For malate (or succinate) to be sole carbon source for bacteroid respiration, pyruvate must be produced to drive the TCAC. Malic enzyme (ME) activity has been demonstrated in a number of different bacteroid species, from both determinate and indeterminate nodules [48,49,50,51]. ME oxidatively decarboxylates malate to pyruvate and, together with MDH, provides an effective means of generating both acetyl-CoA and OAA for bacteroid TCAC operation. In all species examined, bacteroids contain two ME forms, one of which is NAD⁺ specific and another which is NADP⁺ specific. Knockout mutants of the two isoforms in S. meliloti have shown that NAD-ME is essential for nitrogen fixation, but not NADP-ME [51,52]. These studies show that import of malate alone can provide the energy required for nitrogen fixation in bacteroids.

Given the key role of dicarboxylates, especially malate, in the nitrogen-fixing process, elucidating the mechanisms by which they are transported across cell and symbiosome membranes in nodules is crucial to our understanding of nodule function.

The bacteroid C₄-dicarboxylate system encodes three proteins, DctA, B and D. DctA is responsible for transport of dicarboxylates while the other proteins are involved in regulation of dctA expression. DctA is part of a wider transporter family that is present in both prokaryotes and eukaryotes [38]. It acts as a symporter with two protons transported for each malate, utilising the pH gradient across the bacteroid membrane [[6] see below].

While it is clear that dicarboxylate transporters exist on infected cell membranes and the SM, the molecular identity of these remains unknown. As mentioned above, it has been recently shown that succinate transport is not essential for nitrogen fixation [41], suggesting that malate is the most likely substrate for the SM dicarboxylate transporter in planta. By altering pH while maintaining total malate concentration, Udvardi and colleagues determined that the monovalent form of malate was the preferred substrate for the SM transporter [28]. Malate anion uptake into the symbiosome is affected by the rate of bacteroid respiration, energisation of the SM by a P-type ATPase, and phosphorylation of the transporter, probably via a calcium dependent protein kinase [53,54]. When comparing characteristics of the SM and infected cell dicarboxylate transport systems, many similarities arise. Both require energisation across their respective membranes to facilitate transport down a concentration gradient, and in both systems transport of malate and succinate competitively inhibit the transport of the other [27,28]. However, it is important to note that the transport of these two systems is distinct, as phthalonate potently inhibits transport by the SM dicarboxylate transporter but has little effect on dicarboxylate uptake by infected cells [27]. Both, phthalonate and cyanocinnamic acid are strong inhibitors of the SM transporter, distinguishing it from other plant and bacteroid dicarboxylate transporters and providing a pharmacological “signature” for it.

When considering possible candidates for the SM dicarboxylate transporter, it is important to understand the energetics of the symbiosome: proton pumping by the SM-ATPase into the symbiosome space provides a positive inside membrane potential and an acidic interior [18,55,56]. In this regard, the symbiosome resembles a vacuole and in fact symbiosomes largely replace the vacuoles in infected cells [4]. This makes tonoplast anion transporter families excellent candidates for the SM dicarboxylate transporter. Unfortunately, the nature of symbiosome uptake precludes using complementation of, for example, mae1 deficient yeast or E. coli mutants to screen nodule cDNAs as candidates, since the transporter in question transports malate out of the plant cytosol (into the symbiosome) and therefore may not catalyse the uptake required for complementation. However, we can use available proteomic and transcriptomic data to identify potential candidates for further investigation [57,58,59].

Proteomic studies on isolated SMs have identified a number of putative transporters on the SM from pea, soybean, and L. japonicus [7,60,61], but few of these have been shown to act as anion transporters. Clarke and colleagues identified five members of the Nitrate/Peptide transporter Family (NPF) in soybean SM fractions [7]. Some NPF members can transport malate and other organic acids, but they are thought generally to operate as proton symporters [62] and, given the direction of the pH gradient across the SM, they are likely to transport compounds out of the symbiosome. However, there are some exceptions: in the non-legume Alnus glutinosa an NPF member, designated as DCAT1, localises to the symbiotic interface of infected nodule cells and was found to have dicarboxylate transport activity when expressed in E. coli and Xenopus oocytes [63]. More recently, some members of the NAXT subgroup of the NPF family have been shown to be anion excreters in roots, mainly of nitrate. For example, Arabidopsis AtNAXT1 apparently facilitates the uniport of one NO₃^- per H⁺ pumped by the H⁺-ATPase [64], a mechanism reminiscent of the malate uniporter described in soybean symbiosomes (see above). A closer examination of the NPF proteins identified in the soybean SM proteome is clearly warranted.

Other prospective anion transporters have been identified in soybean SM fractions: these include five ABC transporters and four Voltage-Dependent Anion Channels (VDACs) [7]. ABC transporters are found in all known organisms and are capable of transporting a very wide range of substrates [65], including malate [66]. While ABC transporters share some characteristics with the SM dicarboxylate transporter, including phosphorylation-regulated transport [67], ABC transporters actively transport their substrates. In comparison, malate transport into isolated symbiosome appears to be passive, driven by the electrical potential across the SM [6,28]. The presence of VDAC proteins in SM fractions [60] is likely due to mitochondrial contamination of the SM fractions, since Wandrey and colleagues immunolocalised these VDACs in L. japonicus nodules and found they colocalised with mitochondria but not symbiosomes [68].

While no definitive candidates for the SM transporter have been identified in proteomic studies, transcriptomic data can be used to screen for more suitable candidates. Tissue–specific transcriptomic analyses of M. truncatula and L. japonicus [19,69,70] have been performed and allow identification of nodule enhanced transcripts.

The tonoplast Dicarboxylate Transporter (tDT) family shares some similarities with the SM dicarboxylate transporter, transporting malate [71] and being inhibited by carbonyl cyanide m-chlorophenylhydrazone [72]. However, these proteins act as co-transporters, generally exchanging malate with citrate [72,73]. Additionally, no nodule-specific tDT isoforms are evident in published transcriptomic databases, so they are unlikely to be candidates for the SM dicarboxylate transporter.

SLow Anion Channel (SLAC) and the homologous (SLAH) proteins were originally annotated as dicarboxylate carriers due to sequence similarities with MAE1, a yeast dicarboxylate transporter, and may be involved in dicarboxylate transport in guard cells during stomatal opening [74]. Expression of Arabidopsis SLAC1, SLAH2, and SLAH3 in Xenopus oocytes generated anion currents when phosphorylated, with a preference for nitrate but some permeability to malate [75,76]. There is enhanced nodular expression of at least one member of this gene family in both M. truncatula [69] and L. japonicus [19,70], and they warrant further investigation for intracellular location and transport activity.

The published characteristics of the SM dicarboxylate transporter most closely represent those of the Aluminium-activated Malate Transporters (ALMT) family. This family was first identified in wheat where high aluminium concentrations caused an efflux of malate from the roots [77], but ALMTs have been found in a wide range of tissues from a number of different plant species, where they mediate organic acid transport across membranes. Some of these, such as AtALMT9, are located on the tonoplast where they catalyse malate uptake into the vacuole [78]. ALMTs transport a range of anions, display voltage-dependent gating and are regulated by phosphorylation [79], all characteristics shared with the SM transporter [28,54,80]. Takanashi and colleagues identified seven ALMT transcripts in L. japonicus and three were expressed in the nodule [26]. Of these, LjALMT4 shows nodule-specific expression, but it was expressed only in the vascular parenchyma cells. Other ALMT transcripts are also express in nodules of L. japonicus [81], soybean [82,83], and M. truncatula [69], but these remain uncharacterized and have yet to be localised.

γ-Aminobutyrate (GABA) is a non-protein amino acid that is ubiquitous across all kingdoms of life. This four carbon molecule regulates plant growth and development and its levels change in response to both abiotic and biotic stresses [84,85]. The many roles of GABA include maintaining cytosolic pH, acting as an osmolyte, participating in C/N metabolism, protecting against oxygen deficiency, scavenging reactive oxygen species (ROS), defense against pathogens and herbivory, and more recently acting as a signaling molecule [86,87]. In legumes, GABA is present at quite high concentrations in nodules compared to other amino acids but its function remains unclear. In isolated pea nodules, GABA was the second most abundant amino acid detected in bacteroids originating from different bacterial strains such as Bradyrhizobium japonicum, and Rihizobium leguminosarum [88,89,90]. Further, in pea plants, when nodulated roots were incubated with ¹⁵N₂ for 30 min, a rapid labelling of GABA in the cytosol and bacteroid fractions of R. leguminosarum was detected [91].

GABA is synthesized in the cytosol by the decarboxylation of glutamate by glutamate decarboxylase (GAD) via the GABA shunt in normal cells. Under stress, the GABA shunt is upregulated and synthesized GABA is transported across the mitochondrial membrane by the GABA—permease [84,92,93]. In the mitochondria, GABA is converted to succinate that feeds into the TCA cycle maintaining energy production. Studies suggest that rhizobial bacteroids have a modified GABA shunt that lacks the GAD enzyme and in the nodules of snake bean and soybean, this finding has been supported by low activity and oxygen dependency [94,95]. In soybean, approximately 0.7 μmol g⁻¹ of GABA was present in the nitrogen-fixing nodules while only 0.01 μmol g⁻¹ GABA was detected in bacteroids from cowpea that showed very little GAD activity. The absence of GAD in the bacteroids suggests that the most likely source of GABA in the bacteroids is of plant origin and may be derived from branched chain amino acids supplied to the bacteroids [96]. Further, exogenous application of GABA (15 mM) to M. truncatula petioles doubled the concentration of GABA in the nodules, and increased nodule activity and N₂ fixation [97]. Higher concentrations of GABA were detected in both phloem and nodules under normal conditions which increased when nodules were partially excised [97]. In several legumes such as alfalfa, lupin, cowpea and soybean, varying concentrations of GABA have been found in the phloem [98,99,100,101,102]. These observations suggest that GABA is phloem mobile, is translocated from shoot to root and nodules and may have a role in enhancing symbiotic N₂ fixation.

Similar effects of GABA on nitrate uptake have been observed in the non-legume Brassica napus, wherein GABA from the shoots was translocated to the roots and the uptake of nitrate (NO₃^-) during nitrogen deficiency was positively correlated with GABA concentrations in the phloem [103]. In Arabidopsis seedlings supplied with exogenous GABA (50 mM), similar effects on nitrogen metabolism have been observed under limited nitrogen conditions [104].

Of particular relevance to this entry is the GABA regulation of ALMT proteins that are involved in transport of malate [85,105]. A putative GABA binding motif was identified on the ALMT family of proteins and GABA binding to aromatic amino acid residues in the motif was shown to negatively regulate malate efflux [105]. ALMTs have been characterized in the nodules of L. japonicus, and nodule enhanced transcripts have been detected in both M. truncatula and soybean but remain uncharacterized [26,106]. It is probable that GABA in the nodules exerts regulatory control of malate transport mediated by the ALMTs during N₂ fixation.

All of these observations suggest that GABA functions both as a metabolite and signaling molecule in legumes in response to stress and conditions that enhance nitrogenase activity. However, the physiological role of GABA in nodules remains unclear and future studies should explore the role and mechanism of GABA accumulation and regulation of ALMTs in the nodules.