Nitrogen sensing in legumes

Abstract

Legumes fix atmospheric nitrogen (N) in a symbiotic relationship with bacteria. For this reason, although legume crops can be low yielding and less profitable when compared with cereals, they are frequently included in crop rotations. Grain legumes form only a minor part of most human diets, and legume crops are greatly underutilized. Food security and soil fertility could be significantly improved by greater grain legume usage and increased improvement of a range of grain legumes. One limitation for the use of legumes as a source of N input into agricultural systems is the fact that the formation of N-fixing nodules is suppressed when soils are replete with n. In this review, we report what is known about this process and how soil N supply might be sensed and feed back to regulate nodulation.

Legumes, Lotus japonicus, Medicago truncatula, nitrate-sensing, nitrogen, nodulation.

Legumes as an N source

Nitrogen (N) is often the key growth-limiting nutrient in agricultural systems, and chemical fertilizers are applied to crops to improve and maintain yields at optimal levels. On a worldwide basis, a large proportion of the N requirement of crops is provided by legumes through their symbiotic association with N-fixing bacteria, in a process called nodulation. This ability to acquire fixed atmospheric N in N-poor soils greatly reduces or eliminates the need for added fertilizer. Another important advantage of growing a legume crop is that it can increase the N levels of soils, benefiting subsequent crops, and this has led to the near-universal deployment of legumes in crop rotations throughout the world. However, legume nodulation is suppressed by high levels of soil N. This is a homeostatic mechanism which presumably allows the plant to balance the high cost of N fixation with its N requirements (Gibson and Harper, 1985). The application of N fertilizer quickly inhibits both the formation and N fixation activity of nodules (Streeter, 1988). Many agricultural soils contain a high level of residual N (Miller et al., 2007) which limits legume nodulation and N fixation. Furthermore, farmers frequently apply N fertilizer to the seed bed of legumes to help with crop establishment. This practice is likely to inhibit legume nodulation until the soil N supply has been depleted. Therefore, understanding how legumes sense and signal their N supply status to regulate nodulation is of fundamental importance for developing more sustainable agriculture using lower inputs of chemical fertilizer. Based on data from Medicago truncatula and Lotus japonicus genomic studies and advances made using the Arabidopsis model, we are now in a position to address the role of N sensing in the inhibition of legume nodulation and potentially to break the link between N fixation and soil N levels.

Soil forms of N and nodulation

There can be large variations in concentration and form of N supply in soil (Miller et al., 2007) and the plant N status depends on the soil supply. There are two main forms of inorganic N available in the soil, and these are nitrate (NO₃^–) and ammonium. In aerated soils in temperate climates NO₃^– is the dominant N supply form, usually exceeding all other N supply forms by at least 10-fold. For example, in a 1 km transect across fields, soil core analysis at 4 m intervals showed NO₃^– concentrations varied from 10 µM to 6–7 mM (0.14 to 84–98 ppm), while the ammonium concentrations ranged from to 1–2 µM to 0.6 mM (Miller et al., 2007). Both of these N forms can be acquired by roots, and the transporter gene families have been well characterized (Ludwig et al., 2007; Tsay et al., 2007; Orsel and Miller, 2010). As both NO₃^– and ammonium can be utilized by plants, it seems likely that both N forms may therefore be effective in inhibiting nodule formation in the field. In the few experiments where NO₃^– and ammonium sources have been compared for the same species, both N forms were shown to be equally effective at inhibiting nodulation (Dan and Brix, 2009). However, this relative inhibitory effect has not been compared in detail and perhaps cannot be generalized as the amount of legume N fixation and NO₃^– assimilation is reported to depend on the species of plant, the partner symbiotic bacterial strain, and environmental conditions (Becana and Sprent, 1987). In hydroponic culture when the N supply can be controlled, concentrations of both NO₃^– and ammonium <0.5 mM did not inhibit nodulation in the perennial legume Sesbania sesban (Dan and Brix, 2009). Concentrations >5 mM of both NO₃^– and ammonium inhibited nodulation in an Acacia species growing in aeroponic culture (Weber et al., 2007). In the same research using low N concentrations, 0.4 mM ammonium was shown to stimulate nodulation (Weber et al., 2007). Adding external supplies of N has been shown to inhibit the N fixation activity of established nodules (Herdin and Silsbury, 1989), in addition to the formation of new nodules (reviewed by Streeter, 1988; Zahran, 1999). In soybean, 0.5 mM KNO₃ treatment did not inhibit nodulation; increasing the concentration to 7.5 mM inhibited nodulation by 50% (Day et al., 1989). However, the inhibitory effects of KNO₃ treatment on nodulation in soybean in these experiments were shown to depend on the strain of Bradyrhizobium used. Nonetheless, supernodulating soybean mutants have been identified which were less sensitive to inhibition by NO₃^– (Day et al., 1989). Whether N enters a legume from the soil or by fixation from the atmosphere it is assimilated into organic N forms and these molecules are likely to feed back to regulate the first steps (see Fig. 1). Comparing legumes and non-legumes shows how the assimilatory pathways result in the accumulation of different types of organic N molecules and the regulatory feedback mechanisms may reflect this diversity (see later).

Fig. 1.

Diagram showing the interactions between N fixation and nitrate reduction and ammonium assimilation in legume root nodules. NR, nitrate reductase; NiR, nitrite reductase. Nitrate is known to be located in the phloem and xylem. The dashed arrows in the phloem and xylem illustrate the transport directions, while those with an adjacent enzyme illustrate enzymatic reactions. Note: to simplify the diagram, no fate is shown for the short-lived signal NO. (Redrawn and updated from M. Becana and J.I. Sprent, Nitrogen fixation and nitrate reduction in the root nodules of legumes. Physiologia Plantarum. ©1987, with permission from Wiley.) (This figure is available in colour at JXB online.)

Legumes probably evolved in a nutrient-poor environment when the ability to fix atmospheric N would offer a considerable growth advantage. The high energetic cost of N fixation means that N-fixing legumes only have an advantage over other plants at low soil N supply. There is likely to be strong selection pressure for legumes which can quickly respond to changes in soil N supply to maximize the advantage of fixing atmospheric N. Since the onset of the Haber–Bosch process and the chemical production of N fertilizer, human activities have led to much elevated environmental N inputs, leading to extensive changes in natural ecosystems, for example grassland (Stevens et al., 2004). Elevated environmental N inputs may alter the legume–rhizobium symbiotic relationship and it has been proposed that this may cause less mutualistic rhizobia to evolve. For example, legumes inoculated with bacterial strains from N-fertilized plots produced significantly less biomass when compared with unfertilized controls (Weese et al., 2015). However, in most natural environments, legume roots compete with those of non-legumes, and this competition is unlikely to permit the build-up of soil N concentrations that drive this type of change in rhizobia. Furthermore, when legumes are grown next to crops, the mixed culture is likely to result in localized depletion of all N forms around the roots of both plants as they compete for nutrients. The transfer of N from a legume to an adjacently grown crop is likely to be mediated by microbes as the N pools in the shared rhizosphere of both types of plant turn over. The competition between root and microbial uptake of N restricts the concentration available to the legume root and there is no evidence to suggest that legume root N uptake systems are any less efficient than those of non-legumes (see later).

Legume nitrate reductase activity and N accumulation

Legume NO₃^– reduction has been described and previously reviewed (see Becana and Sprent, 1987) and here will only be briefly mentioned in relation to feedback signals for N status and inhibition of nodulation. Temperate legumes assimilate NO₃^– chiefly in the roots when growing under low N supply (Andrews, 1986), but the findings were species dependent and current model legumes (Lotus and Medicago) were not tested. Legume NO₃^– reductase activity and vacuolar storage of NO₃^– increase with external NO₃^– supply, and this substrate induction response is also well known in non-legumes. Vacuolar storage occurs in both roots and shoots; in barley the nitrate stores could be remobilized in 24 h (van der Leij et al., 1998). Legume NO₃^– reductase activity is important for the generation of nitric oxide (NO) as the enzyme mediates production of the signal from nitrite (Rockel et al., 2002). Nitrate reductase activity is found generally in the root and specifically in the nodule (Fig. 1), and nitrite can be measured in the nodule (Becana and Sprent, 1987). A role for reactive oxygen and reactive N species including NO has been demonstrated in early nodulation (Damiani et al., 2016).

One characteristic of legumes is the occurrence of specific N-containing molecules that are found in the phloem and xylem; these depend on the species of plant and include amino acids, ureides, and amides (see Fig. 1). For example, when non-nodulated soybean plants were treated with high concentrations of nitrate, there appeared to be higher concentrations of these compounds in the phloem (Vessey and Layzell, 1987). During nodulation of soybean and some other determinate nodulating legumes, the ureides allantoin and allantoic acid are the major N form exported, and can comprise up to 80% of N transported away from nodules (McClure and Israel, 1979; Streeter, 1979). The transporters for allantoic acid and allantoin transport out of the soybean nodule have been identified (Collier and Tegeder, 2012). Other types of determinate nodulating legumes and indeterminate nodulating legumes export the amides glutamine and asparagine. These molecules may feed back to inhibit root N uptake, fixation, and nodulation in legumes (Serraj et al., 1999; Vadez et al., 2000), possibly through their phloem concentrations (Parsons et al., 1993). Negative feedback regulation from downstream N metabolites is likely to be a common feature found in all types of plant.

Root N transporters and sensing of N supply

The ability of plant roots to sense external NO₃^– allows them to adapt to the dramatic fluctuations of available NO₃^– in the soil, from micromolar to millimolar concentrations (Miller et al., 2007). Modifying the NO₃^– sensing threshold of a legume is a strategy for bypassing nitrate inhibition of nodulation, allowing a normal number of N-fixing nodules to form at higher concentrations of external soil NO₃^–.

Root architecture responses to N

The ability of roots to adjust their architecture to changes in N supply is well known (reviewed by Walch-Liu et al., 2006; Kiba and Krapp, 2016) and this growth plasticity must be mediated by a sensing mechanism. Changes in root architecture can have both localized and systemic effects, and include hormonal changes that generate altered growth (Kiba and Krapp, 2016). High or low NO₃^– treatment (6 mM or 0.5 mM) to the entire root system inhibits or induces the initiation or growth of lateral roots, whereas in N-limited plants, there is a promotion of lateral roots in parts of the roots exposed to NO₃^–-rich patches (Robinson et al., 1999; Zhang et al., 2007; Walch-Liu et al., 2006; Ruffel et al., 2011). In legumes, split root experiments showed that NO₃^– inhibition of nodulation is a localized effect, limited to the treatment site (Hinson, 1975; Carroll and Gresshoff, 1983). However, in common with all plants, legumes have both localized and systemic signals which can modify root architecture (Mohd-Radzman et al., 2013).

Legume inorganic N transporters

Legume membrane transporter families for NO₃^– and ammonium corresponding to those of Arabidopsis have been described (Orsel and Miller, 2010). Legume ammonium transporters (AMTs) have been functionally characterized and their transport properties are very similar to those identified in Arabidopsis (D’Apuzzo et al., 2004). There are two types of AMTs in legumes and non-legumes, AMT1s for high affinity uptake and the AMT2-type transporters may be important in the recovery of ammonium lost from nodule cells by efflux (Simon-Rosin et al., 2003). Two families of plant NO₃^– uptake transporters, NRT1 and NRT2, have been identified. The NRT2s are all high affinity NO₃^– transporters, but the NRT1 family members can transport many different substrates including NO₃^–, amino acids, and peptides. The NRT1 family was renamed as the nitrate peptide family (NPF), but was recently shown to transport the plant hormones auxin, gibberellins, and abscisic acid (ABA) (Krouk et al., 2010; Kanno et al., 2012; Tal et al., 2016), making this name outmoded. The NRT1/NRT2 families were also described in the model legumes, including L. japonicus (Criscuolo et al., 2012). Lotus japonicus was found to have 92 NRT1 (NPF) genes and four NRT2 genes. The expression pattern of three NRT2-type transporters in another model legume species, M. truncatula, has been described (Pellizzaro et al., 2015). Some members of this family require a protein partner (NAR2 or NRT3) for function, and these were identified in legumes too (Tong et al., 2005).

In the non-legume Arabidopsis, one plasma membrane NO₃^– transporter, AtNRT1.1/CHL1/NPF6.3, is specifically associated with sensing and signalling NO₃^– availability. This dual transport and sensing function has led to the idea that the protein encoded by the NO₃^– uptake function gene is a transceptor—a membrane protein that belongs to a transporter family that also acts as a sensor/receptor (Gojon et al., 2011). The protein can transport both NO₃^– and auxin, and competition between the two substrates results in the redistribution of auxin, causing changes in lateral root growth (Bouguyon et al., 2015). Furthermore, phosphorylation at a key threonine residue (T101) in AtNRT1.1 switches the NO₃^– transport operating range from low to high concentrations. Phosphorylation makes the transporter operate in the high affinity range (µM) while dephosphorylation switches the protein to a low affinity transporter with a K_m for NO₃^– in the millimolar range (Liu and Tsay, 2003). This phosphorylation is carried out by the CBL-interacting protein kinase23 (CIPK23; Ho et al., 2009). An Arabidopsis atnrt1.1-deficient mutant exhibits reduced NO₃^– uptake, but also features major changes in gene expression including decreased expression of the high-affinity NO₃^– transporter gene NRT2.1, and failed to produce extra lateral roots in response to localized patches of higher NO₃^– concentation (Muños et al., 2004; Krouk et al., 2006; Remans et al., 2006; Ho et al., 2009). Furthermore, the phosphorylation state of T101 was shown to control the level of expression of NRT2.1 (Ho et al., 2009). The decreased expression of NRT2.1 was shown to be independent of NO₃^– assimilation, supporting a signalling role for NRT1.1 (Muños et al., 2004). Furthermore, a mutation in AtNRT1.1 that replaced a highly conserved P492 with a leucine residue resulted in a protein that was defective in both high and low affinity NO₃^– transport, but the signalling function was unaltered (Ho et al., 2009). A similar sensing mechanism may exist in legumes, and closely related homologues have been identified in N-fixing plants. Interestingly, legumes have two or three close relatives of AtNRT1.1 (Fig. 2A), but not all have the conserved T101 residue that is required to switch transport affinity (Fig. 2B). A role for auxin in the NO₃^– inhibition of nodulation has been demonstrated. When auxin is added with NO₃^–, the negative effect of NO₃^– on nodulation, and particularly Rhizobium infection, is partly offset (Streeter, 1988). Functional orthologues of AtNRT1.1 in legumes are good candidates to mediate the early localized sensing of NO₃^– that must be involved in the inhibition of nodulation.

Fig. 2.

AtNRT1.1 transceptor homologues in legumes. Sequence relationships between NRT1.1 and its close homologues. (A) A phylogenetic tree of Arabidopsis and Medicago truncatula. Medicago has three close homologues (paralogues), Medtr4g101380 (A), Medtr5g012270 (B), and Medtr5g012290 (C). AtNPF6.2 (AtNRT1.4) and AtNPF6.4 (AtNRT1.3) are low affinity nitrate transporters, AtNPF3.1 is a gibberellic acid transporter (Tal et al., 2016). (B) The amino acid residue T101 (underlined) important for signalling is conserved in Medicago NRT1.1a and NRT1.1b, but not in NRT1.1c. (This figure is available in colour at JXB online.)

One NRT1 family member, NIP/LATD, is involved in nodulation and lateral root growth, and transports NO₃^– at high affinity concentrations (250 μM); however, the knock-out mutant for this nodule-expressed NO₃^– transporter still showed NO₃^– inhibition of nodulation like the wild type (Bagchi et al., 2012). Another NRT1 transporter, MtNRT1.3 (MtNPF6.8), mediates the regulation of primary root growth by NO₃^– and can transport ABA (Pellizzaro et al., 2014). Members of this family of transporters could provide an early NO₃^– sensing system that can modify both root architecture and nodulation through a transceptor type of mechanism.

Legume signals for N status

Signals of plant N status must be downstream of and integrated with the primary sensors of soil N availability, like transceptors. Legumes have systemic signalling mechanisms to regulate the level of nodulation. Similar systemic signalling mechanisms are found in non-legumes and many were first identified in Arabidopsis.

Small peptides as signals for N status in nodulation

Legumes have long-distance systemic signalling mechanisms to regulate the level of nodulation (recently reviewed by Djordjevic et al., 2015; Araya et al., 2016). This is in part mediated by CLE peptides, which are induced by NO₃^– (Okamoto et al., 2009; Nishida et al., 2016) and inhibit nodule initiation when overexpressed (Mortier et al., 2010; Reid et al., 2011; Nishida et al., 2016). This effect is systemic and, for most CLE peptides with effects on nodulation, is dependent on their proposed receptor, the leucine-rich repeat receptor kinase HAR1/SUNN/NARK, to which CLE peptides have been shown to bind directly (Okamoto et al., 2009, 2013; Mortier et al., 2010; Saur et al., 2011; Nishida et al., 2016). Root Determined Nodulation 1 (RDN1) is expressed in the vascular cylinder and encodes a close homologue of Arabidopsis hydroxyproline O-arabinosyltransferase which arabinosylates CLE peptides (Ogawa-Ohnishi et al., 2013), and therefore may be required for their function. Mutations in HAR1/SUNN/NARK and RDN1 cause a hypernodulation phenotype that is resistant to NO₃^– inhibition (Caroll et al., 1985; Wopereis et al., 2000; Nishimura et al., 2002; Searle et al., 2003; Schnabel et al., 2005, 2011). However, when the phenotype was checked for the formation of new nodules, nitrate suppression of nodulation was found to be the same in the wild type, sunn, and rdn1 mutants (Kassaw et al., 2015). Intriguingly, grafting experiments show that the hypernodulation and NO₃^–-resistant nodulation phenotypes of these mutants are shoot determined (Buttery and Park, 1990; Hamaguchi et al., 1993; Delves et al., 1986; Krusell et al., 2002; Nishimura et al., 2002; Okamoto and Kawaguchi, 2015). However, although nodule numbers were not repressed at high N in plants with a har1 shoot and wild-type root, the nodules that formed were small and white, suggesting that nodule growth and development are separately regulated from nodule initiation (Okamoto and Kawaguchi, 2015). Earlier work supports this idea as it reported that SUNN was involved in systemic regulation of nodule initiation by N, but was not involved in N-mediated changes in nodule growth (Jeudy et al., 2010). Yet more support for the systemic action of these peptides was the recovery from shoot xylem sap of a mature arabinosylated form of an L. japonicus CLE peptide expressed in transgenic soybean roots (Okamoto et al., 2013).

Some inroads have been made in the understanding of the SUNN/HAR1 signalling pathway. SUNN expression is controlled by the tightly linked LSS (LIKE SUNN SUPERNODULATOR) locus in M. truncatula (Schnabel et al., 2010). Another receptor-like kinase KLAVIER, also required for N- and CLE-mediated mediated repression of nodulation, also acts in the shoot and interacts with HAR1 (Oka-Kira et al., 2005; Miyazawa et al., 2010). The F-box protein TOO MUCH LOVE (TML) was found to be needed downstream of the CLE peptides and HAR1 in the root for repression of nodulation (Magori et al., 2009; Takahara et al., 2013).

Recently, two more receptor-like kinases, CORYNE and CLAVATA2, were shown to interact with SUNN, and the coryne mutant displayed shoot-controlled hypernodulation (Crook et al., 2016). Together these studies suggest that HAR1/SUNN/NARK functions within a receptor complex to control nodule number or novel nodule initiation on the same root. The N-sensing components for this long-distance signalling mechanism remain unknown. Mutants in the gene corresponding to SUNN in soybean were tested in the field, and were found to be hypernodulated and to fix more N, but the plants were stunted (Suganuma et al., 2001). The authors concluded that this was due to high consumption of carbohydrates by the nodules, possible suggesting an imbalance in C/N metabolism in these plants. Some soybean hypernodulation mutants were shown to be less sensitive to NO₃^– inhibition of N fixation (Fujikake et al., 2003).

Another class of small peptides called CEPs (C-TERMINALLY ENCODED PEPTIDEs) are induced when N availability is low and can promote nodulation when overexpressed (Imin et al., 2013; Mohd-Radzman et al., 2016). Furthermore, in Arabidopsis, the LRR-RLK CEPR1 and its homologue CEPR2 were shown to be involved in CEP-mediated systemic NO₃^– signalling (Tabata et al., 2014). CEPR1 is expressed opposite the phloem pole in Arabidopsis roots, and was proposed to mediate lateral root initiation distally in the pericycle opposite the protoxylem (Roberts et al., 2016). It was recently confirmed in legumes that the CEP activity is dependent on the CEPR1 orthologue, CRA2 (Mohd-Radzman et al., 2016). Although specific CEP peptides with roles in nodulation have not yet been identified, loss of CRA2 leads to decreased nodulation and increased lateral root formation (Huault et al., 2014). The exaggerated lateral root formation phenotype is the opposite of that seen during CEP overexpression or when the peptide is added exogenously, and contrasts with the HAR1/SUNN/NARK phenotype. Furthermore, it was recently shown that CEP/CRA2 promotion of nodulation acts via the inhibition of ethylene signalling (Mohd-Radzman et al., 2016). It is therefore possible to speculate that the distal signal operating in CEPR1/CEP signalling in Arabidopsis is ethylene. In wild-type legumes, nodules form opposite the xylem pole, and this radial distribution pattern is determined by ethylene which is produced opposite phloem poles and negatively regulates nodule development (recently reviewed by Guinel, 2015). Further studies on CRA2 expression are needed particularly in the shoot (Huault et al., 2014), but it will be interesting to see whether a similar non-cell-autonomous mechanism may operate to control nodulation in legumes. Together the data demonstrate that N-regulated small peptides both positively and negatively act to regulate nodulation in legumes (Fig. 3).

Fig. 3.

Regulation of nodulation by N status. Nitrogen-replete legumes produce CLE peptides which systemically inhibit nodulation acting through the HAR1/SUNN receptors in the shoot. In this pathway. KLAVIER and TOO MUCH LOVE act downstream of HAR1 in the shoot and root, respectively. In low N conditions, CEP peptides signal via CRA2 which promotes nodulation in an ethylene-dependent manner. High ammonium levels inhibit rhizobial Nod factor biosynthesis (Wang and Stacey, 1990). Plant components are in green, rhizobial components in blue. Inhibited regulation is indicated by broken lines. (This figure is available in colour at JXB online)

Transcription factors and miRNAs in nodulation

Plant N status involves signalling via specific transcription factors (TFs) and miRNAs, and these may be important in nodulation. Several nitrate signals related to root architectural responses (see above) are mediated by specific TFs. For example, in Arabidopsis, local NO₃^– signalling through AtNRT1.1 has downstream ANR1, the MADS-box TF, controlling lateral root branching. Many other TF genes linking N status and root growth have been identified for Arabidopsis and they have been described recently (Guan et al., 2014). In legumes, the nodule inception protein (NIN) is needed for the initiation of nodule development and has DNA-binding and dimerization domains like TFs (Schauser et al., 1999; Borisov et al., 2003). In Arabidopsis, NIN-like proteins (NLPs) were shown to be TFs important for nitrate signaling (Konishi and Yanagisawa, 2013; Marchive et al., 2013). In the legume L. japonicus, NIN and NO₃^– were shown to act antagonistically to regulate gene expression (Soyano et al., 2015). However, a direct link between NLPs and plant N status in legumes has yet to be clearly demonstrated.

Another mechanism for signalling N status in legumes that can influence nodulation may be small RNA signals. In rice, miR444a has been shown to be linked to the N status of the plant and specifically to NO₃^– supply (Yan et al., 2014). Plant nutrient status can be associated with miRNA changes (reviewed by Paul et al., 2015). Legumes have been shown to have miRNAs in common with non-legumes, but also to have some that are specific, including some that are nodule localized (Turner et al., 2012). In soybean it was demonstrated that miR172 regulates the amount of nodulation that occurs (Yan et al., 2013). Furthermore, it was shown that an exogenous application of synthetic miPEP172c can stimulate actual miR172c expression, and this increased the number of nodules formed (Couzigou et al., 2016). The role of miRNAs in regulation of nodulation is clear, but a link with nodulation and N status has yet to be established for legumes.

Conclusions and future prospects

The N-sensing response of legumes is complicated by the different time scales for N responses; clearly the very early localized response to nitrate is likely to be mediated by membrane proteins such as the NPF transceptors. Systemic responses appear to be mediated by small peptides that are perceived by a family of LRR-receptor kinases. The mechanism for translating this sensing response into developmental outcomes is linked to the redistribution of hormones such as auxin, cytokinins, gibberellins and ABA. In this manner, N transporters have the potential to provide a suite of responses that can include the inhibition of nodulation and nodule activity. Reviewing the literature published around this topic shows that very few papers have actually determined N fixation activity, and acetylene reduction is used as proxy measurement. A definitive study showing the time course and concentration required for NO₃^– inhibition of nodulation activity is needed. For example, how good is the evidence that nodule repression occurs in deeper soil layers and less in the top soil as NO₃^– fertilizer leaches through the profile? Soil water supply changes the NO₃^– concentration at the root surface and may help to explain the well-documented water stress inhibition of nodulation (Gil-Quintana et al., 2013). Indeed the soil supply of both water and NO₃^– may be sensed by the same mechanism. Novel sensors for the sensitive detection of NO₃^– (Ho and Frommer, 2016) will be a useful tool for dissecting this relationship. The formation of new nodules on legumes must be sensitive to the plant N status signals, such as peptides and miRNAs, but also includes the metabolic negative feedback signals such as the products of N assimilation. For example, there is evidence that the amino acids, glutamine and asparagine negatively regulate CEP1 expression (Imin et al., 2013). For nodulation and fixation to occur it is important to remember there are actually two partner organisms and the symbiotic bacteria may have a parallel signalling network that is N responsive and important for N fixation in nodules. For example, the flavonoid signals produced by roots depend on the N supply to the plant (Coronado et al., 1995). The crosstalk between the host plant and bacteria may be influenced by the N status of the plant, and this topic is poorly described. From studies with hypernodulation mutants. we know that a C/N imbalance can cause problems for the plant and growth can be impaired (Suganuma et al., 2001). Altering the N-sensing capacity of a legume could result in a similar C/N imbalance stunted phenotype, with active N-fixing nodules under high soil N supply, causing the carbon supply from photosynthesis to limit growth. The prospects for breaking the link between legume N fixation and soil N levels are good as the regulatory controls are becoming well characterized, but both sensing and N status signals will need to be addressed.

Acknowledgements

The authors are supported by grant funding (BB/JJ004553/1 and BB/L010305/1) from the UK Biotechnology and Biological Sciences Research Council and the John Innes Foundation.

Author notes

*

Correspondence: [email protected] or [email protected]

Editor: Michael Considine, University of Western Australia