)2SO4 precipitate. Pre-incubation of GS1 with or without substrates for protein kinases (Mg2+ and ATP) indicated that GS1in vivo is phosphorylated ( Figure 4). Dephosphorylation of GS1 is catalysed by a microcystin-sensitive serine/threonine protein phosphatase in the presence of EDTA ( Figures 4–6 and 7a). The sensitivity of the protein phosphatase to microcystin reveals its possible relation to PP2A, a member of the PPP family of protein (serine/threonine) phosphatases which are ubiquitous in all eukaryotic cells ( MacKintosh, 1998). The in vivo degree of phosphorylation cannot be estimated because of degradation of dephospho-GS1 ( Figure 7a).

Reversible phosphorylation of NR and GS1 is affected in vitro by the same substrates and inhibitors ( Glaab and Kaiser, 1996; MacKintosh et al., 1995 ; Figure 4) and involves ATP hydrolysis ( Figure 5; Kaiser and Spill, 1991; MacKintosh et al., 1995 ). However, the activity of NR and GS1 are oppositely affected by the reversible phosphorylation as dephosphorylation activates NR ( MacKintosh et al., 1995 ) but deactivates GS1 ( Figures 6 and 7c). In addition, phospho-NR is an initial step in NR degradation whereas phospho-GS1 is more protected against degradation than dephospho-GS1 ( Figures 4–7; Botrel and Kaiser, 1997; Ortega et al., 1999 ). Thus, NR and GS1, both key enzymes involved in nitrogen assimilation in plants, are differently affected by reversible phosphorylation.

Our data show that chloroplastic GS is also a phospho-protein interacting with 14-3-3 proteins ( Figures 9 and 10). However, post-translational regulation by reversible phosphorylation during light/dark transitions could not be observed ( Figure 10). Possible explanations could be that the site of GS2 phosphorylation is of non-regulatory significance, as is the case for some phosphorylation sites of SPS ( Toroser et al., 1999 ), or that the phosphorylation of GS2 takes place in the cytosol prior to import into the chloroplasts where reversible phosphorylation presumably does not occur ( Glaab and Kaiser, 1996).

Phosphorylated GS interacts with 14-3-3 proteins

Cytosolic and chloroplastic GS are members of the phospho-protein group interacting with 14-3-3 proteins. This has been shown by different methods, e.g. interaction with digoxygenin-labelled 14-3-3 proteins ( Moorhead et al., 1999 ), interaction with His-tag 14-3-3 proteins or far-Western blotting overlaid with 14-3-3–GFP ( Figures 8a and 9). Phosphorylated enzymes can interact with 14-3-3 proteins and thereby modify their activity state, e.g. phospho-NR is converted into a low-activity state when it interacts with 14-3-3 proteins ( Bachmann et al., 1996a ; MacKintosh et al., 1995 ; Moorhead et al., 1996 ). Phospho-GS1 is converted into a state of higher activity when 14-3-3 proteins are present ( Figure 5, insert, Figures 6 and 7c). Possible 14-3-3 protein specificity was investigated using three different 14-3-3 isoforms from Arabidopsis and one from barley (data not shown), but we were not able to detect any 14-3-3 isoform specificity as seen for NR ( Bachmann et al., 1996b ).

The phosphorylation of GS1 delays its degradation ( Figures 6 and 7a), whereas the interaction between phospho-GS1 and 14-3-3 proteins increases the GS1 activity ( Figures 5–7). The role of GS1 and/or 14-3-3 proteins in regulating senescence has recently been indicated, in that senescence occurred earlier both in tobacco plants over-expressing an exogenous GS1 gene and in antisense potato plants with repressed 14-3-3 protein content ( Vincent et al., 1997 ; Wilczynski et al., 1998 ). Over-expression of a 14-3-3 protein in tomato delayed leaf senescence ( Markiewicz et al., 1996 ).

Reversible phosphorylation of GS1 in roots and leaves

One of the two gs isogenes encoding cytosolic GS, Bngsr2-1, is expressed in both roots and leaves of B. napus, whereas Bngsr1-1 is root specific ( Figure 1 and Finnemann and Schjoerring, 1999). GS1 activity in both roots and senescing leaves increased upon pre-incubation with 14-3-3 proteins (data not shown). Thus, post-translational regulation and interaction with 14-3-3 proteins take place both in roots and leaves of B. napus.

Post-translational regulation of GS has until now only been shown in B. napus (this paper) and in the closely related species B. oleracea ( Moorhead et al., 1999 ). We do not know whether reversible phosphorylation is only related to the B. oleracea genome, but this does not seem to be the case as many gs1 genes from other plant species, e.g. radish, Arabidopsis, rice, maize and pea ( Peterman and Goodman, 1991 ; Sakakibara et al., 1992 ; Sakamoto et al., 1989 ; Tingey et al., 1987 ; Watanabe et al., 1994 ) have 87–97% identity in the deduced amino acid sequence (no other gs isogenes in the same plant species have higher amino acid similarity) and have age-dependent expression patterns identical to Bngsr2-1 ( Figure 1).

Model for post-translational regulation of GS1

A model for post-translational regulation of GS1 is shown in Figure 11. In this model, conversion of GS1 to a phosphorylated form (GS1-P) requires the action of a protein kinase in the presence of both ATP and Mg2+ ( Figures 4–6). Compared to the dephosphorylated form (GS1), GS1-P has an optimal degree of protection against protein degradation ( Figures 6 and 7a). Highest activity is achieved when GS1-P interacts with 14-3-3 proteins ( Figures 5, 6 and 7c). Dephosphorylation is catalysed by a microcystin-sensitive protein phosphatase which dephosphorylates GS1, thereby releasing the 14-3-3 interaction ( Figures 7c and 8c). GS1 is more easily degraded than GS1-P ( Figure 7a), and its activity is unknown. Whether the 14-3-3 protein level remains constant during light/dark transitions is still a matter for discussion ( MacKintosh et al., 1995 ; Markiewiez et al., 1996 ; Weiner and Kaiser, 1999).

Figure 11.

A hypothetical model for the reversible control of cytosolic glutamine synthetase activity by phosphorylation and dephosphorylation.

The regulation of GS1 activity obtained by phosphorylation is light/dark-dependent ( Figure 10) and influenced by the ATP/AMP ratio in vitro as also observed for NR activity (data not shown; Kaiser and Spill, 1991). The ATP/AMP ratio changes in vivo during light/dark transitions ( Kaiser and Spill, 1991), and leads to changes in the phosphorylation status of NR under semi-in vivo conditions ( Kaiser et al., 1993 ; MacKintosh et al., 1995 ). In young photosynthetically active leaves, the requirement of ATP for biosynthesis lowers the ATP/AMP ratio. Decreased phosphorylation of GS1 was measured during the light period ( Figure 10), as also reported for NR ( Agüera et al., 1999 ). During darkness, the respiratory activity increases, thereby increasing the ATP/AMP ratio and both the amount of phospho-GS1 ( Figure 10) and phospho-NR ( Agüera et al., 1999 ; MacKintosh et al., 1995 ). While the rate of photosynthesis decreases during early senescence, respiration is maintained or even increased ( Krömer, 1995), and as a consequence the level of phospho-GS1 increases ( Figure 10).

The maintenance of subcellular compartmentation during this last phase of senescence is critical for the highly organized and controlled catabolic events, allowing remobilization of nitrogen and other phloem-mobile elements ( Table 1; Buchanan-Wollaston, 1997; Feller and Fischer, 1994). The predominant localization of GS1 in the phloem parenchyma ( Edwards et al., 1990 ; Kamachi et al., 1992 ) and the protection of GS1 against degradation by phosphorylation and increased activity by 14-3-3 protein interactions ( Figures 5 and 7b,c; Ortega et al., 1999 ) both point to a role for GS1 in nitrogen remobilization. The increased glutamine content in leaves and phloem during the final stages of senescence ( Table 1) further supports this function of GS1.

Experimental procedures

Plant material

Plants of B. napus (Brassica napus L. ssp. napus cv. global) were cultivated hydroponically as described previously ( Finnemann and Schjoerring, 1999). The plants were grown in growth chamber at 20°C with a photoperiod of 16 h per day. Immediately after collection of leaf material the samples were frozen in liquid nitrogen, homogenized using mortar and pestle, and stored at −80°C.

Isolation of RNA and expression analysis

RNA extraction from frozen leaf material was performed using the RNeasy plant mini kit (Qiagen GmbH, Hilden, Germany). The RNA quality was analysed by gel electrophoresis. To determine the expression level of the GS isogenes from B. napus, a method with multiplex RT–PCR was designed using the Titan one-tube RT–PCR system (Boehringer Mannheim GmbH, Mannheim, Germany). RT–PCR was carried out with the primers 5′-CAGCAC- AATCAACAGAGTTGAGAG-3′ (forward, positions 256–279 of the full-length Bngsl1 cDNA) and 5′-CAGGATACGCTCCAACGGGCC-3′ (reverse, positions 696–716 of the full-length Bngsl1 cDNA); 5′-GCAGAGACTACACTTCTTTGG-3′ (forward, positions 1109–1129 of the full-length Bngsr1-1 cDNA) and 5′-CATGAAAAG- GCCTTTTCAGATCGCC-3′ (reverse, positions 1349–1374 of the full-length Bngsr1-1 cDNA); 5′-GCTTACACTCCAGCGGGCGAACC-3′ (forward, positions 340–363 of full-length Bngsr2-1 cDNA) and 5′-GAGATGTTTTCAATGGATGTG-3′ (reverse, positions 1271–1290 of full-length Bngsr2-1 cDNA) with the following RT–PCR cycle parameters: once at 50°C for 30 min, 94°C for 2 min; 10 times at 94°C for 30 sec, 58°C for 30 sec, 68°C for 5 min; 25 times at 94°C for 30 sec, 58°C for 30 sec, 68°C for 5 min with 5 sec elongation of time per cycle; once at 68°C for 7 min. As an internal standard, the actin gene cloned from B. napus and was used with primers 5′-AAGAGCACCCGGTTCTTCTCAC-3′ (forward, positions 353–374 of full-length BnAc cDNA) and 5′-GTGCGACCACCT- TGATCTTCAT-3′ (reverse, positions 1051–1072 of full-length BnAc cDNA) amplifying an 699 bp fragment. The primers amplified transcripts related to both the A and C genome of the amphidiploid B. napus. Amplified transcripts were analysed on a 1.0% agarose gel and quantified by the Gel Doc 1000 system (Bio-Rad, Richmond, CA, USA) and the Storm 860 Scanner (Molecular Dynamics Inc., Sunnyvale, CA, USA). Different amounts of RNA extracted from both young and senescing leaves were used to optimize the multiplex RT–PCR procedure, resulting in typically sigmoidal curves (data not shown). However, linear curves could be obtained for all isogenes of interest by use of 10–300 ng total RNA, and hence 50 ng total RNA was used in the multiplex RT–PCR procedure. The relative variation between the replicates in the multiplex RT–PCR procedure and the subsequent fragment quantification was less than 4%. Northern blots and RT–PCR gave similar results (data not shown).

Protein extraction, enzymatic assay and Western blotting

Protein extraction and GS activity were carried out as described previously ( Finnemann and Schjoerring, 1999) with minor modifications: partial purification of GS was achieved by di-ammonium sulphate fractionation (40–50% precipitate used), resuspended in extraction buffer (70 m m MOPS pH 6.80, 10 m m MgSO4, 5 m m glutamate, 2 m m DTT, 0.1% Triton X-100, 10% glycerol), and desalting (HiTrap desalting columns, Pharmacia Biotech, Uppsala, Sweden). Pre-incubation experiments with GS extracts were performed in extraction buffer (containing 10 m m Mg2+) with the addition of stock solutions (made with MES pH 6.80) of ATP, ATP-γ-S, Mg2+, EDTA, microcystin LR or phosphorylated Raf1 protein (gift from A.T. Fuglsang, Department of Plant Biology, The Royal Veterinary and Agricultural University, Copenhagen, Denmark). Assay procedures for GS activity measurements were optimized so that pre-incubation effects were not related to non-optimal conditions. 32P-labelling of GS was performed with [γ-32P]ATP (1.11 TBq /mmol) and bands were visualized by a Storm 860 scanner (Molecular Dynamics Inc.). In experiments with the nickel–nitrilotriacetic acid (Ni-NTA) columns (Qiagen Inc., Valencia, CA, USA), an extraction buffer consisting of 50 m m MES (pH 6.70), 10 m m MgSO4, 2 m m mercaptoethanol, 0.1% Triton X-100 and 10% glycerol was used. Protein was quantified using the Bradford method ( Bradford, 1976), and Western blot analysis was carried out according to the manufacturer's protocol (Bio-Rad) using a polyclonal GS antibody ( Cullimore and Miflin, 1984). Visualization of GS was achieved with alkaline phosphatase linked to a goat anti-rabbit IgG (Bio-Rad), and quantification was performed using a Storm 860 scanner (Molecular Dynamics Inc.).

Phloem sap collection and amino acid determinations

Collection of phloem sap was performed according to the method of Lappartient and Touraine (1996). Petioles of leaves were excised in 20 m m sodium EDTA, pH 7.0, using a sharp blade. Immediately thereafter, the leaf petioles were washed in 5 m m sodium EDTA, pH 7.0, and immersed in 2.0 ml 5 m m sodium EDTA, pH 7.0, for 4 h in darkness and 95% relative humidity. The volume of phloem exudate collected could not be estimated, and the results were therefore normalized on the basis of the total amount of amino acids. Determinations of the amino acid content in the phloem exudates and leaf tissues were performed according to Finnemann and Schjoerring (1999).

Determination of NH4+ and chlorophyll in the leaf tissues

Soluble NH4+ in finely ground frozen leaf samples was extracted in 0.1 m H2SO4 for 1 h on ice. After filtration (0.45 µm PVDF microcentrifuge tube filter, Whatman, Maidstone, UK), sample pH was adjusted to 6.0. Leaf NH4+ content was measured by isocratic HPLC ( Finnemann and Schjoerring, 1999). Chlorophyll measurements were performed using the methanol extraction procedure described by Lichtenthaler and Wellburn (1983).

Interactions with 14-3-3 proteins

The 14-3-3 proteins (GF14-ω and 14-3-3–GFP, a gift from A.T. Fuglsang and T. Jahn, Department of Plant Biology, The Royal Veterinary and Agricultural University, Copenhagen, Denmark) were expressed using the expression vector pQE-9 (Qiagen Inc.). Due to an N-terminal Met-Arg-Gly-Ser-6×His tag (His-tag 14-3-3), purification by Ni-NTA columns was possible (Qiagen Inc.). The interaction between GS and His-tag 14-3-3 proteins was shown as follows. Leaf GS extract containing 500 µg total protein was incubated with 50 µg of His-tag 14-3-3 proteins for 1 h on ice. Subsequently, the His-tag 14-3-3:GS interaction was visualized by Ni-NTA column purification (Qiagen Inc.) followed by immunodetection of GS in the eluted fraction using Western blot with a specific GS antibody ( Cullimore and Miflin, 1984). For the far-Western overlay method (developed by T. Jahn, Department of Plant Biology, The Royal Veterinary and Agricultural University, Copenhagen, Denmark), a Western blot was pre-incubated in buffer (50 m m MES pH 6.50, 10 m m MgSO4) prior to incubation with 14-3-3 protein linked to a green florescent protein (14-3-3–GFP). After incubation and washing, visualization was performed using a Storm 860 scanner (Molecular Dynamics Inc.). Subsequently, GS subunits were shown on the same Western blot by immunodetection.


This work was supported by grants from the Danish Ministry of Food, Agriculture and Fisheries, research program ‘Traits for the Crop Plants of the Future’.

EMBL sequence database accession numbers X72751 (Bngsl1), X76736 (Bngsr1-1) X82997 (Bngsr2-1) and AF111812 (BnAc).