Oversensitivity of Arabidopsis gad 1 / 2 mutant to NaCl treatment reveals the importance of GABA in salt stress responses

Salt stress is one of the major problems in agricultural fields. Currently, more than 20% of irrigated agricultural lands are affected by salinity. High concentrations of sodium affects plant growth by competing with the uptake of important ions like potassium (K + ), and posing osmotic stress. Some plant species developed mechanisms such as modifying cellular metabolism to minimize effects of high salt concentrations. Gamma-aminobutyric acid (GABA) accumulation during salt stress is one of the results of modifications in cellular metabolism. However, whether this response is specific or not has not been shown before. Here, it was hypothesized that GABA accumulation is needed to counter the effects of salt stress. For that, GABA-depleted Arabidopsis gad1/2 mutant was investigated for altered response under salt stress. Indeed, the double mutant was oversensitive to 150 mM NaCl treatment. Furthermore, the mutant was oversensitive to osmotic stress; since the double mutant showed reduced shoot water content after 300 mM mannitol treatment. Comparison of metabolites between salt-treated wild type and gad1/2 mutant showed that GABA shunt plays a central role in modulating the carbon and nitrogen metabolism. Taken together, the findings show that GABA accumulation under salt stress conditions plays an important role to overcome the high salt concentration damage.


INTRODUCTION
Salinity is one of the several abiotic stresses that pose threat to the productivity of agricultural lands.Previous reports show that a quarter of irrigated agricultural lands worldwide are severely affected by salinity, and a further 1.5 million hectares of lands are abandoned every year because of salinity (Munns and Tester, 2008).There E-mail: derejeworku1@yahoo.com.Tel: +4915218927122.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License are several reasons for salinity problems in soils; however, irrigation water combined with poor drainage stand out as the most serious in damaging agricultural fields.High sodium level in soils affects plant growth in two ways; first, it increases the osmotic potential of the soils thereby reducing the uptake of water into roots (osmotic stress); second, it competes with the influx of important ions like K + , inhibiting many physiological and biochemical processes such as nutrient uptake and assimilation (ionic stress) (Munns and Tester, 2008;Hasegawa et al., 2000).To counter the effects of salinity-induced cellular damages, plants have developed mechanisms such as osmotic tolerance and Na + exclusion (Munns and Tester, 2008).These tolerance mechanisms involve the interplay of several genes engaged in ion transport, detoxification process and metabolite production (Rus et al., 2001;Apse et al., 1999;Hayashi et al., 1997;Szekely et al., 2008).
Gamma-aminobutyric acid (GABA) is one of the metabolites that accumulate in response to abiotic stresses (Kinnersley and Turano, 2000).It is produced from the decarboxylation of glutamate by glutamate decarboxylase (GAD) within the cytosol and catabolized within the mitochondrial matrix by the activity of GABA-T and SSADH (Shelp et al., 2012).Despite the presence of five copies of GAD genes in Arabidopsis genome, two paralogs, GAD1 and GAD2, are responsible for more than 90% of GABA produced in shoots and roots under normal growth conditions.Arabidopsis double mutants impaired in GAD1 and GAD2 functions contained low GABA in shoots and roots (Scholz et al., 2015;Mekonnen et al., 2016).However, the induction of GAD4 transcripts under hypoxia (Miyashita and Good, 2008), drought (Urano et al., 2009) and cold (Kaplan et al., 2007) treatments suggest some specific functions of the GAD4 enzyme under stress conditions.Renault et al. (2010) also showed the induction of GAD4 transcript under salt stress in a dose dependant manner and suggested the specificity of the response.However, whether this transcriptional induction of GAD4 is important for GABA accumulation and salt stress tolerance of plants is still not clear.
GABA accumulation in response to salt stress has been reported in Arabidopsis (Renault et al., 2010), alfa alfa (Fougere et al., 1991), and tobacco (Binzel et al., 1987) plants.In Arabidopsis, GABA accumulation involves the up-regulation of almost all GABA-shunt genes at transcriptional level.Furthermore, the in vitro activity of GAD and GABA-T proteins were enhanced in response to salt stress (Renault et al. 2010).Despite the rapid accumulation of GABA under salt stress, little is known about its specificity and function under such conditions.Renault et al. (2010) reported the specificity of the GABA shunt response under salt stress since, the pop2-1 mutant impaired in GABA catabolism was oversensitive to ionic stress.The authors concluded that the Mekonnen 253 metabolism of GABA rather than GABA accumulation is needed to counter the ionic component of the salt stress.This provokes a question regarding the importance of GABA accumulation during salt stress.A recent report showed the importance of GABA accumulation than it's metabolism in drought stress response.GABA-depleted Arabidopsis gad1/2 mutant exhibited oversensitive phenotype following drought stress treatment.Functional complementation with a third mutation in pop2 gene specifically increased the GABA levels in shoots and reversed the drought oversensitive phenotype of the double mutant (Mekonnen et al., 2016).The pop2 mutation abolishes the GABA transaminase activity and leads to the accumulation of GABA.
The present work was conducted with the main objective of determining the importance of GABA accumulation under salt stress; first, by phenotypic characterization of GABA-depleted mutants under salt stress; second, by analyzing expression of major salt stress responsive genes; third, by analyzing the changes in C:N metabolism of mutant plants.

Salt stress treatment
Seeds of wild type (Col-0) and gad1/2 mutant (in Col-0 background) were sown on soil, cold treated and germinated in the greenhouse.Then, individual seedlings were transplanted to small pots filled with soil mix and kept in the greenhouse.The pots containing individual plants were arranged in a completely randomized design.Two weeks after transplanting, the salt stress treatment was initiated with 150 mM NaCl.The treatment was given by immersing the perforated tray containing pots in another tray filled with the salt solution.
Similarly, for control treatment plants on one of the trays were given water throughout the experiment.The treatment was given every third day for two weeks.Each treatment was replicated five times per genotype.The experiment was repeated once.

Osmotic stress treatment
Seeds from wild type and gad1/2 mutant were sown and germinated in the green house as described previously.The seedlings were transplanted to pots filled with soil mix and transferred to growth chamber (16/8 light/dark cycle, ~70% RH).For osmotic stress treatment, four-week-old plants were watered without and with 300 mM mannitol.The experiment was arranged in a completely randomized design as described previously.For quantification of the shoot water content, shoot samples were collected beginning one day after the onset of the stress treatment.Determination of the shoot water content was carried out as described previously (Mekonnen et al., 2016).Each treatment was replicated five times per genotype.

RNA extraction, cDNA synthesis and qRT-PCR
Leaf materials were collected from control and 150 mM NaCl treated four-week-old wild type and gad1/2 plants, and snapfrozen in liquid nitrogen.RNA extraction was carried out as described previously (Scholz et al., 2015).RNase treatment, cDNA synthesis and qPCR were performed as described previously (Mekonnen et al., 2016).Primer pairs that specifically detect GAD3 and GAD4 transcripts were used (Renault et al., 2010).The primers used for specific amplification of ABA2, ABA3, AAO3, AKT1, GORK, KAT1 and HAK5 transcripts are listed in the appendix.

GC-MS and HPLC analysis of metabolites
GC-MS: Leaf materials (~100-200 mg) were collected from fourweek-old wild-type and gad1/2 plants treated without or with 150 mM NaCl, and immediately frozen in liquid nitrogen.Metabolite extraction and measurement was carried out as described previously (Scholz et al., 2015).
HPLC: Shoot and root samples were collected from four-week-old wild-type and gad1/2 plants treated without or with 150 mM NaCl and immediately frozen in liquid nitrogen.For root sample collections, the entire plant was taken from the pot and soil was immediately removed by immersing the below soil part of the plant in water.Then, roots were briefly dried by placing them between tissue papers.The extraction and measurement of amino acids were performed as described previously (Mekonnen et al., 2016).

ICP MS measurement of Na + and K +
At least 500 mg of fresh shoot materials were harvested from fiveweek-old wild-type and gad1/2 Arabidopsis plants treated without and with 150 mM NaCl, and freeze dried in liquid nitrogen.The extraction and measurement of Na + and K + was carried out as described previously (Mekonnen et al., 2016).

Experimental design and statistical analysis
For phenotypic, chemotypic and expression analysis the experiments were arranged in a completely randomized design (CRD).For quantitative data, the statistical significance between treatment means was assessed using student's t-test.

Analysis of GAD transcripts and GABA accumulation
To determine the compensatory expression of GAD3 and GAD4 under salt stress, transcript abundances were investigated in shoots and roots of the gad1/2 mutant.The salt stress treatment increased the abundance of GAD4 transcripts in shoots of wild type plants, although it is not statistically significant (Figure 1A).In contrast, the abundance of GAD4 transcript was unchanged in shoots of gad1/2 mutant (Figure 1A).GAD3 transcript was not detected in shoots and roots of both genotypes under normal growth conditions and was not induced either by salt stress treatment (Figure 1A and B).In roots of gad1/2 mutant, the GAD4 transcript was induced by nearly four-fold following the salt treatment.To determine if the induction of GAD4 expression under salt stress was reflected at a metabolite level, the content of GABA in shoots and roots of both genotypes was determined.The GABA amount increased in shoots and roots of the wild type by 2-and 2.5-fold respectively, following the salt stress treatment (Figure 1C and D).However, in gad1/2 shoots the GABA level increased to a detectable level following the salt stress treatment (Figure 1C).Surprisingly, in gad1/2 roots the GABA content was not altered by the salt treatment, despite a four-fold increase in GAD4 transcript (Figure 1D).

Phenotypic analysis of the gad1/2 mutant under salt and osmotic stress conditions
The impact of GABA depletion on the phenotype of gad1/2 mutant under salt stress was investigated.Interestingly, after two weeks of salt stress treatment gad1/2 shoots were completely wilted, a characteristic of osmotic stress induced phenotype (Figure 2A).Ionicstress induced features such as chlorosis and necrosis were not visible.To confirm the effects of the osmotic stress component, wild type and gad1/2 plants were subjected to 300 mM mannitol.Under control conditions, the shoot water content was similar between the genotypes during four days of the osmotic stress treatment period (Figure 2B).However, mannitol treatment significantly reduced the shoot water content of the gad1/2 mutant compared to the wild type (Figure 2B).This difference was evident on the third and fourth days after the initiation of the osmotic stress treatment.

Expression analysis of ABA biosynthetic genes
To examine if the salt stress in gad1/2 mutant leads to induction of salt stress responsive genes, the transcript abundances of ABA synthesizing genes ABA2, ABA3 and AAO3 were analyzed.Both genotypes upregulated the expressions of all three genes after 150 mM NaCl treatment (Figure 3A to C), confirming the occurrence of salt stress effects.The expression of ABA3 was induced nearly by five-fold in shoots of gad1/2 mutant following salt stress treatment (Figure 3B).The level of ABA3 induction in gad1/2 mutant was significantly greater than the wild type (~three-fold) after 150 mM NaCl treatment (Figure 3B).Similarly, the expression of AAO3 was increased by four-fold in shoots of gad1/2 mutant following the salt treatment and it is nearly double to the amount of transcripts measured in wild type shoots under the same treatment conditions (Figure 3C).At control conditions, there was no difference in the expression of all three genes between the genotypes (Figure 3).

Measurement of Na + and K + accumulations
To substantiate the salt oversensitive phenotype of gad1/2 mutant with ionic data, the Na + and K + amounts were determined in shoots.Both genotypes accumulated high amounts of sodium in shoots after 150 mM NaCl treatment; although, it was greater in gad1/2 mutant (Figure 4A).In wild type plants the amount of sodium increased from 2.5 to 12 mg/g DW -1 whereas, in gad1/2 mutant the sodium level increased from 2.5 to 20 mg/g DW -1 following salt stress treatment (Figure 4A).As expected, the potassium content was reduced in both genotypes after salt stress treatment, but it was sever in gad1/2 mutant (Figure 4B).In gad1/2 shoots, the level of potassium was reduced from 28 to 21 mg/g DW -1 which is statistically significant (Figure 4B).The combined effects of high sodium and less potassium accumulations in gad1/2 mutant following salt treatment greatly affected the sodium and potassium homeostasis (Figure 4C).

To examine if low K
+ content in gad1/2 mutant corresponds to the expression pattern of the potassium   transporters, the transcript abundances of some potassium channels (KAT1, AKT1, GORK and HAK5) were analyzed.Under control conditions, there was no difference in shoot transcript abundances of HAK5, KAT1 and GORK genes between the wild type and gad1/2 mutant (Figure 5B to D).However, the expression of AKT1 was reduced by four-fold in gad1/2 shoots under control conditions (Figure 5A).In roots, the transcript abundances of all K + channels, except GORK, were similar between the genotypes under control conditions (Figure 5E, G&H).However, GORK transcript was threefold higher in gad1/2 roots under control conditions (Figure 5F).The treatment of 150 mM NaCl did not affect the expressions of all genes in shoots of the wild type (Figure 5A to D).However, in gad1/2 mutant the transcript abundances of HAK5 and GORK were altered by the salt treatment (Figure 5B and C).In roots of wild type plants, the expressions of AKT1 and GORK were induced by the salt treatment (Figure 5E and F); whereas, in gad1/2 mutant only HAK5 transcript was significantly induced (Figure 5G).

Effects of salt treatment on the metabolic composition of gad1/2 mutant
The importance of the GABA shunt in maintaining the balance of C and N metabolism under salt stress conditions was investigated.The accumulations of less abundant amino acids such as lysine, leucine and isoleucine were significantly increased in gad1/2 shoots compared to the wild type after 150 mM NaCl treatment (Table 1).In contrast, the contents of most abundant amino acids like aspartate, glutamate and asparagine were not altered by the salt treatment (Table 1).The shoot glutamate content, 3.46 µg/g FW -1 in wt and 4.99 µg/g FW -1 in gad1/2, was significantly different between the two genotypes under control conditions; however, this difference was alleviated by the salt stress treatment (Table 1).Despite a considerable difference in shoot amino acid compositions between the wild type and gad1/2 mutant under control conditions, the differences in roots were relatively minimum (Table 2).After salt stress treatment, the root amino acid contents of both genotypes were the same (Table 2).Investigation of the salt stress effect on TCA cycle intermediates revealed major changes.The contents of fumarate, oxaloacetate, malate and citrate were reduced by more than 50% (Figure 6A).In contrast, in gad1/2 mutant the abundance of oxaloacetate, malate and citrate was increased by three or more-fold (Figure 6B).Surprisingly, after salt stress treatment 100-fold higher amount of proline was measured in shoots of gad1/2 mutants compared to wild type.However, no proline was detected in shoots of both genotypes under normal growth conditions (Figure 6C).

Induction of GAD4 transcript did not compensate for the loss of gad1 and gad2
Arabidopsis genome contains five genes encoding glutamate decarboxylase (Shelp et al., 1999).These Table 1.HPLC measurement of amino acids (µmol/g FW) in shoots of four-week-old wild type and gad1/2 mutant treated without and with 150 mM NaCl.five paralogs show organ specificity in terms of expression (Shelp et al., 2012).For example, GAD1 expresses mainly in roots, whereas GAD2 transcript was detected in all organs (Zik et al., 1998).Similar to GAD2, GAD4 transcript was detected in various organs such as shoots, roots, flowers and siliques (Scholz et al., 2015;Renault et al., 2010).The induction of GAD4 expression in response to salt stress (Renault et al., 2010), hypoxia (Miyashita and Good, 2008), drought stress (Urano et al., 2009) and cold treatment (Kaplan et al., 2007) has been shown previously.In the present work, GAD4 transcript was induced in wild type plants treated with 150 mM NaCl, an observation in line with a previous report (Renault et al., 2010).The significant induction of GAD4 transcript in roots of gad1/2 mutant after salt stress treatment might suggest the severity of the stress in the gad1/2 mutant.Despite the induction of GAD4 expression in shoots and roots of the gad1/2 mutant, its effect was not reflected at the metabolite level.This probably suggests a minor contribution of GAD4 to the total GABA pool in tissues.To date, only few metabolic pathways are known to be regulated at transcriptional level (Urano et al., 2009).GABA synthesis in plants is regulated posttranslationally by the Ca 2+ /CaM complex (Bouché et al., 2004).Therefore, the induction of GAD4 expression may not necessarily lead to high GABA levels.

The gad1/2 mutant is oversensitive to the salt and osmotic stress
The induction of GABA shunt activities and GABA accumulation in response to various abiotic stresses such as drought, cold and hypoxia have been reported previously (Kinnersley and Turano, 2000;Miyashita and Good, 2008).Subsequently, the phenotypes of some GABA-shunt mutants were characterized under abiotic stress conditions.Renault et al. (2010) examined the responses of GABA-rich pop2-1 mutants to salt stress treatments and observed oversensitivity to only the ionic component of the salt stress.Recently, GABA-depleted gad1/2 mutant was characterized under drought stress conditions and exhibited over-sensitive phenotype compared to the wild type (Mekonnen et al., 2016).Here, the responses of the gad1/2 mutants to 150 mM NaCl treatment were investigated.Interestingly, the mutant plants exhibited a wilty-phenotype as shown by the loss of turgor and bending of the inflorescence.Salt stress generates different types of effects due to its osmotic and ionic components.The osmotic stress effect is usually rapid and manifested earlier than the ionic component (Munns et al., 1995).In the present work, the osmotic stress component of the salt stress treatment was visible two weeks after the onset of the treatment, as shown by the bending of inflorescences and shrinking of leaves.Furthermore, the lower shoot water content in gad1/2 mutant following 300 mM mannitol treatment suggests the greater effect of the osmotic stress.GABA is one of the most abundant amino acids in roots and accounts up to 7% of the total free amino acids under normal growth conditions (Bouché et al., 2004).The increment of GABA amount by two-fold in wild type roots following salt stress (Figure 1D) probably suggests the importance of GABA in defense against the osmotic stress effect.The function of GABA as osmoticum has been suggested previously (Bown and Shelp, 1997).Furthermore, high GABA accumulating Arabidopsis pop2 mutants showed tolerance to mannitol induced osmotic stress (Renault et al., 2010).Similarly, exogenous application of GABA to maize seedlings increased the endogenous GABA level and subsequently improved tolerance to salt stress (Wang et al., 2017).Sensitivity to the ionic component of salt stress requires the accumulation of Na + to a toxic level.For example, salt overly sensitive (sos3) mutants treated with 100 mM NaCl accumulated 60 mg/g DW -1 Na + (Zhu et al., 2007).Similarly, the salt oversensitive pop2-1 seedlings accumulated ~50 mg/g DW -1 of sodium after 150 mM NaCl treatment (Renault et al., 2010).In gad1/2 mutant, the amount of sodium ion might have increased by 8-fold after 150 mM NaCl treatment however; the concentration (20 mg/g DW -1 ) was comparatively smaller than the amount measured in salt sensitive mutants.The measurement of lower quantity of Na + in gad1/2 mutant after salt stress treatment may not necessarily reflect an alteration in sodium uptake from the soil.It could also be due to a difference in treatment conditions and age of the plant.In fact, GABA has been shown to involve in the uptake of Na + .Arabidopsis seedlings treated with 1 and 10 mM exogenous GABA accumulated more Na + than the control (Essah et al., 2003).However, in the present work low endogenous GABA of gad1/2 mutant did not correspond to low sodium level.Similarly, high endogenous GABA accumulating pop2-1 mutant accumulated similar amount of sodium compared to the wild type following NaCl treatment (Renault et al., 2010).These observations suggest that endogenous GABA do not influence sodium uptake in plants.In contrast, potassium levels seem to have correlation with the endogenous GABA level.High GABA containing pop2 mutants accumulated more K + ion (Renault et al. 2010) than low GABA containing gad1/2 mutant (Figure 4B).Despite a notable change in Na + /K + ratio of gad1/2 mutant following salt stress, the ionic stress symptoms were not visible within the experimental period probably due to the rapid and severe effect of the osmotic stress.
The severity of the salt stress effect in gad1/2 mutant was reflected by the significant increase in transcription of ABA synthesizing genes.Plants accumulate ABA in vegetative tissues when encounter adverse environmental conditions such as drought and salt (Xiong and Zhu, 2003;Zhu, 2002;Seo and Koshiba, 2002).Kefu et al. (1991) showed the rise of ABA levels in barley and cotton leaves following treatment of 75 mM NaCl.
Here, the transcripts of ABA synthesizing genes (ABA3 and AAO3) were significantly induced in gad1/2 mutant following 150 mM NaCl treatment, indicating the severity of the stress.ABA3 and AAO3 are involved in the last step of the ABA biosynthesis pathway (Seo et al., 2000;Bittner et al., 2001).Despite the lack of sufficient experimental evidences on the interaction between the GABA shunt and ABA pathway, the proper functioning of the ABA pathway was required for the GABA effect on 14-3-3 expressions (Lancien and Roberts, 2006).

GABA-depletion
differentially regulated the expressions of AKT1 and GORK in gad1/2 mutants Potassium is one of the most important and abundant cation in plants, and accounts up to 10% of the plant dry matter (Leigh and Wyn Jones, 1984).It plays a central role in many biological processes therefore; it is required in large quantity.To meet this requirement, plants should posses an effective mechanism that involves potassium channels to absorb potassium from the medium through their roots and translocate to the aerial parts (Gierth and Mäser, 2007).In Arabidopsis, three major potassium transporter family proteins, in addition to channels, have been identified (Maser et al., 2001).Despite the existence of diverse channels and transporters in Arabidopsis, two proteins, AtAKT1 and AtHAK5, contribute 84 and 78% of low affinity and high affinity transport of potassium, respectively, in wild type roots (Gierth and Mäser, 2007).In the present work, the difference in shoot K+ content between the Wt and gad1/2 mutant was surprising considering the similarity in root AKT1 and HAK5 expressions under control conditions.In fact, the transcript data alone may not fully explain the difference in shoot K + level.The difference could also be due to a change in its translocation to shoots.AKT1 has been shown to involve in potassium retrieval from xylem sap (Lagarde et al., 1996).The down regulation of AKT1 expression in GABA-depleted gad1/2 shoots probably reduced the potassium retrieval from xylem which in turn has a feedback effect on the loading of potassium to the xylem vessel in roots.Furthermore, the up-regulation of GORK expression in shoots and roots of gad1/2 mutant without and with 150 mM NaCl treatment, respectively, probably led to the efflux of potassium from the plant.Previous works showed that GORK effluxes potassium from roots to the medium, and the process is enhanced under stress conditions (Demidchik, 2014;Jayakannan et al., 2013).The enhancement of GORK activity under salt stress is associated with the depolarization of the plasma membrane due to the uptake of more Na + .Alternatively, the reduced accumulation of potassium in gad1/2 mutant could be due to the lack of GABA-inhibition effect on anion channels like ALMT.The efflux of anions depolarizes membranes and this leads to inhibition of K + uptake.The negative regulation of ALMT proteins by GABA has been reported recently (Ramesh et al. 2015).In general, the low potassium content in shoots of gad1/2 mutant could be the combined effects of reduced potassium retrieval from the xylem, an efflux of potassium from roots Mekonnen 261 and reduced uptake from the soil.

Disruption of the GABA-shunt altered the stressinduced modification of cellular metabolism in gad1/2 mutant
Plants respond to environmental stresses with various kinds of cellular responses such as modification of cell wall architecture, adjustment to membrane system and changes in cell cycle and cell division (Krasensky and Jonak, 2011).Modification in cellular metabolism is also a response that plants undertake during abiotic stresses.Alfalfa roots exposed to 100 and 150 mM NaCl accumulated more amino acids and less organic acids compared to the control (Fougére et al., 1991).Similarly, Arabidopsis plants accumulated more amino acids and less carbohydrate when treated with 150 mM NaCl (Renault et al., 2010).In the present work, the accumulations of amino acids were generally increased in shoots and roots of both genotypes after NaCl treatment (Tables 1 and 2).However, the increment was significant in shoots of gad1/2 mutants than the wild type.
The rise of amino acid levels under stress conditions could come from production or stress-induced protein degradation (Krasensky and Jonak, 2011).
Glutamate is a precursor for many of the amino acids produced in cells (Bouché and Fromm, 2004).In wild type plants, the glutamate content was increased by 44 and 133% in shoots and roots, respectively, after 150 mM NaCl treatment.Therefore, it is not surprising to see an increment in amino acid levels.In gad1/2 mutant, the glutamate content was increased by 35 and 34% in shoots and roots, respectively, after salt treatment.With a relatively smaller change of glutamate content, the accumulation of amino acids in gad1/2 shoots might come from the degradation of proteins induced by stress.Besides, the disruption of the GABA-shut also affects the TCA cycle intermediate.The direct link between the GABA-shunt and the TCA cycle has been shown in tomato (Studart-Guimaraes et al., 2007) and potato (Araújo et al., 2008).The reduction of organic acids (Fougére et al., 1991) and accumulation of GABA (Renault et al., 2010) under salt stress suggests that GABA-shunt is the preferred route for C and N metabolism under such conditions.The increment of organic acid levels in gad1/2 mutants following salt treatment further confirms the importance of the GABAshunt under stress conditions.Proline is another compound that accumulates in response to stresses (Singh et al., 1972;Renault et al., 2010).Despite a modest increase in wild type plants following salt stress, the amount of proline was remarkably increased in gad1/2 mutant.In wild type plants, GABA-shunt seems to be the preferred route for glutamate metabolism than proline syntheses under salt stress condition.Such preferential accumulation of GABA over proline has been reported in tobacco plants exposed to water stress (Liu et al., 2011).GABA and proline are synthesized from a common precursor, glutamate (Krasensky and Jonak, 2011).These observations suggest that GABA plays a central role in salt induced modifications of cellular metabolism.
Taken together, the accumulation of GABA under salt stress plays a certain role to counter salt induced damage.However, further experiments should be conducted to determine the specific function of GABA under salt stress conditions.

Figure 1 .
Figure 1.Analysis of GAD transcripts and GABA content in shoots and roots of four-week-old wild type and gad1/2 mutants; transcript abundance of GAD3 and GAD4 in shoots (A) and root (B) of wild type and gad1/2 mutants treated without and with 150 mM NaCl; GABA content in shoots (C) and roots (D) of wild type and gad1/2 mutants treated without and with 150 mM NaCl; Values are means of three biological replicates for transcript analysis and at least five biological replicates for GABA measurement; error bars show the standard error of means; different letters on top of the error bars show significant differences between treatments after student's t-test, P<0.05.

Figure 2 .
Figure 2. Phenotypic analysis of wild type and gad1/2 mutants treated without and with 150 mM NaCl (A) and 300 mM mannitol (B); shoot phenotype of five-week-old wild type and gad1/2 mutants treated without and with 150 mM NaCl for two weeks; pictures are representative of plants of similar phenotype; shoot water content measurement in wild type and gad1/2 mutants treated without and with 300 mM mannitol; values are means of five biological replicates; different letters on top of the error bars show significant differences between treatments after student's ttest, P<0.05; D2 represent the second day after the initiation of the treatment.

Figure 3 .
Figure 3. Transcript analysis of ABA synthesizing genes; the transcript abundances of ABA2 (A), ABA3 (B) and AAO3 (C) genes were determined in shoots of four-weeks-old wild type and gad1/2 mutants treated without and with 150 mM NaCl; values are means of three biological replicates; error bars show the standard error of means; different letters on top of the error bars show significant differences between treatments after student's t-test, P<0.05.

Figure 4 .
Figure 4. ICP-MS measurement of shoot Na + (A) and K + (B) contents in five-week-old wild type and gad1/2 mutant treated without and with 150 mM NaCl; calculation of the Na + /K + ratio (C); values are means of at least ten biological replicates; error bars show the standard error of means; different letters on top of the error bars show significant differences between treatments after student's t-test, P<0.05.

Figure 5 .
Figure 5. Transcript analysis of major potassium channels; the relative expression of AKT1 (A), GORK (B), HAK5 (C), KAT1 (D) in shoots and AKT1 (E), GORK (F), HAK5 (G), KAT1 (H) in roots of four-week-old wild type and gad1/2 mutants treated without and with 150 mM NaCl; the expressions were normalized against actin; values are means of three biological replicates; error bars show the standard error of means; different letters on top of the error bars show significant differences between treatments after student's t-test, P<0.05, RE-relative expression.

Figure 6 .
Figure 6.GC-MS measurement of TCA cycle intermediates and proline; the shoot TCA cycle intermediates in four-week-old wild type (A) and gad1/2 (B) plants treated without and with 150 mM NaCl; proline content in shoots of wild type and gad1/2 mutants treated without and with 150 mM NaCl (C); values are means of at least eight biological replicates; error bars in (C) show the standard error of means; different letters on top of the error bars show significant differences between treatments after student's t-test, P<0.05; n.d., not detected.

Table 2 .
HPLC measurement of amino acids (µmol/g FW) in roots of four-week-old wild type and gad1/2 mutant treated without and with 150 mM NaCl.
ns * Values are means of at least six biological replicates; asterisks show statistical significance after student's t-test; * P<0.05; ns, non significant