Glutamine Synthetase is a Molecular Target of Nitric Oxide in Root Nodules of Medicago truncatula and is regulated by Tyrosine Nitration

Nitric oxide (NO) is emerging as an important regulatory player in the Rhizobium legume symbiosis, but its biological role on nodule functioning is still far from being understood. To unravel the signal transduction cascade and ultimately NO function it is necessary to identify its molecular targets. This study provides evidence that glutamine synthetase (GS), a key enzyme for root nodule metabolism, is a molecular target of NO in root nodules of Medicago truncatula, being regulated by tyrosine nitration in relation to active nitrogen fixation. In vitro studies, using purified recombinant enzymes produced in E. coli, demonstrated that the M. truncatula nodule GS isoenzyme (MtGS1a) is subjected to NO-mediated inactivation through tyrosine nitration and identified tyrosine 167 as the regulatory nitration site crucial for enzyme inactivation. Using a sandwich enzyme-linked immunosorbent assay it is shown that GS is nitrated in planta and that its nitration status changes in relation to active nitrogen fixation. In ineffective nodules and in nodules fed with nitrate, two conditions in which nitrogen fixation is impaired and GS activity is reduced a significant increase in nodule GS nitration levels was observed. Furthermore, treatment of root nodules with the NO donor sodium nitroprusside (SNP) resulted in increased in vivo GS nitration accompanied by a reduction in GS activity. Our results support a role of NO in the regulation of nitrogen metabolism in root nodules and places GS as an important player in the process. We propose that the NO-mediated GS post-translational inactivation is related to metabolite channeling to boost the nodule antioxidant defenses in response to NO. Immunosorbent Assay (ELISA) was developed as a tool to compare the amount of in vivo nitrated GS in nodules collected under different physiological conditions. This assay uses an anti-GS antibody as a capture antibody to immuno-separate GS from the other nodule proteins and a specific anti-3-nitrotyrosine antibody for detection and quantification of the relative amount of nitrated GS in the different nodule extracts. The relative quantification of GS nitration was based on standards using pure recombinant MtGS1a, which was previously subjected to in vitro nitration by incubation with 100 μ M PN. Since the GS polypeptide content and GS activity change under different physiological conditions, the nodule protein extracts were also

6 species, and exclusively expressed in developing seeds (Seabra et al., 2010). The other three GS genes are expressed in root nodules, but MtGS1a is highly up regulated in the central infected cells, accounts for the production of over 90% of the total nodule GS activity and encodes the isoenzyme responsible for the assimilation of the ammonia released by nitrogen fixation (Carvalho et al., 2000a).
The essentially of GS for plant life predicts the existence of several precise and tightly controlled mechanisms for the regulation of enzyme activity. The regulation of GS at the transcriptional level has been well studied, and it is generally accepted as the primary regulatory point, but a number of post-translational regulatory controls were shown to be critical for the regulation of GS activity (Hoelzle et al., 1992;Temple et al., 1996Temple et al., , 1998Ortega et al., 1999Ortega et al., , 2001. In higher plants, GS activity is known to be modulated by oxidative turnover (Ortega et al., 1999, Palatnik et al., 1999, Ishida et al., 2002 and phosphorylation along with 14-3-3 interaction (Finnemann and Schjoerring, 2000;Man and Kaiser, 2001;Riedel et al., 2001;Lima et al., 2006aLima et al., , 2006b. In E. coli, the regulation of GS activity involves cyclic adenylylation and nitration of tyrosine residues (Berlet et al., 1996). Tyrosine nitration has also been implicated in the regulation of GS in animals (Gorg et al., 2007), being conceivably a similar mechanism for the regulation of the plant enzyme.
Protein tyrosine nitration is a nitric-oxide mediated post-translational modification which consists on the addition of a nitro (NO 2 ) group to one of the equivalent ortho carbons of the aromatic ring of tyrosine residues (Radi, 2004) that can alter the conformation and structure of proteins, the catalytic properties of enzymes and their susceptibility to proteolysis (Abello et al., 2009). Protein tyrosine nitration has been best studied in animal systems and little is known regarding the functional effects of protein tyrosine nitration in plants (Corpas et al., 2009), but recent proteomic studies revealed the existence of a high number of nitrated proteins in sunflower (Chaki et al., 2009) and in Arabidopsis (Cecconi et al., 2009;Lozano-Juste et al., 2011), suggesting that tyrosine nitration also represents an important NO-mediated regulatory mechanism in plants.
The formation of NO and its involvement in the legume-rhizobia symbiosis has been the subject of much research in the last few years and it is believed that it plays an important, but yet unknown, signaling role on the symbiosis (Wang and Ruby, 2011). The model legume M. truncatula and its symbiotic partner Sinorhizobium meliloti provided the background for many of those studies. NO production has been located at the infection sites during the initial stages of nodule development of M. truncatula and seems to be required for the establishment of the symbiosis (Del Giudice et al., 2011), whereas in fully 7 developed root nodules it appears to be confined to the nodule fixation zone, suggesting an involvement of NO in root nodule metabolism (Baudouin et al., 2006;Horchani et al., 2011). Proposed sources of NO in root nodules include NO synthase-like proteins (Cueto et al., 1996;Baudouin et al., 2006;Leach et al., 2010), nitrate reductase and electron transfer chains from both plant and bacteria (Kato et al., 2003;Mesa et al., 2004;Meakin et al. 2007;Horchani et al., 2011, Gupta et al., 2011. NO appears to directly affect nodule metabolism by inhibiting nitrogenase activity. The involvement of NO in nitrogenase inactivation has been demonstrated in soybean and Lotus after nitrate supply (Kanayama et al., 1990;Meakin et al., 2007;Kato et al., 2010). In Lotus japonicus, the artificial application of the NO donor sodium nitroprusside (SNP) decreased nitrogen fixation, whereas the application of a NO scavenger (cPTIO) enhanced nitrogen fixation (Shimoda et al., 2009;Kato et al., 2010). In spite of the inhibitory effect of NO on nitrogenase, it appears that NO production is required for the nodule development and functioning and the plant antioxidant systems appear to be crucial to maintain nodule functioning (Pauly et al., 2006;Keyster et al., 2010;Leach et al., 2010;Sanchez et al., 2011). The levels of NO inside the nodule appear to be controlled by leghemoglobin which is able to scavenge NO and in this way may protect nitrogenase from inactivation (Kanayama et al., 1990;Herold and Puppo, 2005;Meakin et al., 2007;Kato et al., 2010;Sanchez et al., 2011). Nitroso-leghemoglobin complexes (LbNO) have been detected in nodules of soybean and Lotus (Kanayama et al., 1990;Mathieu et al., 1998;Meakin et al., 2007;Sanchez et al., 2010). This NO-scavenging function has also been attributed to nonsymbiotic haemoglobins in Lotus japonicus, which are induced upon symbiotic infection and accumulate in nitrogen fixing nodules (Shimoda et al., 2009).
There is still much controversy and uncertainty around the possible roles of NO on the legume-rhizobia symbiosis and the identification of its molecular targets is a major asset to start dissecting its mechanisms of signaling and action. In this study we provide evidence that glutamine synthetase is a molecular target of NO in root nodules of M. truncatula and that its activity is post-translationally regulated by tyrosine nitration in relation to active nitrogen fixation. A model is proposed to integrate the post-translational regulation of GS by tyrosine nitration within the context of root nodule metabolism.

8
The major protein post-translational modifications induced by NO are nitrosylation of cysteines and nitration of tyrosine residues (Radi, 2004). To investigate whether nodule glutamine synthetase could be affected by any of these modifications, we started by evaluating in vitro the effect of two reactive nitrogen species producers on the M. truncatula nodule isoenzyme MtGS1a and the plastid located isoenzyme MtGS2a. The plant proteins were produced in E. coli with a His-tag, purified by Ni-affinity chromatography and used to evaluate GS activity following incubation with increasing concentrations of either peroxynitrite (PN) or tetranitromethane at pH 8 (TNM). GS activity was found to be inhibited by the two compounds in a dose dependent manner for the two isoenzymes ( Fig. 1), reaching total inactivation after incubation with 100 μM of either PN or TNM. To assess whether the inhibitory effect could be due to tyrosine nitration, the treated enzymes were analyzed by western blotting using a specific anti-3-nitrotyrosine antibody. Immunoreactivity was found to increase with increasing PN or TNM concentration for both isoenzymes (Fig.1).
To further investigate whether the cause of the NO-induced GS inactivation is nitration or other oxidative processes triggered by the reactive nitrogen species, we tested the effect of a number of reagents known to selectively modify tyrosine, cysteine or methionine residues, on the activity of each GS isoenzyme. The widely used tyrosine O-acetylation agent N-acetyl-imidazole was used to specifically modify tyrosine residues, the alkylating sulphydryl reagent iodocetamide and the S-nitrosylation reagent S-nitrosoglutathione were used as cysteine modification agents and hydrogen peroxide as a potent methionine and cysteine oxidant. The activity of MtGS1a was not significantly affected by cysteine alkylation ( Fig. 2A) nitrosylation ( Fig. 2B) or oxidation (Fig. 2C). Conversely, the acetylation of tyrosine residues by N-acetyl-imidazole clearly induced MtGS1a inactivation (Fig. 2D). These results indicate the existence of tyrosine residue(s) which upon modification lead to MtGS1a inactivation. Interestingly, this effect appears to be specific for the cytosolic isoenzyme MtGS1a, treatment of the plastid located MtGS2a isoenzyme with cysteine modifying reagents results in enzyme inactivation ( Fig. 2A,B,C), whereas N-acetyl-imidazole has a minor inhibitory effect on this isoenzyme (Fig. 2D).
Additional evidence that MtGS1a inhibition is a result of tyrosine modifications is given by the finding that cysteine alkylation, induced by incubation of the enzyme with iodoacetamide, does not prevent MtGS1a from being inactivated by either PN or TMN 9 (Fig. 3). Taken together, these results suggest that NO-dependent inactivation of MtGS1a is due to tyrosine nitration whereas MtGS2a activity is inhibited by thiol residue nitrosylation.
The flavonoid epicatechin has been shown to selectively prevent nitration of tyrosine residues (Schroeder et al., 2001), we therefore examined whether this flavonoid could protect the enzymes from inactivation. Both MtGS1a and MtGS2a, were incubated with increasing concentrations of epicatechin before treatment with 50μM peroxynitrite (Fig. 4).
Epicatechin was able to protect MtGS1a from inactivation (Fig. 4A), but not MtGS2 (Fig.   4B). Furthermore, the incubation of both GS isoenzymes with epicatechin, prior to peroxynitrite treatment resulted in decreased immunoreactivity against nitrotyrosine, confirming that this flavonoid does indeed prevent tyrosine nitration. These results substantiate that MtGS1a inactivation is a consequence of tyrosine nitration mediated by reactive nitrogen species. As for MtGS2a, peroxynitrite and TNM induced tyrosine nitration but this modification does not seem to be the cause of enzyme inactivation as epicatechin was able to prevent nitration of the protein, but not the inactivation induced by the reactive nitrogen species producers (Fig. 4B).
Tyrosine nitration is generally considered an irreversible process whereas S-nitrosylation is rapidly reversible (Besson-Bard et al., 2007). In order to examine whether the NOinduced GS inactivation could be reversed, we incubated the NO-treated enzymes with DTT and looked for activity recovery. Following incubation with DTT, the activity of MtGS2a could be totally recovered, while MtGS1a inactivation could not be reversed (Fig.   5), further strengthening the idea that the MtGS1a NO-inhibitory effect is due to tyrosine nitration whereas MtGS2 inactivation is likely to result from cysteine nitrosylation.

Identification of Tyrosine 167 as a Relevant Nitration Site
Once MtGS1a and MtGS2a appear to be differentially regulated by NO-mediated modifications, either through tyrosine nitration or cysteine modifications, respectively, we searched for tyrosine residues present in MtGS1a but not in MtGS2a. An alignment of the aminoacid sequences of the two proteins identified three tyrosine residues on MtGS1a, Tyr108, Tyr167 and Tyr263, which are substituted in MtGS2a (Fig. 6). From the analysis of the three dimensional structure of MtGS1a (Seabra et al., 2009), it is predicted that only the residues Tyr167 and Tyr263 are placed within a local environment favorable to nitration (Abello et al., 2009) and in addition their modification is likely to affect GS activity. To investigate whether any of these residues could be a regulatory nitration site www.plantphysiol.org on August 20, 2017 -Published by Downloaded from Copyright © 2011 American Society of Plant Biologists. All rights reserved. and account for the specific susceptibility of MtGS1a to inactivation by nitration, we mutated residues Tyr167 and Tyr263 to phenylalanine by site-directed mutagenesis and tested the susceptibility of the mutated proteins to inactivation induced by TNM. The mutated recombinant proteins were found to be active and their specific activity was similar to the non-mutated enzymes. However, the inhibitory effect of TNM was significantly reduced for the Tyr167 mutated protein (MtGS1aY167F) whereas mutation of Tyr263 does not appear to affect TNM-induced GS inhibition, with the protein showing a pattern of inhibition comparable with the non mutated enzyme (Fig. 7A). These results demonstrate that Tyr167 is a relevant site for the inhibition of MtGS1a activity by nitration. However, we cannot exclude the existence of additional tyrosine residues important for the inactivation process or the contribution of other oxidative modifications induced by NO, as the mutated protein was still susceptible to inhibition by TNM, although to a much lower extent (20% inactivation). Pre-treatment of the mutated enzyme with 20 μM epichatechin could not prevent this inactivation (Fig. 7B), suggesting that Tyr167 is the only tyrosine implicated in the regulation of MtGS1a activity by nitration and that the observed 20% reduction in GS activity results from oxidative modifications different from tyrosine nitration. Western blot analysis of the mutated proteins treated with TNM (Fig. 7C) showed a reduction in nitrotyrosine imunoreactivity of the two mutated enzymes (Tyr167 and Tyr263) relative to the non-mutated enzyme, indicating that the two residues are subjected to nitration, but only Tyr167 appears to be critical for GS inactivation.

In vivo GS nitration in Effective versus Ineffective Root Nodules
To begin studying the physiological significance of GS tyrosine nitration for the root nodule functioning, we compared GS nitration in planta in situations where nitrogen fixation is impaired. For this purpose, a Sandwich Enzyme-Linked Immunosorbent Assay (ELISA) was developed as a tool to compare the amount of in vivo nitrated GS in nodules collected under different physiological conditions. This assay uses an anti-GS antibody as a capture antibody to immuno-separate GS from the other nodule proteins and a specific anti-3-nitrotyrosine antibody for detection and quantification of the relative amount of nitrated GS in the different nodule extracts. The relative quantification of GS nitration was based on standards using pure recombinant MtGS1a, which was previously subjected to in vitro nitration by incubation with 100 μM PN. Since the GS polypeptide content and GS activity change under different physiological conditions, the nodule protein extracts were also subjected to GS activity measurements and to immunoblot analysis using anti-GS antibodies. The cytosolic GS immunoreactive bands were quantified by densitometry and the values were plotted. The analyses were performed in at least four biological replicates and for each biological sample GS nitration, GS quantity and GS activity were compared.
In ineffective root nodules formed by the rhizobium mutant FixJ -GS activity is significantly reduced (Fig. 8A), which is an expected result since the Rhizobium mutant is unable to produce ammonium for GS assimilation, yet the GS polypeptide content is only slightly reduced (Fig. 8B). Interestingly, the amount of nitrated GS was found to be significantly higher in the mutant nodules when compared to wt nodules (Fig. 8C). Thus, the reduction on GS activity in ineffective root nodules can be related with an increase in in vivo GS nitration strongly suggesting a direct relationship between GS nitration and nitrogen fixation.

In vivo GS Nitration in Roots and Root Nodules Following Nitrate Application
Nitrate is known to induce NO production inside the nodule leading to the inhibition of nitrogenase activity (Kanayama et al., 1990;Meakin et al., 2007;Kato et al., 2010). To evaluate the effect of nitrate application on GS nitration in roots and nodules of M. truncatula, nodulated plants (19 DAI) were fed with nitrate for two days, and the levels of GS nitration were compared in nitrate treated and non-treated plants, and both in roots and root nodules. Following nitrate application GS activity was found to be significantly reduced in root nodules (Fig. 9A), and this reduction was accompanied by a decrease in GS polypeptide content (Fig. 9B). The amount of cytosolic GS polypeptides (39 kDa) was 50% reduced and the plastid located polypeptides (42 kDa) were reduced to undetectable levels.
Interestingly, in spite of the observed reduction in GS abundance the amount of nitrated GS was found to be significantly increased in the nitrate treated nodules (Fig. 9C), further suggesting a relationship between GS nitration and active nitrogen fixation.
In roots, nitrate induced an increase in root GS activity (Fig. 9A), as well as in the amount of cytosolic and plastid located polypeptides (Fig. 9B), reflecting the fact that the roots start to assimilate nitrate. In opposition to what was observed in the nodules, the amount of nitrated GS in the root, which is mainly composed of the isoenzyme MtGS1b (Carvalho et al., 2000b) does not seem to be affected by the nitrate treatment (Fig. 9C). In spite of the higher production of nitric oxide, which is likely to occur following nitrate www.plantphysiol.org on August 20, 2017 -Published by Downloaded from Copyright © 2011 American Society of Plant Biologists. All rights reserved.
feeding, the root enzyme does not appear to be regulated by nitration. This result supports a specific regulation of the nodule isoenzyme MtGS1a by tyrosine nitration.

In vivo GS nitration in Root Nodules treated with SNP
To further evaluate whether GS nitration is the cause of GS inactivation in vivo, we applied the widely used NO-donor sodium nitroprusside (SNP) to root nodules of Medicago truncatula and followed GS activity, GS polypeptide content and GS nitration in nodules treated with 1 mM and 10 mM SNP. The GS polypeptide content was not affected by the SNP treatment ( Fig. 10B) but GS activity was found to be reduced (Fig. 10A). The reduction in GS activity can be directly correlated with an increase in the levels of nitrated GS evaluated by ELISA (Fig. 10C).

DISCUSSION
There is increasing evidence for a role of NO in the legume-rhizobia symbiosis, but the nature of this role is still far from being understood. Elucidation of mechanisms of action of NO in root nodules will inevitably require the identification of its molecular targets. This study represents a first step in that direction as it provides evidence that the key metabolic enzyme glutamine synthetase (GS) is a target of NO in root nodules of M. truncatula, being post-translationally regulated by tyrosine nitration in relation to active nitrogen fixation.
Glutamine synthetase is abundantly present in root nodules and plays a pivotal role in the assimilation of the ammonium released by nitrogen fixation. We have previously shown that in M. truncatula this function is fulfilled by the cytosolic isoenzyme MtGS1a, which is abundantly present in the infected cells of the nodule fixation zone (Carvalho et al., 2000a), coinciding with the major site of NO formation in this model species (Baudouin et al., 2006;Horchani et al., 2011). The enzyme is thus accessible to the oxidative effects induced by this reactive compound. NO can affect protein function and in this way may signal fundamental physiological processes, with S-nitrosylation of cysteine residues and tyrosine nitration being the major protein post-translational modifications induced by NO (Besson- Bard et al., 2007). We therefore, started by evaluating whether the M. truncatula nodule enzyme MtGS1a could be affected by either of these modifications. In vitro studies using purified recombinant proteins demonstrated that the nodule enzyme MtGS1a is subjected to tyrosine nitration and that this modification provokes a loss of enzyme activity. To our knowledge this is the first report on GS inactivation by tyrosine nitration in plants, but 13 previous studies reported an in vitro inhibition of the mammalian and bacterial GS activity by tyrosine nitration (Berlet et al.;1996;Gorg et al., 2007). In Arabidopsis, the plastid located GS (GS2) has been previously identified both as an S-nitrosylated protein, (Lindermay et al., 2005) and as a nitrated protein in more recent proteomic studies (Cecconi et al., 2009;Lozano-Juste et al., 2011), but the effect of either of these modifications on the activity of the enzyme was not assessed before. The present study demonstrates that the activity of the cytosolic enzyme is modulated by tyrosine nitration and suggests that the plastid located GS is susceptible to both nitration and S-nitrosylation, but the cause of GS2 inactivation appears to be S-nitrosylation. However, we cannot rule out the possibility that GS2 inactivation could result from other cysteine oxidative processes triggered by reactive nitrogen species. The finding that two isoenzymes sharing a high degree of sequence homology and a remarkably conserved active site fold are differentially modified by NO strengthens the idea that the NO signaling effects are specific under different physiological contexts. In addition to a differential susceptibility of individual GS isoenzymes to NO, the differential localization of the isoenzymes in a specific plant tissue is likely to be implied in the NO-mediated regulation of GS activity.
Tyrosine nitration is considered a selective process, and typically only one or two of the tyrosine residues present in a protein become preferentially nitrated, depending on the 14 fact that a mutation of this residue to phenylalanine results in a significantly reduced NOmediated inhibitory effect. These results strongly suggest that Tyr167 is the only tyrosine implicated in the modulation of MtGS1a activity by nitration, but Tyr263 constitutes an additional nitration target without apparent functional significance as mutation of this residue to phenylalanine does not affect the susceptibility of the enzyme to NO, in spite of its reduced nitrotyrosine imunoreactivity in response to NO treatments. The nitration sites have been appointed for some plant proteins in a recent proteomic approach, but the functional significance of nitration at a particular residue was not demonstrated (Lozano-Juste et al., 2011). Here, by site directed mutagenesis we present unequivocal evidence that Tyr167 is the regulatory nitration site responsible for enzyme inactivation. To our knowledge this is the first report which unambiguously identifies a nitration site in a plant enzyme and determines its functional significance.
Contrary to S-nitrosylation, tyrosine nitration is generally considered an irreversible process, however it has been reported that the inhibition of mammalian GS by tyrosine nitration could be reversed by a putative denitrase existent in spleen extracts from lipopolissacaride (LPS) treated mice (Gorg et al., 2007). We tested whether the NOinduced inactivation of MtGS1a could be reversed by incubation of the in vitro nitrated protein with protein extracts from different organs of the plant and even from spleen extracts from LPS treated mouse (data not shown). All our attempts failed to recover MtGS1a activity and thus the inhibitory effect of NO on MtGS1a activity does not appear to be reversible. Conversely, the NO-mediated inactivation of MtGS2a could be readily reversed by incubation with reducing agents, further suggesting that the plastid located enzyme is inactivated by S-nitrosylation.
Due to the interest in the physiological significance of GS nitration for the root nodule functioning, we investigated whether GS was nitrated in vivo in root nodules and whether the levels of GS nitration could be related to lower levels of GS activity in vivo. Central to these studies was the development of a sandwich ELISA to quantify the amount of in vivo nitrated GS in root nodules under different physiological conditions. Using this method, the levels of in vivo GS nitration were compared in two situations where nitrogen fixation is impaired, GS activity is known to change and NO is expected to be produced namely, in nodules fed with nitrate and in effective versus ineffective nodules. Interestingly, a direct relationship could be established between reduced nodule GS activity, increased GS nitration and reduced nitrogen fixation activity, strongly suggesting that GS is posttranslationally inactivated by NO-mediated nitration in response to lower nitrogen fixation rates. Further evidence for the effect of in vivo GS nitration on the reduction of GS activity is given by the finding that nodules treated with the NO-donor SNP show reduced GS activity, which can be directly correlated with increased GS nitration. Since the GS polypeptide content was not altered by the SNP treatment, the reduction in GS activity can be directly attributed to the increased GS nitration levels. Given that the ammonium released by nitrogen fixation is assimilated in the cytosol of the infected cells by glutamine synthetase, it seems reasonable that the enzyme activity is modulated in response to the cell requirements to shut down ammonia assimilation, if it is not being produced.
It is noteworthy that the NO-induced inactivation of GS by nitration appears to be somehow specific in the root nodules. Following nitrate application, the root enzyme behaved in an opposite manner in relation to the nodule enzyme. Root GS activity increased following nitrate treatment, a process accompanied by de novo synthesis of the enzyme. However, in spite of the elevated GS polypeptide content there were no detectable changes in the amount of nitrated GS in the root, further suggesting a specific regulation of the nodule enzyme by NO. As referred above, in M. truncatula roots, GS is mainly composed of a different cytosolic isoenzyme MtGS1b which is mainly located in the root cortex, whereas MtGS1a is confined to the root vascular tissues (Carvalho et al., 2000b). It is likely that both the formation of NO at the sites of expression of each individual isoenzyme and the differential susceptibility of the two isoenzymes to NO, account for the differential regulation of GS in roots and root nodules.
All together the results presented in this work indicate that GS is post-translationally regulated by tyrosine nitration in root nodules of M. truncatula in relation to nitrogen fixation. A model for the regulation of the nodule enzyme is proposed in Fig. 11. According to this model, the inhibition of GS activity by tyrosine nitration would be directly related to NO-induced nitrogenase inhibition. In view of the fact that elevated levels of NO in root nodules lead to decreased production of ammonium for GS assimilation, the enzyme would be shut down by post-translational inactivation through tyrosine nitration in response to the signal NO, the same signal that shuts down nitrogenase. On the other hand, it is proposed that GS inhibition is involved in the nodule antioxidant response to NO. Glutamate, a substrate for GS activity is also the precursor for the synthesis of glutathione (GSH), which is known to be highly abundant in root nodules of several plant species and to play a major role in the antioxidant defense participating in the ascorbate/GSH cycle (Matamoros et al., 1999;Matamoros et al., 2003;Becana et al., 2010) deleterious effects of reactive nitrogen species. Consistent with this model, the synthesis of the two enzymes involved in GSH production from glutamate, gamma-glutamylcysteine synthetase (γ-ECS) and glutathione synthetase (GSHS) were found to be up regulated by NO in root nodules (Matamoros et al., 2003;Innocenti et al., 2007) In conclusion, this study identifies GS as a target of NO in root nodules of M. truncatula and shows that the enzyme is in vivo post-translationally regulated by tyrosine nitration in relation to nitrogen fixation. These findings are significant as the identification of the molecular targets of NO is determinant for the assembly of the signal transduction cascade that will ultimately help to unravel NO function. The present study also highlighted subtle differences between the regulatory effects of NO on different GS isoenzymes and under different physiological contexts, corroborating the idea that NO signaling is specific in determined physiological backgrounds. This specificity is probably governed by the compartmentalization and dynamics of NO production as well as the differential regulation of individual GS isoenzymes by NO. Particularly interesting is the clear relationship that was observed between reduced nitrogen fixation and inactivation of nodule GS by tyrosine nitration. This NO-mediated post-translational inactivation, which is possibly related to metabolite channeling to the cell antioxidant defenses, provides a link between NO signaling and nitrogen metabolism in root nodules.

Protein Extraction from Plant Tissues
Roots or root nodules of Medicago truncatula were homogenized at 4 o C in a mortar and pestle with two volumes of extraction buffer (PBS, pH 7.4 containing 0.05% Triton X-100 and 1mM PMSF) and the homogenates centrifuged at 13 000 g for 20 min, at 4 o C.

Expression and Purification of Recombinant GS in E. coli
The Medicago truncatula GS isoenzymes MtGS1a and MtGS2a were produced in  GS activity was determined using the transferase assay (Cullimore and Sims, 1980). One unit of transferase activity is equivalent to 1 μmol min -1 γ-glutamyl hydroxamate produced at 30 o C. Activity data were expressed as the mean ± SD of at least three independent experiments, with triplicate technical assays for each experiment. The specific activity of the E.coli produced purified GS isoenzymes is 144 μmol min -1 mg prot -1 for MtGS1a and 40 μmol min -1 mg prot -1 for MtGS2, respectively.

In vitro GS Nitration
Purified GS (2.5 μg) was diluted in 0.02 M potassium phosphate buffer and incubated with either peroxynitrite or tetranitromethane at concentrations 0-100 μM in a final volume of 100 μL. Treatment with PN (diluted in 0.001 N NaOH) was performed for 10 minutes at room temperature at pH 7.4. Control assays were performed by incubating the enzyme with 50 μM decomposed PN, for this purpose PN was added to the buffer 10 min before addition of the enzyme. The addition of PN to the buffer system did not lead to a pH change.
Treatments with TNM (diluted in ethanol) were performed for 1 hour at room temperature, at pH 8. Control assays were performed by adding equivalent volume of ethanol. To evaluate the protective effect of epicatechin on GS nitration, the purified enzymes were incubated for 30 minutes at 30ºC with epicatechin at concentrations 0-30 μM, prior to the addition of the nitrating agents. GS activity was assessed in the absence of nitrating reagents to evaluate the effect of epicatechin on the enzyme activity. Aliquots were taken for determination of enzyme activity and in some cases for western blot analysis.

Reaction of GS with S-Nitrosoglutathione, Hydrogen Peroxide, Iodoacetamide and N-Acetyl-Imidazole
We assessed the effects of several chemical reagents known to modify cysteine (Snitrosoglutathione, hydrogen peroxide and iodoacetamide) and tyrosine residues (N-acetylimidazole). To this end, 2.5 μg of purified GS, diluted in 0.02 M potassium phosphate buffer (pH 7.4), was incubated with various concentrations of each chemical reagent for 30 minutes at 30ºC. GS activity was assayed as described above. To assess the contribution of thiol oxidation to the peroxynitrite and TNM -induced GS inactivation purified His-tag with 50 μM of each of the NO donor reagents.

Protein Determination, SDS-PAGE and Western Blot Analysis
Soluble protein concentration was measured by the Coomassie dye-binding assay  with a anti-3-nitrotyrosine specific antibody (NO 2 -Tyr). Loading controls were obtained by immunoblotting the membranes with anti-GS1 or anti-GS2 antibody (GS1 and GS2), respectively. GS activity was normalized to that found in the absence of NO donor and is represented as mean value ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively. imidazole (D) and assayed for GS transferase activity. GS activity was normalized to that found in the absence of reagents and is represented as mean value ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively.

Figure 3. Effect of iodoacetamide on MtGS1a inactivation induced by TMN and PN.
Purified MtGS1a was treated with 500 μM iodoacetamide, followed by incubation with 50 μM peroxynitrite or 50 μM TNM at pH 8 and assayed for GS activity. GS activity was normalized to that found in absence of iodoacetamide and nitrating agent. GS activity is represented as mean values ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 . 29 and nitrating agent and is represented as mean values ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively. was additionally treated with 500 μM GSNO, followed by incubation with 10 mM DTT and assayed for GS activity. GS activity was normalized to that found in absence of reagents and is represented as mean values ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively.  The effect of epicatechin in the inactivation of mutated GS1aY167F was assessed by incubation of the mutated and non-mutated enzymes with 20 μM epicatechin followed by treatment with 50 μM TNM, pH 8 (B). GS activity was normalized to that found in absence of TNM and epicatechin and is represented as mean value ± SD of at least three independent experiments, assayed in triplicate. Overall nitration of mutated and nonmutated TNM treated enzymes was evaluated by western blots immunodetected with antinitrotyrosine antibody (NO 2 -Tyr). Loading controls were obtained by immunoblotting the membranes with anti-GS antibody (GS) (C). The average initial GS activity was 0.36 μmol min -1 μg prot -1 .     increasing concentrations, assayed for GS transferase activity and immunodetected with a anti-3-nitrotyrosine specific antibody (NO 2 -Tyr). Loading controls were obtained by immunoblotting the membranes with anti-GS1 or anti-GS2 antibody (GS1 and GS2), respectively. GS activity was normalized to that found in the absence of NO donor and is represented as mean value ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively.     and MtGS2a (B) were incubated with different concentrations of epicatechin, followed by treatment with 50 μM peroxynitrite, assayed for GS transferase activity and immunodetected with anti-nitrotyrosine antibody (NO 2 -Tyr). Loading controls were obtained by immunoblotting the membranes with anti-GS1 and anti-GS2 antibody (GS1 and GS2), respectively. GS activity was normalized to that found in absence of epicatechin and nitrating agent and is represented as mean values ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively.   Purified MtGS1a or MtGS2a were treated with 50 μM PN, and MtGS2a was additionally treated with 500 μM GSNO, followed by incubation with 10 mM DTT and assayed for GS activity. GS activity was normalized to that found in absence of reagents and is represented as mean values ± SD of at least three independent experiments, assayed in triplicate. The average initial GS activity was 0.36 μmol min -1 μg prot -1 and 0.1 μmol min -1 μg prot -1 for MtGS1a and MtGS2a, respectively.   Purified GS1a and mutated GS1aY167F and GS1aY263F were incubated with TNM, pH 8, at increasing concentrations and assayed for GS transferase activity (A). The effect of epicatechin in the inactivation of mutated GS1aY167F was assessed by incubation of the mutated and non-mutated enzymes with 20 μM epicatechin followed by treatment with 50 μM TNM, pH 8 (B). GS activity was normalized to that found in absence of TNM and epicatechin and is represented as mean value ± SD of at least three independent experiments, assayed in triplicate. Overall nitration of mutated and nonmutated TNM treated enzymes was evaluated by western blots immunodetected with antinitrotyrosine antibody (NO 2 -Tyr). Loading controls were obtained by immunoblotting the membranes with anti-GS antibody (GS) (C). The average initial GS activity was 0.36 μmol min -1 μg prot -1 .