γ-Aminobutyrate (GABA) Regulated Plant Defense: Mechanisms and Opportunities

Global climate change and associated adverse abiotic and biotic stress conditions affect plant growth and development, and agricultural sustainability in general. Abiotic and biotic stresses reduce respiration and associated energy generation in mitochondria, resulting in the elevated production of reactive oxygen species (ROS), which are employed to transmit cellular signaling information in response to the changing conditions. Excessive ROS accumulation can contribute to cell damage and death. Production of the non-protein amino acid γ-aminobutyrate (GABA) is also stimulated, resulting in partial restoration of respiratory processes and energy production. Accumulated GABA can bind directly to the aluminum-activated malate transporter and the guard cell outward rectifying K+ channel, thereby improving drought and hypoxia tolerance, respectively. Genetic manipulation of GABA metabolism and receptors, respectively, reveal positive relationships between GABA levels and abiotic/biotic stress tolerance, and between malate efflux from the root and heavy metal tolerance. The application of exogenous GABA is associated with lower ROS levels, enhanced membrane stability, changes in the levels of non-enzymatic and enzymatic antioxidants, and crosstalk among phytohormones. Exogenous GABA may be an effective and sustainable tolerance strategy against multiple stresses under field conditions.


Introduction
The world population is predicted to be 9-10 billion people by 2050, so that a 60 −1 10% increase in global food production will be required as more marginal lands are being used for agricultural purposes [1]. Furthermore, at the current rate of global warming, the temperature is projected to increase by 1.5−2.4 • C [2]. With a 1.5 • C increase, heavy precipitation and associated flooding will intensify and be more frequent in most regions of Africa, Asia, North America and Europe. Additionally, more frequent and/or severe droughts will occur in a few regions on all continents except Asia. With further global warming, every region is projected to increasingly experience concurrent and multiple climatic changes (e.g., salinity, O 2 deprivation, acidity, heavy metals), which will adversely affect plant growth and development. These changing climatic conditions could facilitate the geographic expansion and aggressiveness of phytopathogens and modify host susceptibility [3]. Therefore, it is imperative to develop crop production systems that are more sustainable under stress conditions [4].
Cytosolic GABA may also be derived indirectly from the metabolism of polyamines (PA) in organelles [43,44] (Figure 1). In Arabidopsis, the primary PA, putrescine (Put), is generated from the secondary PA, spermidine (Spd), and the tertiary PA, spermine (Spm), via FAD-dependent PA oxidases (PAO) in the peroxisome [45]. Put is produced in the plastid from arginine via arginine decarboxylase, agmatine imidohydrolase and carbamoylputrescine amidohydrolase [17,44]. 4-Aminobutanal (ABAL) can be derived in the peroxisome from both Spd and Put via two copper-dependent diamine oxidase activities (AtCuAOα3 and AtCuAOζ based on the terminology in Tavladoraki et al. [46]) [47]. Terminal oxidation of Put in the plastid requires a plastidial CuAO, likely with a preference for diamines as substrates [17].
Early support for the existence of ABAL/pyrroline dehydrogenase in plants is available from the production of radiolabeled GABA from exogenously supplied radiolabeled Put, and the suppression of GABA production by the addition of aminoguanidine, a CuAO inhibitor [44,48]. More recent research demonstrated the conversion of ABAL into GABA via ABAL/pyrroline dehydrogenase activities (plastidial AtALDH10A8 and peroxisomal AtALDH10A9) [49,50]. Both AtALDH10A8 and AtALDH10A9 have strong alkaline pH optima, are NAD-dependent and use ABAL, as well as 3-aminopropanal and γ-trimethylaminobutyraldehyde (all with K m s in the low micromolar range) as substrates to produce GABA, β-alanine and γ-butyrobetaine, respectively [50,51]. They are also prone to substrate inhibition. ataldh10A8,9 seedlings are phenotypically normal, and the GABA and Glu levels are similar across the mutants and WT [50].
Proline is another potential indirect source of cytosolic GABA [52]. It is derived in the plastid/cytosol from Glu via NADP-dependent ∆ 1 -pyrroline-5-carboxylate synthetase and NADP-dependent ∆ 1 -pyrroline-5-carboxylate reductase [53]. Proline reacts with a hydroxyl radical, resulting in H-abstraction from the amine group, then spontaneous decarboxylation of proline and formation of pyrrolidin-1-yl [52]. Pyrrolidin-1-yl can easily be converted to ∆ 1 -pyrroline/ABAL and oxidized to GABA via ABAL/pyrroline dehydrogenase activity. To date, there is no direct evidence for the contribution of proline to GABA production in planta.

Conversion of GABA to Succinate or γ-Hydroxybutyrate
The mitochondrial-localized bidirectional amino acid transporter/GABA permease (AtBAT1/GABP) links the anabolic and catabolic portions of the GABA shunt [54] (Figure 1). However, GABA uptake by mitochondria isolated from the atgabp mutant is not eliminated, suggesting the possible existence of other mitochondrial GABA carriers with overlapping or redundant functions [16]. The substrate preference for AtGABP requires clarification (arginine, Glu and lysine, but not GABA and proline [55]; GABA, but not proline [54]; K m Spd = 55 µM, K m Put = 85 µM, K m arginine = 1.4 mM [56]).
The conversion of SSA into succinate in the mitochondrion is catalyzed by SSADH activity [71,72] (Figure 1). The AtSSADH is reversible with a pH optimum of 9-9.5 and feedbackregulation by NADH and ATP (K m SSA = 15 µM, K m NAD = 130 µM, K i NADH = 122 µM, K i ATP = 8 mM). The atssadh mutant overaccumulates GABA (28 nmol g −1 FM) and H 2 O 2 by 2-and 4-fold, respectively [73,74]. Succinate can contribute to the production of C skeletons and NADH via the TCAC and the generation of ATP via the mitochondrial electron transport chain (mETC), which in turn, prevents the accumulation of ROS [25,28,74]. Notably, SSADH and nearly every enzyme involved in the TCAC and mETC are succinylated, but the supply of succinate is limited during oxidative stress by lower OGDH and succinyl-CoA ligase activities [75,76]. These findings support the hypothesis that GABA shunt activity is necessary for the modification and regulation of respiratory activities to ensure an adequate ATP supply and minimize the generation of ROS (73).
SSA can be reduced to γ-hydroxybutyrate (GHB) via NADPH-dependent glyoxylate/SSA reductases in the mitochondrion/plastid (AtGLYR2/SSR2) and cytosol (AtG-LYR1/SSR1) [77][78][79][80][81][82][83] (Figure 1). These proteins have affinity for SSA in the low millimolar range, glyoxylate in the low micromolar range, and NADPH in the low micromolar range, and are competitively inhibited by NADP + (K i = 1-3 µM) [78,[82][83][84][85]. The atglyr1,2 mutants and NAD kinase 1 overexpression line accumulate less and more GHB, respectively, with submergence than the WT [86]. The growth of plantlets or roots of various Arabidopsis lines with altered GLYR activity responds differentially to SSA or glyoxylate under chilling conditions [83]. Together, these findings are consistent with an elevated rate of SSA conversion to GHB with cold and low O 2 , and suggest that AtGLYR1,2 are part of an adaptive response to stress-induced changes in redox balance [83]. Notably, the rice osglyr1/2 double mutant displays stunted growth under photorespiratory conditions, compared to the WT [87], validating our earlier hypothesis that the GLYRs reduce both glyoxylate and SSA in planta [88]. The two GLYRs function in a redundant manner, which would be consistent with the diffusion of SSA and glyoxylate from their sites of origin. The fate of GHB in plants is uncertain; however, it could be linked to acetyl-CoA and fatty acid metabolism [12,16].

Biosynthesis of Glutamate from Succinate and 2-Oxoglutarate
Succinate is converted to citrate and 2-OG in the TCAC [25,28] (Figure 1). Two potential routes have been proposed for diverting these two important TCAC intermediates to the generation of cytosolic Glu [19]. The first involves the export of citrate from the mitochondrion via the dicarboxylate/tricarboxylate carrier (AtDTC) and its conversion to isocitrate and then 2-OG in the cytosol. The second involves the export of 2-OG via a 2-OG/malate translocator (AtOMT) [19]; contrary to a recent suggestion [89], we could find no evidence in the literature for a mitochondrial-specific OMT. In both cases, 2-OG would be converted to Glu via cytosolic transaminase activities. A third route is also possible, involving the direct synthesis of Glu from 2-OG via the mitochondrial glutamate dehydrogenase (GDH) and then export via the uncoupling proteins AtUCP1,2 [90,91]. The UCP is known to decrease the electrochemical gradient across the mitochondrial inner membrane, and prevent the over-reduction of the mETC [92].
Glu availability is an essential regulator of GAD activity, and the generation of cytosolic Glu from succinate bypasses two reactions of the TCAC (i.e., 2-OGDH and succinyl-CoA ligase). However, it does not exclude participation in the cytosolic GAD reaction of Glu originating in the plastid via the glutamine synthetase/glutamate synthase (GS/GOGAT) cycle [93] (Figure 1). The movement of 2-OG and Glu across the plastidial inner membrane is mediated by dicarboxylate translocators (AtDIT) [94].

GABA Transport
Arabidopsis grows efficiently on GABA as the sole N source, providing the first strong evidence for GABA uptake by plant cells [95] (Figure 1). Two types of plasma membranelocated amino acid transporters, amino acid permease 3 (AtAAP3) and proline transporter 1,2,3 (AtPROT1,2,3), transport GABA (K m = 12.0 and 1.7-5 mM, respectively [96,97]). AtPROT2-and SlPROT1-mediated GABA transport is inhibited by proline and quaternary ammonium compounds [95,97]. AtGAT1 is a high-affinity plasma membrane-localized, proton-coupled transporter that is apparently specific for GABA (not transporting Glu or Asp) (K m = 10 µM) [98,99]. Endogenous GABA in the atgat1 mutant is unaffected by the addition of exogenous GABA, but it increases in the WT, confirming that AtGAT1 plays a role in GABA influx into the cell [99].
GABA is released from asparagus mesophyll cells via an unknown mechanism [100], but evidence is emerging for the bi-directional transport of GABA across the plasma membrane via the wheat root, aluminum-activated malate transporter (TaALMT1) [18,101,102] ( Figure 1). AtALMT1 is highly homologous to TaALMT1, and there are plant ALMTs that encode channels with a preference for malate, chloride or nitrate [14]. However, their capacity to transport GABA has not yet been investigated.
To date, two organellar transporters for GABA have been described. The mitochondrial AtGABP was the first (see Section 2.1.2). Another one, SlCAT9, is localized to the tonoplast and links the cytosol with the vacuolar compartment, operating strictly in a stoichiometric exchange mode with Glu and aspartate so that the osmolarity of the vacuole does not change [103] (Figure 1).

Precursor-Product Relations and Flux in the GABA Shunt
Kaplan et al. [104] showed that exposure to 4 • C results in the sharp accumulation of both GABA and succinate from 12 to 24 h in the aerial portion of Arabidopsis plants. Subsequently, GABA declines, but succinate remains steady for 72 h. GHB accumulates sharply from 24 to 48 h and then declines to 96 h. Glu accumulates in a linear fashion from 12 to 96 h, whereas Put does so from 24 h. Furthermore, Espinoza et al. [105] demonstrated that exposure to 4 • C increases the GABA and alanine levels from 2 to 30 h; the GABA level remains steady for the remaining 28 h. Proline and Put begin to accumulate shortly thereafter (from 18-22 to 58 h), whereas the Spd level does not change. Interestingly, the succinate and malate levels decrease early in the time course and then remain steady. In contrast, the levels of 2-OG and glutamine (Gln) initially decline and then increase.
Some generalizations are possible from these two studies: (i) the cold-induced accumulation of GABA, alanine, succinate and GHB are not correlated with the availability of Put and proline; (ii) succinate and GHB may be concomitantly generated from SSA; (iii) SSA and succinate turnover may be restricted or an alternate source of succinate exists; and, (iv) Glu/Gln metabolism is altered under cold conditions. A definitive explanation for the increasing accumulation of Glu is not possible, despite the accumulation of three well-known products of Glu metabolism (i.e., GABA, Put and proline), though protein hydrolysis is known to be stimulated by low temperature [106].
Treatment with 50-150 mM NaCl for 6 d increases the GABA level in soybean (Glycine max (L.) Merr.) roots by 11-to 17-fold, compared to the control, as well as the diamine oxidase activity by 52-86%, but decreases the levels of Put, Spd and Spm [48]. Aminoguanidine inhibition of diamine oxidase activity increases the Put level from 28 to 51 nmol g −1 fresh mass (FM), but decreases the GABA level from 10.8 to 6.6 µmol g −1 FM. While these findings could be interpreted as support for the derivation of GABA from Put [48], the molar stoichiometry (∆GABA/∆Put) deviates markedly from the 1:1 ratio expected if Put is a major source of GABA, suggesting that aminoguanidine also interferes with the generation of GABA from Glu.
Other studies have modeled the stress-induced changes in flux through the GABA shunt using suspension cultures. For example, the mean GABA, alanine, proline, Glu and Gln pools in control tomato (Solanum lycopersicum L.) cells are 1.2, 2.0, 0.12, 0.84 and 6.77 µmol g −1 FM, respectively, over a 2-d period, whereas they are 13.0, 6.26, 31.2, 1.22 and 2.71 µmol g −1 FM, respectively, in cells adapted to water stress (25% polyethylene glycol 6000) [107]. Computer simulation of 15 N-labeling kinetics reveals that adaptation to water stress increases the N flux into GABA and alanine, suggesting high pyruvate availability and rapid turnover of both amino acids. The rate of GABA synthesis and catabolism, respectively, are 0.80 and 0.785 µmol h −1 g −1 FM in control cells, and 2.4 and 1.26 µmol h −1 g −1 FM in adapted cells. About 76% of the GABA is located in a metabolically inactive pool in unadapted cells, but only 38% in adapted cells. The proline pool increases by 300-fold due to greater synthetic rates from Glu and restricted oxidation, and the metabolically inactive Gln storage pool becomes depleted. Notably, the rate of nitrogen assimilation doubles, even though the total soluble protein (on a dry mass basis) decreases by 30%.
In related research, the mean GABA, alanine, proline, Glu and Gln pools in unadapted cowpea (Vigna unguiculata (L.) Walp) cells at 26 • C are 0.27, 5.37, 0.42, 1.79 and 4.11 µmol g −1 FM, respectively [108]. The GABA pool size in cells transferred to 42 • C increases to 1.85 and 3.24 µmol g −1 FM after 2 h and 1 d, respectively, and the other amino acids also accumulate, but less extensively than GABA. Total free amino acid levels increase approximately 1.5-fold after 1 d at 42 • C. The computer simulation suggests that heat shock induces a 63-fold increase in the rate of GABA synthesis over the first 2 h, without any change in the rate of GABA catabolism. The rate of GABA synthesis over the next 22 h increases 7-fold, and this is accompanied by a 3-fold increase in GABA catabolism. The rates of alanine and proline synthesis increase by 0.6 and 1.7-fold, respectively, over the 1-d period. The size of the free amino acid pool increases within 1 d, and the rate of protein synthesis decreases by 20-30%. An accelerated rate of protein degradation may also contribute to the effects of heat shock on the amino acid perturbations [108].
The stress-activated acceleration of flux through the GABA shunt can be explained by increases in GAD activity and GABA synthesis due to elevated levels of Glu (e.g., reduction in ATP availability, recycling of NH 3 and protein synthesis; increase in protein degradation) [93,[109][110][111], and the Ca 2+ /CaM complex or H + in the cytosol. The GOGATor GDH-mediated regeneration of Glu from 2-OG in the plastid and mitochondrion, respectively, is sufficient to sustain the increased formation of cytosolic GABA for a limited duration. However, the relative importance of PAs and proline to GABA accumulation remains controversial [35].

Respiration, Redox Balance and Reactive Oxygen Species
Recent studies have assessed the interaction between GABA catabolism and respiration. For example, increases in GAD activity and GABA level (1-fold) in leaves of tomato plants after 5 d of exposure to 200 mM NaCl, together with the 25% decrease in succinate [112], suggest that the SSADH-mediated production of succinate is insufficient to sustain the average rate of respiration. An increase in H 2 O 2 level supports this interpretation. Che-Othman et al. [89] also found a 1-fold increase in the GABA level in wheat seedlings 3 to 11 d after exposure to 150 mM NaCl, compared to the control. Furthermore, increases in GAD activity (pH 5.8), and the levels of succinate, 2-OG, Glu and Gln, together with decreases in the activity and abundance of pyruvate dehydrogenase and OGDH, and the levels of citrate, aconitate, fumarate and malate suggest that: (i) elevated GABA shunt activity accounts for an approximate 20% increase in the respiration rate of salt-stressed leaves, despite the lower potential for pyruvate oxidation by TCAC; and (ii) GOGAT and GDH are likely indicators of N assimilation into Glu [89].
Respiration in soybean roots is reduced by 40%, after 6 h of hypoxia, but the ATP and pyruvate supply is enhanced via an activated glycolytic pathway, and the cytosolic NAD + is regenerated via fermentation reactions [113]. The direct flux of pyruvate into the TCAC is low, and the conversion of succinate to fumarate is markedly decreased due to restricted pyruvate dehydrogenase and succinate dehydrogenase activities. Pyruvate accumulation is minimized via the alanine transaminase-and GABA-TP-mediated formation of alanine, and the alanine transaminase reaction generates 2-OG for use by OGDH and succinyl CoA ligase to produce another ATP. The NAD + required for oxidation of 2-OG is apparently produced by the anti-clockwise operation of the TCAC malate dehydrogenase, which increases malate accumulation. The carbon flux from SSA to succinate is not eliminated, even though the NADH/NAD + ratio is presumably altered to some degree. GABA accumulates, at least in part, due to the stimulation of GAD activity by bound Ca 2+ /CaM or lower cytosolic pH. Overall, both GABA and succinate appear to be temporary storage metabolites that readily supply the TCAC when hypoxia is mitigated [114].
The hypoxia-induced increase in NADPH/NADP + ratio might be attributed to increases in the activities of NAD kinases [86] and the oxidative pentose phosphate pathway [106]. In addition, there is evidence for NADPH-mediated reduction of alternative oxidases and then activation by either pyruvate or succinate [115], NADPH-mediated removal of SSA via SSR activity [78,83], and NADPH oxidase-mediated generation of H 2 O 2 [116,117]. With the exception of hypoxia, abiotic stresses cause stomatal closure and decrease CO 2 fixation, leading to the underutilization of NADPH, over-reduction of the photosynthetic electron transport chain, and the generation of ROS. Stomatal closure also increases Rubisco activity, leading to the glycolate oxidase-mediated generation of H 2 O 2 [88]. Table 1 describes examples wherein the phenotype of GABA pathway mutants is modified under stress conditions. In general, the atgad2 single and atgad1/2 double mutants of Arabidopsis do not accumulate GABA in response to salinity, drought or hypoxia, but they become hypersensitive to the stress [40,69,70,118]. This suggests that AtGAD4-, PA-or proline-derived GABA is insufficient to counter the stress under consideration, though ataldh10A8,9 single mutants have been shown to decrease the GABA level and confer a hypersensitive salinity phenotype [50]. In contrast, the GABA level in the atgaba-t mutant increases with salinity or hypoxia and the phenotype is more tolerant to the stress [70,118]. Furthermore, the loss in GABA accumulation in the atgad1/2 mutant results in more RBOHF/NADPH oxidase-mediated H 2 O 2 production and K + efflux via the GORK and Shaker-type outward rectifying K + channels than in the atgaba-t mutant [70,117,118]. Surprisingly, another study reported that the GABA level in the shoot of the atgaba-t mutant doubles with salinity (28 µmol g −1 DM), while the succinate and 2-OG levels decrease or are unchanged, but the mutant becomes hypersensitive to salinity [68]. The atssadh mutant overaccumulates GABA, H 2 O 2 and GHB in the absence of stress, and displays hypersensitivity to heat stress [73,74].

Genetic Manipulation of Endogenous GABA Modifies the Abiotic Stress Phenotype
Knockdown of tomato SlGAD1-4 prevents the salinity-induced GABA accumulation and confers a hypersensitive salinity phenotype, whereas knockdown of SlGABA-T1-3 slightly increases the GABA level, while conferring a hypersensitive salinity phenotype [112] (Table 1). In contrast, knockdown of SlSSADH increases the GABA level and confers a salinity tolerant phenotype. Furthermore, the GABA and H 2 O 2 levels are higher and lower, respectively, than the corresponding levels in the SlGAD1-4 and SlGABA-T1-3 knockdown lines (12.5 and 0.5 µmol g −1 FM, respectively, vs. 6 and 0.75-0.8 µmol g −1 FM). Thus, the tolerance is inversely related with the H 2 O 2 level.
Overall, these findings suggest that the capacity to tolerate stress is associated with H + consumption during GABA synthesis, pH regulation of H + -ATPase, activation of the GABA shunt and TCAC, down-regulation of NADPH oxidase, and GABA inhibition of GORKmediated K + efflux ( [21,60,61,118]; also see Section 3.1). Notably, GABA is implicated in activating the transcription of 14-3-3 proteins, which are known activators of H + -ATPases and GORK [119,120].

Root Malate and K + Efflux
Wheat root TaALMT1 is responsible for malate efflux, but this is inhibited by GABA binding to the cytosolic surface of the protein [18,101,102] (Figure 1). Indeed, both malate and GABA likely traverse the protein, though not simultaneously or through the same pore [18]. The GABA-binding domain is conserved in the GORK channels of the root epidermis [20]. The atgork1 mutant displays an hypoxia (waterlogging)-tolerant phenotype, greater K + retention and hypoxia-induced Ca 2+ signaling, and does not show any change in K + efflux in response to GABA [20,117] (Table 2). Therefore, one would expect the stress-induced accumulation of GABA to reduce the export of malate and K + from root cells [13,18,19,102]. Notably, the GABA-binding domain is conserved in most ALMTs [14] and further research is required to establish if they transport GABA, as well as anions.

Stomata Functioning
ALMTs and GORK facilitate the functioning of stomata [121] (Figure 1). During the opening, malate and Cl-fluxes into the guard cell vacuole are mediated by tonoplast AtALMT6 and malate-activated AtALMT9, respectively. During the closure, the efflux of malate and K + is mediated by plasma membrane-localized AtALMT12 and AtGORK, respectively. The drought-induced increase in abscisic acid (ABA) level upregulates GORK activity, resulting in stomatal closure and drought tolerance. The opening of AtALMT6 and AtALMT4 mediates malate efflux from the vacuole. Various Ca 2+ -permeable channels contribute to the elevation of cytosolic Ca 2+ during stress [121,122].
A role for GABA accumulation in drought resistance has been shown using Arabidopsis mutants with altered GABA levels. The atgad1/2 double mutant has impaired stomatal closure and a drought-susceptible phenotype [69] (Table 1). Furthermore, the light-todark transition is slower, and GABA levels are lower than those in the WT without or with drought. Subsequent research with atgad2 and atalmt9 mutants demonstrated that the drought-induced accumulation of GABA suppresses light-induced stomatal opening, whereas it has no effect under constant light [40] (Tables 1 and 2). Stomatal opening in the almt12 mutant showed WT sensitivity to GABA, whereas dark-induced stomatal closing is insensitive to GABA [40] (Table 2). These findings indicated that GABA accumulation in the cytosol of the guard cell reduces stomatal reopening and transpirational water loss, thereby improving drought tolerance. Further research is required to investigate iGABA inhibition of ALMT4,6 by cytosolic GABA contributes to the regulation of stomatal movement [123] (Table 2).

Overexpression of Malate Efflux Is Linked to Aluminum Tolerance
The accumulation of free aluminum (Al 3+ ) ions in the soil solution under low pH limits plant growth and productivity [124]. Ramesh et al. [101] have monitored malate efflux in roots of near-isogenic, Al 3+ tolerant (ET8) and sensitive (ES8) lines of wheat. In the absence of Al 3+ , GABA has no effect on malate efflux under acidic conditions. However, malate efflux increases in response to Al 3+ treatment, and decreases in response to Al 3+ and exogenous GABA in ET8, but not in ES8. While the GABA level is higher in ET8 than in ES8 under acidic conditions, Al 3+ reduces the GABA level in both lines to similar levels. This suggests that high endogenous GABA inhibits the activity of TaALMT1 in the absence of Al 3+ , and that both malate and GABA are exported from the cytosol when TaALMT1 is Al 3+ activated, resulting in the sequestration of Al 3+ and modulation of the plant sensitivity to Al 3+ [102,124]. Table 2 summarizes various examples wherein ALMT1 overexpression enhances the efflux of malate in response to Al 3+ treatment under acidic conditions [101,[125][126][127][128]. Further research is required to establish the precise role of GABA efflux under acidic conditions in the presence of Al 3+ [102]. It could be related to the alleviation of ammonium toxicity [129].

Interaction between Plants and Other Organisms
Recent advances in our knowledge of the interactions between plants and other organisms have been reviewed in detail [16,61,[130][131][132]. GABA inevitably accumulates in the host plant in response to bacterial and fungal infection, and infestation by invertebrate pests; however, the mechanism of action for GABA appears to differ. For example, the development of insect larvae and root-knot nematodes is delayed, presumably by disrupting the function of neuromuscular junctions [133][134][135][136][137], whereas in bacterial pathogens, quorum sensing is down-regulated, modulating the hypersensitive response in the host [138][139][140][141][142][143][144]. On the other hand, accumulated GABA may boost host endurance against fungal pathogens by sustaining TCAC activity and reducing oxidative damage [132,[145][146][147]. Table 3 briefly summarizes examples wherein genetic manipulation of endogenous GABA modifies biotic stress resistance. For example, transgenic tobacco and Arabidopsis plants with elevated GABA levels are more resistant to infection by Agrobacterium and Pseudomonas than WT plants [123,124], as well to predation by insect larvae and root-knot nematode [119][120][121][122]. In contrast, tomato plants with extremely low GABA levels are more susceptible to infection by Ralstonia [128]. Together, these findings provide strong support for the role of GABA in plant defense [114][115][116].

Genetic Manipulation of Endogenous GABA Modifies the Biotic Stress Resistance
Recently, Deng et al. [148] reported that activation of the mitogen-activated protein kinase (MPK3/MPK6) signaling cascade greatly induces GABA biosynthesis in Arabidopsis ( Table 3). The gad1/2/4 triple and gad1/2/4/5 quadruple mutants, in which the GABA levels are extremely low and the Glu and alanine levels are compromised, are more susceptible to both Pst and Pst-avrRpt2. Functional loss of AtMPK3/AtMPK6, their upstream AtMKK4/AtMKK5, or their downstream substrate, WRKY33, suppresses AtGAD1 and AtGAD4 expression after Pst-avrRpt2 treatment. These findings lend support for involvement of the MPK3/MPK6 signaling cascades in the induction of GAD and plant immunity against bacterial and fungal pathogens [149].  Table 4 briefly summarizes many examples from the literature wherein tolerance to hypoxia, drought, salinity, chilling, heat, osmotic stress and proton stress, as well as heavy metals (i.e., aluminum, arsenic and chromium) is improved in vegetative organs by the application of exogenous GABA [39,. The application of exogenous GABA typically increases the level of endogenous GABA, and elicits a diverse range of biochemical, molecular and physiological responses. The activity of the GABA shunt is increased to sustain the TCAC and energy production, though the precise response depends on the organ (shoot vs. roots) and the stress under consideration. Activities of N assimilation (including protein degradation) and PA pathways can also be modulated (also see [185]). Elevated endogenous GABA is also responsible for further increasing the stress-induced levels of non-enzymatic (ascorbic acid, GSH, phenols) and enzymatic (e.g., ascorbate oxidase, superoxide dismutase, ascorbate peroxidase, monodehydroascorbate reductase, glutathione reductase, glutathione peroxidase, glutathione S-transferase, catalase) antioxidants and osmolytes (sugars, organic and amino acids, including proline). These result in lower levels of ROS, malondialdehyde (a product of ROS-mediated peroxidation of membrane polyunsaturated fatty acids and marker for the depletion of antioxidant systems), and protein carbonylation (product of protein peroxidation and a marker of oxidative damage), lower activities of NADPH oxidase, lipoxygenase and polyphenol oxidase, and restoration of ion homeostasis (electrolyte leakage), which is indicative of membrane stability.    There is also evidence for the GABA-induced production of nitric oxide (NO) ( Table 4), which could be associated with the enhancement of antioxidant defense, as well as regulation of epigenetic mechanisms and gene transcription [157,174,186]. The stress tolerance could also be related to GABA-induced changes in pathways associated with other phytohormones such as ABA (ABA receptors), ethylene (ACC oxidase, ACC synthase), PAs (arginine decarboxylase, free and conjugated forms, S-adenosylmethionine decarboxylase) and salicylate (Table 4) [60,156,172,175,187], which can regulate metabolic homeostasis and influence the expression of stress factors (miRNAs, transcription factors, heat shock proteins) with known and yet-to-be-determined functions [156,183,188]. It is known that the exogenous application of GABA, ABA and salicylate alleviate the drought-induced damage to membranes and leaf water status in creeping bentgrass by affecting similar metabolic pathways, yet cause differential changes in metabolite accumulation [155]. Additionally, NO and nitrate reductase are jointly needed for salicylate-induced water-stress tolerance in pepper plants [189]. The osmotin protein, which belongs to the PR-5 family of pathogenesisrelated proteins, is known to inhibit the activity of defensive cell wall barriers in fungi [190]. Cinnamyl alcohol dehydrogenase is involved in lignin biosynthesis and alkenyl reductase and can detoxify cytotoxic substrates such as aldehydes [155]. Further research is required to investigate the effects of exogenous GABA on alternative respiratory pathways involved in the scavenging, regulation and homeostasis of ROS and NO [92,191], the action of other phytohormones [23,156,187,[192][193][194][195][196][197], and post-translational modifications and epigenetic regulation of gene expression [147,198] during stress.

Concluding Remarks
Plants must endure a wide variety of abiotic and biotic stresses under field conditions. To prevent significant yield losses, many crop improvement programs strive to develop stress-tolerant cultivars. Plants may respond uniquely to different or simultaneous stresses, so breeding tolerance against one stress may be at the expense of tolerance to another. With climate change, plants will likely experience more extreme weather events or multiple stresses, including plant diseases. It would therefore be beneficial to develop a tolerance strategy against multiple stresses.
GABA metabolism has garnered considerable attention in recent years, in part because it often accumulates in response to a variety of abiotic (cold, heat, drought, salinity, salinityalkalinity, osmotic, low O 2 , heavy metal toxicity) and biotic (invertebrate pests, bacteria, fungi) stresses (Figure 2). These findings are typically attributed to stimulation of GABA anabolism or inhibition of GABA catabolism. However, there are clear cases in which GABA pathway activity is promoted to sustain respiration and the generation of energy, without the accumulation of GABA. On the other hand, accumulated GABA may bind to AMLT and GORK, interfering, respectively, with the transport of malate in stomatal guard cells and K + in root epidermal cells, thereby enhancing plant tolerance to drought and hypoxia. With very few exceptions, genetic manipulation of GABA metabolism and receptors, respectively, reveal positive relationships between GABA levels and abiotic/biotic stress tolerance, and between malate efflux from the root and heavy metal tolerance.
Common plant responses to avoid or tolerate abiotic and biotic stresses include stomatal closure and corresponding decreases in photosynthesis, and reduced leaf growth and root length, as well as greater ROS activity. These responses are coordinated by phytohormones such as ABA, NO, ethylene, salicylate and jasmonate. Thus, the enhancement of endogenous GABA by either genetic engineering or the application of exogenous GABA reduces the stress-induced ROS level and restores or partially restores the morphophysiological features of the unstressed phenotype by promoting or modifying activities of the GABA shunt, TCAC, antioxidant, secondary metabolism and phytohormone pathways ( Figure 2). Furthermore, elevated plant GABA adversely affects the activity of phytopathogens by various mechanisms. Therefore, exogenous GABA might function under field conditions as an effective and sustainable tolerance strategy against the multiple abiotic and biotic stresses that could be exacerbated by climate change.
Low temperature and controlled atmosphere conditions (low O 2 , elevated CO 2 ) are extensively employed to extend the postharvest life of horticultural commodities, especially botanical fruits. However, these crops may suffer from chilling injury and other physiological disorders, as well as fungal decay. The use of exogenous GABA to improve the marketability of stored horticultural commodities will be described in a companion paper. abiotic/biotic stress tolerance, and between malate efflux from the root and heavy metal tolerance.

Figure 2.
Abiotic and biotic stress-induced changes in morphological and physiological features are restored or partially restored by increasing endogenous GABA and GABA shunt activity. Depending upon the nature of the abiotic stress, the entry of pyruvate into the tricarboxylic acid cycle via pyruvate dehydrogenase, the catabolism of 2−oxoglutarate and succinate via 2−oxoglutarate dehydrogenase and succinate dehydrogenase, respectively, and activity of cytochrome oxidase in the mitochondrial electron transport chain may be restricted, thereby limiting the generation of molecules associated with energy transfer (NADH, FADH and ATP). These limitations are to some extent overcome by increasing the levels of endogenous GABA, by either GAD overexpression or application of exogenous GABA. The elevated endogenous GABA increases flux through the shunt, which in turn, increases the level and/or entry of succinate into the non−cyclic tricarboxylic acid cycle, mitochondrial electron transport chain, and accordingly, ATP generation. The elevated GABA also modifies the activity of stress-induced pathways involving enzymatic and non−enzymatic antioxidants, N assimilation, secondary products, and phytohormones (NO; ethylene and ABA are not shown) by uncharacterized mechanisms. The activity of phytopathogens is directly inhibited by accumulated GABA. Symbols: X, biochemical reaction potentially inhibited by stress; white filled arrows, increasing or decreasing activity of pathway(s) affected by GABA; black filled arrow, increasing level of metabolite affected by stress. Abbreviations: ALMT, aluminum−activated malate transporter; Cyt Ox, cytochrome oxidase; NADK, NAD kinase; OPPP, oxidative pentose phosphate pathway; PDH, pyruvate dehydrogenase; SAM, S−adenosylmethionine; SDH, succinate dehydrogenase; TCAC, tricarboxylic acid cycle; TH, transhydrogenase; see Table 4 legend for the remaining abbreviations.
Common plant responses to avoid or tolerate abiotic and biotic stresses include stomatal closure and corresponding decreases in photosynthesis, and reduced leaf growth and root length, as well as greater ROS activity. These responses are coordinated by phytohormones such as ABA, NO, ethylene, salicylate and jasmonate. Thus, the enhancement of endogenous GABA by either genetic engineering or the application of exogenous GABA reduces the stress-induced ROS level and restores or partially restores the morphophysiological features of the unstressed phenotype by promoting or modifying activities Figure 2. Abiotic and biotic stress-induced changes in morphological and physiological features are restored or partially restored by increasing endogenous GABA and GABA shunt activity. Depending upon the nature of the abiotic stress, the entry of pyruvate into the tricarboxylic acid cycle via pyruvate dehydrogenase, the catabolism of 2−oxoglutarate and succinate via 2−oxoglutarate dehydrogenase and succinate dehydrogenase, respectively, and activity of cytochrome oxidase in the mitochondrial electron transport chain may be restricted, thereby limiting the generation of molecules associated with energy transfer (NADH, FADH and ATP). These limitations are to some extent overcome by increasing the levels of endogenous GABA, by either GAD overexpression or application of exogenous GABA. The elevated endogenous GABA increases flux through the shunt, which in turn, increases the level and/or entry of succinate into the non−cyclic tricarboxylic acid cycle, mitochondrial electron transport chain, and accordingly, ATP generation. The elevated GABA also modifies the activity of stress-induced pathways involving enzymatic and non−enzymatic antioxidants, N assimilation, secondary products, and phytohormones (NO; ethylene and ABA are not shown) by uncharacterized mechanisms. The activity of phytopathogens is directly inhibited by accumulated GABA. Symbols: X, biochemical reaction potentially inhibited by stress; white filled arrows, increasing or decreasing activity of pathway(s) affected by GABA; black filled arrow, increasing level of metabolite affected by stress. Abbreviations: ALMT, aluminum−activated malate transporter; Cyt Ox, cytochrome oxidase; NADK, NAD kinase; OPPP, oxidative pentose phosphate pathway; PDH, pyruvate dehydrogenase; SAM, S−adenosylmethionine; SDH, succinate dehydrogenase; TCAC, tricarboxylic acid cycle; TH, transhydrogenase; see Table 4 legend for the remaining abbreviations.

Supplementary Materials:
The following are available online at https://doi.org/10.5281/zenodo.52 95721, Table S1: Important Arabidopsis thaliana genes associated with GABA metabolism and signaling.  Institutional Review Board Statement: Not applicable. This study did not involve humans or animals.
Informed Consent Statement: Not applicable. This study did not involve humans or animals.

Conflicts of Interest:
The authors declare no conflict of interest.