Role of glutamate metabolism in bacterial responses towards acid and other stresses

Authors

  • C. Feehily,

    1. Bacterial Stress Response Group, Department of Microbiology, School of Natural Sciences, National University of Ireland, Galway, Ireland
    Search for more papers by this author
  • K.A.G. Karatzas

    Corresponding author
    • Bacterial Stress Response Group, Department of Microbiology, School of Natural Sciences, National University of Ireland, Galway, Ireland
    Search for more papers by this author

Correspondence

Kimon-Andreas G. Karatzas, School of Food Biosciences, University of Reading, Reading RG6 6AD, UK. E-mail: k.karatzas@reading.ac.uk

Summary

Glutamate plays a central role in a wide range of metabolic processes in bacterial cells. This review focuses on the involvement of glutamate in bacterial stress responses. In particular, it reviews the role of glutamate metabolism in response against acid stress and other stresses. The glutamate decarboxylase (GAD) system has been implicated in acid tolerance in several bacterial genera. This system facilitates intracellular pH homoeostasis by consuming protons in a decarboxylation reaction that produces γ-aminobutyrate (GABA) from glutamate. An antiporter system is usually present to couple the uptake of glutamate to the efflux of GABA. Recent insights into the functioning of this system will be discussed. Finally, the intracellular fate of GABA will also be discussed. Many bacteria are capable of metabolizing GABA to succinate via the GABA shunt pathway. The role and regulation of this pathway will be addressed in the review.

Introduction

Glutamate is an important molecule for all living organisms, which plays a role in various metabolic processes. It is a nonessential amino acid involved in protein synthesis and other fundamental processes such as glycolysis, gluconeogenesis and the citric acid cycle (Berg et al. 2007). It is also a key metabolite because it serves to link nitrogen and carbon metabolism (Berg et al. 2007). Catabolism of glutamate occurs mainly by the action of either glutamate dehydrogenase or glutamate decarboxylase (GAD) (Berg et al. 2007). The first enzyme, among other roles, is important for the assimilation of ammonia to amino acids, while the second is important for resistance mainly against acid but also other stresses (Berg et al. 2007). As glutamate metabolism in bacterial stress responses is at the core of this review, it will focus in the decarboxylation of glutamate to γ-aminobutyrate (GABA) and the subsequent catabolism of GABA through the GABA shunt.

The GAD is an enzyme that catalyses the decarboxylation of glutamate to GABA (O'Byrne and Karatzas 2008). The decarboxylation of glutamate is the first step of the GABA shunt pathway that leads to the production of succinate by the sequential actions of GABA-aminotransferase (GABA-AT) and succinic semialdehyde dehydrogenase (SSDH; Zhu et al. 2010). The GAD enzyme is present in a wide variety of organisms from all kingdoms of life. In mammals, glutamate is the main excitatory and GABA the main inhibitory neurotransmitter (Petroff 2002). The conversion of the one to the other by GAD plays an important role in various brain functions.

However, in plants and micro-organisms, the GAD system appears to play other roles (Maras et al. 1992). In plants, the GAD system is expressed in response to a variety of stress conditions such as temperature shock, hypoxia or increasing levels of Ca2+ (Shelp et al. 1999). In micro-organisms, the system seems to be related mainly with resistance against acidic conditions, although a role involving other stresses has been seen also. The GAD system has been described as the most important mechanism of acid resistance in the highly acid tolerant bacterium Escherichia coli (Capitani et al. 2003; Foster 2004). Acid resistance would allow various foodborne pathogens or spoilage bacteria to grow on acidic foods. Furthermore, this property is also a virulence or probiotic factor, as it allows pathogens or probiotic organisms, respectively, to pass through the extremely acidic conditions of the stomach barrier. However, a balance between acid resistance and other functions is essential as increased acid and stress resistance have been shown to impair colonization or virulence (Karatzas et al. 2005, 2007, 2008). The decarboxylation of glutamate consumes a proton, and therefore, micro-organisms take advantage of this property to remove protons from the intracellular milieu under acidic conditions (O'Byrne and Karatzas 2008). It operates in a cycle that starts with the import of an extracellular glutamate molecule (Glte) in exchange for an intracellular GABA (GABAi) through the glutamate/GABA antiporter (O'Byrne and Karatzas 2008). These antiporters are usually a part of the GAD system in most micro-organisms (Table 1). Subsequently, the imported glutamate is decarboxylated by the GAD, with the incorporation of a proton in the α-carbon of this molecule in the place of the carboxyl group to form GABA (Karatzas et al. 2012; Fig. 1). The bond is highly stable and it cannot be subjected to ionization, preventing the release of the proton in the intracellular milieu. Subsequently, the GABA molecule that carries the removed proton is either exported by the antiporter as extracellular GABA (GABAe) or remains inside the cell (GABAi) as has been shown previously (Karatzas et al. 2010). The current model for the removal of protons by the GAD system has been debated, however, and other hypotheses have been put forward, which are analysed here. In bacteria, the GAD system is usually accompanied with one or two glutamate/GABA antiporters (Table 1). The antiporter increases the availability of glutamate to the GAD enzymes increasing the capacity of the system to remove protons and, therefore, the efficiency of the system.

Figure 1.

Schematic representation of the function of the extracellular (GADi) and intracellular (GADi) GAD system in Listeria monocytogenes under severe acidic conditions (adopted from Karatzas et al. 2012). GadD2 and GadD3 convert Glti to GABAi (GADi). GadD2 also processes Glte imported by the GadT2 antiporter (GADe) which is transformed to GABA, which is then exported by GadT2. Under these conditions, GadD1 and GadT1 do not play a role in acid resistance (Cotter et al. 2005; Karatzas et al. 2012). GABA, γ-aminobutyrate; GAD, glutamate decarboxylase.

Table 1. The variability in the architecture of the bacterial glutamate decarboxylase (GAD) system
SpeciesDecarboxylasesAntiportersReference
  1. N/A represents not available information regarding the presence of the gene in the organism.

  2. a

    Presence of GAD based on bioinformatic evidence.

Escherichia coli GadA, GadBGadCSmith et al. (1992)
Shigella flexneri GadA, GadBGadCWaterman and Small (2003a)
Mycobacterium tuberculosis GadA Cole et al. (1998), Cotter et al. (2001)
Listeria monocytogenes GadD1, GadD2, GadD3GadT1, GadT2Cotter et al. (2005), Karatzas et al. (2012)
L. monocytogenes ser. 4 and various strainsGadD2, GadD3GadT2Cotter et al. (2005), Karatzas et al. (2012)
Lactobacillus reuteri GadBGadC1, GadC2Su et al. (2011)
Lactococcus lactis GadBGadCSu et al. (2011)
Lactobacillus plantarum GadBGadCSu et al. (2011)
Clostridium perfringens  GadC 
Methanocaldococcus jannaschii a GadBN/AKezmarsky et al. (2005)
Brucella abortus a GadBGadCRoop et al. (2003)
Bifidobacterium dentiumGadBGadCVentura et al. (2009)
Bifidobacteria (other species than Bif. dentium)Not presentNot presentVentura et al. (2009)

The by-product of the GAD system is GABA, which is the substrate of the GABA shunt pathway. The GABA shunt pathway was first described during the study of guinea pig brain cells in 1970 (Balazs et al. 1970). This pathway involves the catabolism of GABA derived from the GAD system to succinate via a two step enzymatic process (Fig. 2). In the first step, a GABA-AT catalyses the reversible conversion of GABA to succinic semialdehyde (SSA), where the amino group of GABA is donated to an α-ketoglutarate molecule with the subsequent production of glutamate (Zhu et al. 2010). The second step is catalysed by an SSA dehydrogenase and converts the SSA to succinate (Zhu et al. 2010). Despite over forty years of study, the role of the GABA shunt pathway in bacteria has not been fully elucidated, although involvement in glutamate metabolism, anaplerosis and oxidative stress response has been suggested (de Carvalho et al. 2011). The pathway has been shown to play an important role in mammalian cells, including a mechanism for the recovery of four of the five carbons lost from the TCA cycle, although there is a discrepancy in the estimation of metabolic activity of this cycle. Cells lose approximately 8% of the energy they would normally generate by passage through the alternative TCA cycle reactions (Waagepetersen et al. 1999). The GABA shunt pathway may also be a way to generate energy from glutamate without the production of ammonia as the enzymes of the pathway are pyridoxal phosphate dependent (Waagepetersen et al. 1999).

Figure 2.

Depiction of the possible catabolism of intracellular GABA through the GABA shunt. The Glti is transformed to GABAi through glutamate decarboxylase under acidic conditions (left side of the vertical line). Subsequently, GABA transaminase (GABA-AT) could convert GABA to glutamate and in parallel α-ketoglutarate is converted to succinate semialdehyde (SSA). Subsequently, SSA could be transformed to succinate by the effect of succinate semialdehyde dehydrogenase (SSDH). GABA, γ-aminobutyrate.

The glutamate decarboxylase system

The GAD system is present in various bacterial species (Table 1), and it has been shown to play a role in their acid resistance (Cotter et al. 2001; Capitani et al. 2003; Su et al. 2011). With the recent developments in genomics, and the increasing availability of genomes of various organisms, it became clear that the GAD system seems to be present even in archaea like Methanocaldococcus jannaschii (Kezmarsky et al. 2005). Despite this wealth of information, it has not been possible to associate the presence of the GAD system with a specific kind of micro-organisms or an environmental niche. It has been suggested that the GAD system is important for survival in the low pH of the stomach, although Salmonella sp. (Foster 2004) and Campylobacter sp. (unpublished data) that are responsible for most of the cases of food-related diseases around the world (WHO 2012) do not possess a GAD system.

Furthermore, the architecture of the GAD system seems to be subject to high variability between species as some possess one, two or even three decarboxylases accompanied with none, one or two antiporters (Table 1). The case of Mycobacterium tuberculosis is quite interesting as it seems to possess a GAD that is not accompanied by an antiporter (Cole et al. 1998; Cotter et al. 2001). In addition, in other cases, there is variability within the same species. For example, Listeria monocytogenes normally possesses three decarboxylases (GadD1, D2, D3) and two antiporters (GadT1, T2), but serotype 4 and a few other strains do not have the GadD1 and GadT1 (Cotter et al. 2005; Karatzas et al. 2012). More interestingly, Bifidobacterium dentium possesses a GAD system, in contrast to all other Bifidobacteria spp. that do not seem to possess one, according to the genomes of this species that have been deciphered (Ventura et al. 2009).

The GAD enzyme, the heart of the GAD system, is a pyridoxal 5′-phosphate (PLP)-dependent enzyme with an acidic optimum pH (3·8–4·6) which forms a hexamer as demonstrated by its crystal structure (Capitani et al. 2003). At neutral pH, the enzyme (GadB) is localized exclusively in the cytoplasm, while at acidic conditions, it is recruited to the membrane where it is able to work synergistically with the glutamate/GABA antiporters. Residues 13–57 of GadB act as hexamerization arms, while residues 3–15 assume different conformations in the neutral- and low-pH forms, playing a significant role in the acid pH-driven association of GadB to the membrane (Capitani et al. 2003). However, there is evidence showing that a GAD from L. monocytogenes (probably GadD3) deviates from this model, processing solely pools of Glti and being unable to act synergistically with the antiporters (Karatzas et al. 2012).

The function of the glutamate decarboxylase system

In various publications, it has been mentioned that the conversion of glutamate to GABA is beneficial for the cell because GABA is less acidic than glutamate (Tramonti et al. 2002; Lei et al. 2011) due to the fact that the pI of the former is close to 7·0 (neutral), while that of the latter is approximately 3·1 (acidic). However, how could this affect their buffering capacity? The glutamate and GABA have similar pKa values for the side chain carboxyl group (pKa = 4·0) and the α-amino group (pKa = 10; Foster 2004). The only difference lies in the fact that glutamate possesses an α-carboxyl group that has a pKa of 2·1. This suggests that a possible difference in the ability of these two compounds to remove protons would occur in a pHe <3 that is lethal for most bacteria. In addition, pHi values would never reach to such low levels. Furthermore, in contrast to GABA, glutamate incorporates protons at 2 < pH < 3 because of its additional α-carboxyl group suggesting that Glte is more beneficial than equimolar levels of GABAe. This is exemplified by a simple experiment where we constructed the titration curves of 100 mmol l−1 aqueous solutions of monosodium glutamate and GABA for 10% HCl w/w (Fig. 3). It is clear that up to pH 3·0, both solutions have the same buffering capacity. But below pH 3·0, the glutamate solution has a higher buffering capacity as it requires more than triple the amount of HCl required compared to the GABA solution, to reduce the pH from 3·0 to 2·0 (Fig. 3). This suggests that it is more beneficial for the cells not to convert glutamate to GABA. However, it is evident that this conversion is beneficial because it occurs in a wide range of pH values and it incorporates one proton from the intracellular milieu in the molecule, in the place of the α-carboxyl group (Fig. 1).

Figure 3.

Titration curves of 100 mmol l−1 aqueous solutions of monosodium glutamate (MSG) and γ-aminobutyrate (GABA) for a solution of 10% HCl w/w. (▲) 100 mmol l−1 GABA and (□) 100 mmol l−1 MSG.

The futile proton hypothesis

Until now, there is little experimental work demonstrating the functionality of the commonly accepted model for GAD system function and the fate of the protons. In contrast, the futile proton hypothesis has been proposed by various investigators in the field (Castanie-Cornet et al. 1999; Storz and Hengge 2000; Tucker et al. 2002; Waterman and Small 2003b), which proposes that if the pKa of glutamate and GABA are taken in account, the GAD system is unable to remove intracellular protons under severe acid stress and the acid resistance might be through an unknown mechanism. A careful investigation into the pKa of glutamate, GABA and the range of pH values intracellularly and extracellularly leads us in two different possibilities that are demonstrated in Fig. 4. The first scenario (A) refers to a mild pH stress where the extracellular pH is >4·5. In this case, Glte is mainly deprotonated and as such is imported in the cell, where the decarboxylation of glutamate removes one intracellular proton (H+*). Subsequently, the GABA carries out this proton, and thus, there is a benefit for the cell. However, in the second scenario (B) where the pH is very low (e.g. 3·5), the chain carboxyl group of glutamate is mainly protonated. This means that the antiporter imports the glutamate in a protonated state, and once intracellularly, this proton (H+) is released in the intracellular milieu that counteracts the benefit from the proton consumed by the decarboxylation process. The latter scenario supports the futile proton hypothesis. However, the GAD system has an obvious increase in acid resistance probably through an unknown mechanism. It has been suggested that this mechanism is a pathway for endogenous glutamate generation or regeneration coupled with the GAD system (Castanie-Cornet et al. 1999; Tucker et al. 2002). The intracellularly generated glutamate might not introduce any protons in the intracellular milieu, and therefore, its decarboxylation will remove protons with a net gain for the cell. A possible candidate pathway that could generate intracellular glutamate (Glti) might involve a glutaminase that converts glutamine to glutamate (Castanie-Cornet et al. 1999; Tucker et al. 2002; Waterman and Small 2003b). Glutamine's only carboxyl group has a pKa of 2, and it would be deprotonated under normal acidic conditions (pH >2). Therefore, its import would not introduce any protons in the intracellular milieu, and the glutamate decarboxylation would result in a net loss of protons and a benefit for the cell. Another alternative pathway could be the GABA shunt that leads to the regeneration of intracellular glutamate by the action of GABA-AT (Fig. 2).

Figure 4.

Representation of two different scenarios on the fate of protons derived from glutamate and source of the protons removed by glutamate decarboxylation. Scenario (a) describes the fate of glutamate, γ-aminobutyrate (GABA) and protons under mild acidic conditions. In this case, both carboxyl groups of glutamate are deprotonated as their pKa is lower than the extracellular pH. In the case of scenario (b), the extracellular pH is very low (<3·5) in the lethal range for most bacteria. In this case, the side chain carboxyl group is mainly protonated (H) because its pKa is higher than the extracellular pH. In this protonated state, glutamate is imported in the cell. Subsequently, this proton (H+) is released in the intracellular milieu, and it is consumed during the glutamate decarboxylation to GABA and thus exported out of the cell. In this case, no proton is removed from the intracellular milieu, and therefore, no benefit exists for the cell, as described by the futile proton hypothesis.

The intracellular glutamate decarboxylase system (GADi)

All of the above suggest the conversion of intracellular glutamate to GABA by the GAD system. Such a process would also remove intracellular protons from the cell under acidic conditions. Some workers have previously demonstrated the presence of GABAi in E. coli (Castanie-Cornet et al. 1999), Shigella flexneri (Waterman and Small 2003b) and L. monocytogenes (Karatzas et al. 2010). However, the significance of this finding has been overlooked. Until now, all studies on the GAD system assume that the exported GABAe represents the total GAD activity. The antiporter works in such a way that, for every molecule of glutamate that is imported, one molecule of GABA is exported. Therefore, the GAD system through the utilization of Glte should not be able to affect the levels of Glti and GABAi. This suggests that GABAi derives directly through the decarboxylation of Glti under acidic conditions and that the GAD system utilizes Glti (Fig. 1). This is supported by data showing that under acidic conditions, mechanisms that increase the levels of Glti are upregulated. In Streptococcus mutans, it has been demonstrated that the import of glutamate through the GlnQHMP transporter increases acid resistance (Krastel et al. 2010), while in L. monocytogenes, the most highly acid-induced genes are coding for a putative glutamate synthase (lmo1734) and a putative ABC transporter of glutamine (lmo1738 and lmo1740) a glutamate precursor (Satorhelyi 2005). Because the utilization of the Glte and Glti are two independent processes, we have previously proposed the division of the GAD system in extracellular (GADe) and intracellular (GADi; Karatzas et al. 2012). At first glance, GADi does not seem to be efficient because of the limited availability of the Glti substrate compared to the vast levels of Glte. However, GADi plays a significant role in the acid resistance of L. monocytogenes removing in an average between strains, a quarter of the protons removed by the overall GAD system (Karatzas et al. 2012). Furthermore, EGD-e, the most studied strain of L. monocytogenes, appears to have a defective GADe unable to export GABA, although it has an active GADi, accumulating GABAi under acidic conditions (Karatzas et al. 2012). The effect of GADi in acid resistance might also explain the function of the GAD system in Myco. tuberculosis that possesses a GadA but no glutamate/GABA antiporters (Cole et al. 1998; Cotter et al. 2001). The important role of the GADi plays in acid resistance remains to be elucidated in other micro-organisms.

Regulation of the glutamate decarboxylase system

The regulation of the GAD system is complex and it has been studied extensively in E. coli. In this organism, RpoS activates gadY, whose small-RNA (sRNA) product stabilizes gadX (Sayed et al. 2007). This increases the GadX levels, which together with GadW binds and activates the gadA and gadBC promoters in vitro (Tramonti et al. 2006). However, in vivo, the latter process requires gadE (Gong et al. 2004), which is activated by GadX and GadW (Sayed et al. 2007), and in turn, activates gadA and gadBC (Tucker et al. 2003; Hommais et al. 2004; Weber et al. 2005). Similarly to the architecture of the GAD system, its regulation is subject to variation. In L. monocytogenes, all regulatory elements from E. coli (gadE, gadY, gadX, gadW) have not been identified, while upregulation of gadT2D2 and gadD3 but not gadD1T1 is SigB-dependent and takes place during entrance in the stationary phase (Cotter et al. 2001; Wemekamp-Kamphuis et al. 2004; O'Byrne and Karatzas 2008). However, exposure of stationary phase cells of L. monocytogenes 10403S to acidic conditions upregulates gadD1T1 and gadD3, but not gadT2D2 operon. However, this is peculiar as this operon plays an important role under severe acidic conditions (Karatzas et al. 2010).

Furthermore, various compounds and conditions seem to activate the GAD system. Apart from acidic conditions, it has been demonstrated that sodium (Richard and Foster 2007) and polyamines (Jung and Kim 2003) lead to the upregulation of the GAD system in E. coli. In both E. coli (Blankenhorn et al. 1999) and L. monocytogenes (Jydegaard-Axelsen et al. 2004), hypoxia has been shown to upregulate the GAD system. Interestingly, the GAD system functionality is variable between different bacteria and probably depends on the environmental niches they are present. For example, although the GAD system of E. coli is functional in minimal media (Castanie-Cornet and Foster 2001), in L. monocytogenes, the system is nonfunctional despite the presence of glutamate (Karatzas et al. 2010). Variability exists also between strains of the same species as L. monocytogenes LO28 utilizes the GAD system in rich media like TSBY, but not in BHI although the 10403S strain utilizes the GAD system in both media efficiently (Karatzas et al. 2012).

The γ-aminobutyrate shunt pathway

The conversion of glutamate to GABA by GAD is the first step of the GABA shunt pathway. However, in bacteria where most of the GABA is exported by the GAD antiporter, the GABA shunt deals with the steps that follow the generation of GABA. The accumulation of intracellular GABA under acidic conditions suggests that its removal should involve a catabolic pathway. This could be assumed as the glutamate/GABA antiporters are only able to remove quantities of GABA that are equimolar to those of glutamate they import, leaving the excess in the cell. Therefore, the first enzyme in the GABA shunt is a GABA-AT (Zhu et al. 2010). This involves the reversible conversion of GABA to SSA, where the amino group of GABA is donated to an α-ketoglutarate molecule. The by-product of this reaction is glutamate (Fig. 2). It has been shown that plants can use pyruvate and glyoxylate as alternative amino group acceptors (Bouché and Fromm 2004; Clark et al. 2009); however, bacteria appear to only have α-ketoglutarate-dependent GABA-AT activity, with the exception of some Rhizobium spp. (Prell et al. 2002). The sole reliance on α-ketoglutarate as the receiving oxoacid is also the case for GABA-ATs present in mammals (Shelp et al. 1999). The importance of this enzyme in bacteria is often highlighted by the presence of more than one gene coding a functional GABA-AT. Studies with E. coli and Pseudomonas syringae have identified the possession of up to three genes encoding functional GABA-ATs (Buell et al. 2003; Kurihara et al. 2010). In E. coli, it appears that induction of GabT, the primary GABA-AT, occurs at pH 9·0 (Stancik et al. 2002) suggesting that this pathway may be more active at alkaline pH conditions.

The second enzyme in the pathway is a SSDH (Zhu et al. 2010). This irreversible step involves the oxidation of SSA to succinate with the formation of CO2. Similarly to GABA-AT, bacteria can possess up to three SSDH encoding genes (Buell et al. 2003). Each SSDH can possess different characteristics and cofactor specificities. For example, in E. coli, there are currently two described SSDH enzymes, GabD and YneI; the former is dependent on NADP+, and YneI preferentially utilizes NAD+ (Donnelly and Cooper 1981b; Fuhrer et al. 2007). YneI can, however, use NADP+, but the activity utilizing this cofactor is only 15% (Donnelly and Cooper 1981a). The presence of this varying NAD(P)+ dependency has also been well described in several species of Pseudomonas (Jakoby and Scott 1959; Padmanabhan and Tchen 1969), as well as both NADP+- and NAD+-dependent SSDH activity shown in Bacillus thuringiensis (Zhu et al. 2010). Partially purified SSDH from plants shows an optimal alkaline pH of 9 (Shelp et al. 1999). This correlates well with the optimal pH for the GABA-AT if the two enzymes are to work together in a metabolic pathway.

Metabolism through the GABA shunt is an important source of nitrogen for bacteria. Arginine, ornithine, agmatine and putrescine are all used as nitrogen sources by E. coli (Schneider et al. 2002). Arginine, ornithine and agmatine are first converted to putrescine, and the pathway subsequently converts putrescine to GABA that is then catabolized via the GABA shunt. In the absence of either the gabT or gabD genes, GABA-AT and SSDH activity remain. This activity is induced through the use of putrescine as a sole nitrogen source (Kurihara et al. 2010). Furthermore, puuE, a gene forming an operon with other members of the putrescine metabolic pathway, was shown to act as a secondary GABA-AT (Fuhrer et al. 2007). It is interesting to note that none of the other genes in the puu gene cluster encode an SSDH; however, YneI, which has been predicted as a secondary SSDH (Kurihara et al. 2010), is induced along with PuuE by putrescine (Zaboura and Halpern 1978). Despite this evidence for the role of the GABA shunt in nitrogen metabolism, E. coli cannot use GABA as a sole source of nitrogen (Belitsky and Sonenshein 2002). The reasons for this are unclear, as E. coli does possess a transporter capable of transporting GABA. Bacillus subtilis has also been shown to utilize GABA; however, unlike E. coli, B. subtilis cannot catabolize putrescine or generate GABA through GAD because of the absence of any GAD encoding genes (Dover and Halpern 1972). Therefore, it imports extracellular GABA via a GABA permease gapP (Ferson et al. 1996). This pathway can serve as the sole nitrogen source for B. subtilis. Prell et al. (2002) suggested that the pathway may be used in nitrogen metabolism for bacteroides, in particular Rhizobium leguminosarum. Some atmospheric nitrogen that is fixed to ammonium during its symbiotic relationship with legume nodules is retained by the bacteria. Results show that GabT is specifically induced by the bacteroid in the nodule state, indicating the enzyme may be involved in catabolizing the glutamate derived from ammonium (Prell et al. 2002).

The regulation of the GABA shunt pathway is a subject that little is known about. Much of the initial work investigating this pathway was performed in E. coli. It was observed that the pathway is regulated by a form of catabolite repression (Dover and Halpern 1972). Further work has since localized the genes to a region called the gab cluster, which contains the four genes gabT, gabD, gabP and gabC (Metzer et al. 1979; Bartsch et al. 1990). Transcriptional control of these genes is complex and involves the csiD-ygaF-gabDTP regulon. This operon is reported to be controlled by the three promoters csiDp, gabDp1 and gabDp2 (Metzner et al. 2004). The first two promoters are σS-dependent, while the third is likely to be induced under limited nitrogen availability by Nac/σ70 (Metzner et al. 2004). More recent work with B. thuringiensis (Zhu et al. 2010) identified the role of a sigma factor, σ54, in controlling the GABA shunt. This regulates a transcriptional activator, GabR, which itself is also autoregulated. The complete mechanism for regulation has not yet been elucidated in these bacteria; however, the gab cluster in B. thuringiensis is a good example of how difficult is to make comparisons and generalizations on GABA shunt regulation, even between species of the same genus. Both gabD and gabT form a GABA-inducible operon in B. subtilis (Belitsky and Sonenshein 2002); however, these two genes are individually regulated in B. thuringiensis by GabR (Zhu et al. 2010). The involvement of ATP as a noncompetitive inhibitor of Arabidopsis thaliana SSDH increases the ways by which the GABA shunt in bacteria could be regulated. This implies that the energy status of the cell itself could be involved in the regulation of the pathway (de Carvalho et al. 2011).

While much of the focus of the GABA shunt research has been on its role in carbon and nitrogen metabolism, there is ground to suggest that it plays a much broader role in bacteria and has a significant impact in terms of both survival under environmental stresses and towards their ability to cause disease. As discussed earlier, the GABA shunt can operate as an alternative pathway to certain steps of the TCA cycle, providing succinate. Many bacteria including L. monocytogenes, Myco. tuberculosis and B. thuringiensis lack a complete set of genes required for the TCA cycle (Glaser et al. 2001; Tian et al. 2005a,b). In fact it has been demonstrated that the SSDHs of Myco. tuberculosis along with an α-ketoglutarate decarboxylase can lead to the generation of succinate in a pathway which is required for normal growth of the bacteria (Metzner et al. 2004). While Myco. tuberculosis does not use the first stage of the GABA shunt, it does indicate that shunt-derived succinate could be of beneficial value to bacteria. Furthermore, the TCA cycle has been proposed as an important process in the formation of spores in Bacillus species (Rutberg and Hoch 1970); however, Aronson et al. (1975) describe that the GABA shunt can successfully complement the loss of α-ketoglutarate dehydrogenase in this process (Tian et al. 2005a).

As the gab gene cluster in E. coli is under the control the multistress response factor σS, it is thought that the GABA shunt may play a role in stress response (Metzner et al. 2004). The mechanism of its involvement is unclear; however, it has been suggested that the control of the GABA shunt may affect the glutamate levels that are seen to increase in response to osmotic stress (Metzner et al. 2004). This may well be the case as the first step in the pathway generates glutamate. Furthermore, a putative SSDH in L. monocytogenes, namely Lmo0913, has been shown to be induced in response to NaCl (Abram et al. 2008b) that would be required to remove the toxic SSA accumulated by the generation of glutamate in this manner. The GABA shunt may also play a role in acid tolerance. The antiport-independent accumulation of GABA in response to acid by L. monocytogenes (Karatzas et al. 2010) would necessitate the catabolism of this metabolite. In fact deletion of lmo0913 in this bacterium results in an acid-sensitive strain (Abram et al. 2008a), highlighting the potential link between acid stress survival and the GABA shunt.

A further, yet understudied role for the pathway has been seen to link GABA metabolism and virulence. A mutation in GABA-AT has been reported to reduce bacterial virulence in the plant pathogen Ps. syringae (Tian et al. 2005a). Another micro-organism, Agrobacterium tumefaciens, uses GABA in damaged plant tissues as signalling molecule that regulates virulence functions (Chevrot et al. 2006). Levels of GABA increase in wounded plant tissues and they signal the modulation of quorum sensing in the bacterium, thereby affecting its virulence on plants. A role for GABA metabolism has been outlined in plant fungal pathogens such as Fusarium graminearum and Cladosporium fulvum (Solomon and Oliver 2002; Carapito et al. 2008). These fungi alter the metabolism of the plant-host to provide the GABA as a nitrogen source. The GABA shunt is also involved in the formation of crystals and spores in B. thuringiensis (Zhu et al. 2010).

Overall, the GABA shunt pathway is a much more important metabolic route than the attention and knowledge of it would suggest. Apart from its role in glutamate cycling, it has been shown to play a role in both secondary carbon and nitrogen metabolism of bacteria. This simple two step pathway is quite variable among bacteria, with the number of genes encoding the pathway as well as the regulation of these genes varying greatly. This makes the establishment of a general model for the GABA shunt in bacteria difficult and necessitates the need for species-specific studies. The majority of our knowledge in this area has been attained through work with mammals, typically with an emphasis on novel drug discovery. The GABA shunt is, however, showing to be a much more important pathway in bacteria than it would appear.

The role of glutamate metabolism in other stresses

It is well known that the GAD system and glutamate catabolism play a role in plants under hypoxia (Shelp et al. 1999). However, there is not so much evidence linking the GAD system and hypoxia in micro-organisms. In previous work, it has been demonstrated that anaerobic conditions trigger the transcription of the gad genes in L. monocytogenes (Jydegaard-Axelsen et al. 2004) and E. coli (Blankenhorn et al. 1999). Furthermore, the gad genes have been shown to play an important role in the survival and growth of L. monocytogenes in various modified atmosphere-packaged foods (Francis et al. 2007). This suggests that the GAD system might play a role in survival or growth under oxygen limitation. The role of the GAD system under anaerobic conditions might be linked with the acidification of the medium and the intracellular milieu that occurs during fermentation. This remains to be investigated. However, in E. coli, upregulation of GadA has also been reported anaerobically under alkaline conditions (pH 9; Blankenhorn et al. 1999). This might suggest that the GAD system plays an additional role apart from increasing the intracellular pH during fermentation. The production of GABA during fermentation has been studied extensively in lactic acid bacteria (Siragusa et al. 2007). The lactic acid produced during fermentation by the bacteria lowers the pH and thus protects the product from growth of pathogenic bacteria. The acidic conditions created lead the lactic acid bacteria to utilize the GAD system and export GABA. Interestingly, this property has lately gained interest as the GABA contained in foods has been shown to have beneficial health effects like the reduction in blood pressure (Inoue et al. 2003; Hayakawa et al. 2004; Gobbetti et al. 2010).

In E. coli O157:H7, it has been demonstrated that glutamate and arginine play an important role in the protection against oxidative stress (achieved through H2O2 or diamide) under acidic conditions (Bearson et al. 2009). Furthermore, the authors demonstrated that this protection required gadC and adiA for the glutamate- and arginine-dependent acid resistance systems, respectively (Bearson et al. 2009). In Saccharomyces cerevisiae, it has been shown that expression of a GAD homologue is required for normal oxidative stress tolerance (Coleman et al. 2001; Capitani et al. 2003). In addition, they showed that exposure to H2O2 increased the expression of SSDH, which catalyses GABA (Coleman et al. 2001). This suggests that the activity of the glutamate catabolic pathway (GAD system and GABA shunt) promotes resistance against oxidative stress. An explanation for the role of the GAD system in oxidative stress might be related to its effects on the intracellular pH. It is well known that oxidation reactions are affected by the pH, and therefore, the increase in pH achieved by the GAD system might change the route of oxidative reactions and processes (Coleman et al. 2001). It has also been suggested that the NADH or NADPH produced by succinate SSDH during the catabolism of GABA (formed by the GAD system) affects redox changes in the cell (Coleman et al. 2001).

Glutamate and GABA are two of the most prominent compatible solutes present in bacteria (Csonka 1989; Wood 2011). Compatible solutes are defined as intracellular organic solutes, which, at high concentration, allow ‘conventional’ enzymes to function efficiently (Brown 1990). They play diverse roles in various stresses protecting macromolecular structures helping to maintain metabolic functions under multiple stresses (Brown 1976) including solvent, chaotrope and hydrocarbon-induced stresses (Bhaganna et al. 2010), high temperature, freeze-thaw treatment or drying (Welsh 2000). Each individual compatible solute has physicochemical properties that confer protection in relation to each of the stress mechanisms. Under osmotic stress, cells accumulate potassium ions and increase glutamate de novo synthesis to counterbalance the high levels of potassium (Goude et al. 2004). The high levels of potassium glutamate provide temporary protection (Gralla and Vargas 2006; Wood 2011). The potassium salts of glutamate are thought to be the primary intracellular ions that play a role in the stabilization of DNA–protein interactions (Wood 2011). However, high concentrations of these salts disturb cellular metabolism, and therefore, they are less preferred osmolytes for long-term protection against hyper-osmotic shock. The high internal concentration of potassium glutamate is apparently not retained under conditions of restored growth in hyper-osmotic conditions (Csonka 1989). However, the high levels of potassium glutamate act as an intracellular signal of osmotic stress inducing the uptake of osmoprotectants (Goude et al. 2004). It has been shown that levels of GABA increase in response to osmotic stress in a variety of bacteria, but this was rather because of uptake from the medium than synthesis (Measures 1975).

The GAD system of Lactobacillus plantarum has been shown to be upregulated in response to bile (Bron et al. 2006). Furthermore, other genes encoding proteins involved in glutamate metabolism are regulated by bile, namely glutamate dehydrogenase and glutamate tRNA-ligase (lp_1169 and lp_0609; Bron et al. 2006). However, the homologue of GadC (lp_2799) was not upregulated, which might be though due to sufficient levels of the antiporter protein to carry out the glutamate/GABA antiporter process (Bron et al. 2006). Furthermore, in L. monocytogenes, disruption of gadT1 encoding for a glutamate/GABA antiporter resulted in sensitivity to bile (Begley et al. 2002). It is suggested that the GAD system may play a role in combating the low-pH stress arising from intracellular dissociation of bile salts (Begley et al. 2002).

The GAD system has also been shown to play a role in nisin resistance of L. monocytogenes (Begley et al. 2010). This is achieved through the GadD1 that has been shown to play a role in growth under mild acidic conditions. Mutants missing GadD1 were impaired in their ability to tolerate lethal and sublethal concentrations of the lantibiotic nisin (Begley et al. 2010). This decarboxylase seems to be important for nisin resistance of other Gram-positive bacteria including Lactococcus lactis subsp lactis that produces nisin. Interestingly, Lactococcus lactis subsp cremoris that has a nonfunctional GadB because of a frameshift mutation is unable to grow in the presence of nisin (Nomura et al. 2002; Begley et al. 2010).

Conclusions

Overall, glutamate metabolism plays an important role in resistance to acid stress and multiple other stresses in well-studied bacteria like E. coli and L. monocytogenes. However, the great variability in the usage of this mechanism between different species and even strains of the same species suggests that more work needs to be performed to fully understand its contribution in stress resistance. Furthermore, as more genomes of micro-organisms become available, it is revealed that this pathway affects the stress resistance of more micro-organisms than previously thought. This can form the basis for further work in the understanding of the stress resistance of these organisms and reveal possible novel functions for glutamate metabolism.

Acknowledgements

The authors are grateful to their colleagues and especially to Dr Conor O'Byrne in Microbiology at NUI Galway for helpful discussions and comments on the manuscript. The work was supported by a Science Foundation Ireland Starting Investigator Research Grant (SIRG; 09/SIRG/B1570) and a Marie Curie European Reintegration Grant (ERG 265154) awarded to K.A.G. Karatzas.

Ancillary