An obvious level of investigation is the level of the organs specialized in the N acquisition (roots and nodules). Nevertheless, a possible control by the N demand of the whole organism relations between these organs and the other parts of the plant has to be considered. Experimentally, split root systems allow one: (1) to evaluate the effect on the whole plant of short term fluctuations of N availability applied on a localized part of the roots; and (2) to evaluate the systemic responses triggered at the whole plant level that are revealed on the untreated side. The responses of roots normally supplied with various nitrogen sources (1 mM NO3−, 1 mM NH4+, or air) to contrasted level of N supply on the other side have been investigated. Three types of N nutritional situations were compared referred as “Control”, “N-sufficient” and “N-limited”. In “Control plants” both sides of the split-root were normally supplied with 1 mM NO3−, 1 mM NH4+ or air. In “N-sufficient” plants one side of the root system was supplied with a high level of nitrogen (10 mM NH₄NO₃) whereas in “N-limited” plants the N source was absent of this compartment (air was substituted by Argon/O₂ in case of nodulated plant). Short term responses (4 days) affecting the acquisition machinery (NO3−, NH4+, uptake and symbiotic N₂ fixation) were studied at first.

Nitrate, ammonium and N₂acquisition are repressed by systemic signaling: The addition of high level of NH₄NO₃ to one side of the root system induced a strong repression of the acquisition capacity of the other side whatever the nature of the N source (NO3−, NH4+, air) at that side (Fig. 1). This indicates that: (i) the three pathways involved in N acquisition are under feedback repression mechanism(s); and (ii) these feed back mechanism(s) involve systemic signaling pathways that adjust the acquisition capacities of the roots to the N demand of the whole plant. Parts of the regulatory mechanisms may be common for the three acquisition pathways.

Molecular responses associated with these physiological responses have been characterized by a transcriptomic approach: To further investigate the molecular basis of the short term control of root N acquisition to variations of N status, transcriptomes of roots and shoots of N-limited and N-sufficient plants supplied with NO3−, NH4+, or air were compared using affymetrix chips and real time Q-PCR platform. Gene lists of differentially expressed genes have been identified for the three sources [49] (Fig. 2). In roots, important groups of these genes are coding proteins related to the specific acquisition machinery of roots or nodules (transporters, assimilatory enzymes). Some are related to root or nodule development suggesting that both the function and the structures are under the control of systemic N signaling. Despite a similarity of the responses of the three pathways involved in N acquisition to these treatments at the physiological level, the transcriptomic study has revealed that at the molecular level genes networks involved in these responses in root and in shoots are rather specific of the N source present in the environment of the roots. Of course this does not rule out the possibility that some elements involved in the sensing of the plant N status might be common for the three N sources status and that these elements constitute potential key molecular components of the systemic signals regulating root N intake. All together these results may suggest that the plant response may be the result of the interaction of two signaling pathways (presumably systemic) related to the nature of the N source and to the level of downstream N metabolites respectively that may determine the response of plant roots and leaves to N availability modulation.

When plants experience a local N limitation, onlyNO3−acquisition is able to display an efficient adaptive response that may efficiency compensate the N deficit: Upon a short (i.e. 4 days) NO3− limitation on one side of split root systems, nitrate fed plants maintain their overall N acquisition and growth. This compensation for a temporary N deficit is mediated by the stimulation of the NO3− uptake affinity and capacity of the other side of the root system still fed with NO3− (Fig. 3). This short term response is easily interpreted as the result of a release of the feed back repression exerted by N metabolites. However, despite the fact that such mechanism exists for the three types of N acquisition intriguingly only NO3− fed plants display this capacity to compensate a local N limitation. The consequence is important for the plant because only NO3− acquisition may allow the whole plant to maintain its growth under such conditions, whereas in NH4+, N₂ fixing plants, local N limitation results rapidly in a decrease of the overall N acquisition of the plant and therefore limits plant growth. As NH4+ and N₂ fixation require important C metabolites flux to the roots to be assimilated, a logical cause might have been a limitation of compensatory response by carbon availability. However, sucrose feeding has ruled out this hypothesis, suggesting that the absence of response was presumably due to an overall limitation of the NH4+ uptake and N₂ fixation capacities of the plant roots that are not as efficient as in the case of NO3−.

Molecular responses identified by the transcriptomic approach are associated to the functional response of the organs: Representative subpopulations of root transcripts that have been identified in our transcriptomic studies (comparison between N-sufficient vs N-limited plant) were selected and studied in response to local N limitation (comparison between “Control” and “N-limited” plants). Interestingly, the differential expression of these genes is well correlated to the adaptive response observed in those plants: only in the case of NO3− fed plants a large proportion (>50%) of the genes identified as responding to systemic repression signals were found also differentially expressed in response to local N limitation. In contrast, in the case of NH4+ or N₂ fed plants that do not display functional response, most of the genes identified as regulated by the systemic signal (N-sufficient vs N-limited) comparison did not respond to local N limitation [49].

A second level of integration is to take into account developmental responses that may affect (both in time and space) structures involved in N acquisition (nodules and roots). Although response involving initiation of new structures may be initiated very rapidly after beginning of N shortage, to be efficient they generally require more time than response modulating the activity of pre-existing structure. Therefore N-limitation treatments were applied on NO3− fed and N₂ fed plants in split root systems on a larger time scale than before (up to 16 days) in order to potentially reveal these responses. The long-distance effects of N limitation were studied by monitoring N intake fluxes (¹⁵N labeling), root and nodule morphometry, nodule initiation (by the use of pENOD11::GUS reporter gene), C allocation within the plant (¹³C labeling).

In contrast with nitrate fed plants, N₂ fixing plants lack the ability to trigger a fast up-regulation of N₂ fixation. Our results indicate that in both NO3− fed and N₂ fed plants, N limitation triggers a systemic signaling which results in developmental response that may compensate the N deficit. In NO3− fed plants, roots of N-limited plants strongly proliferate in the nitrate containing compartment after 14 days (as compared to nitrate fed roots of control plants) to the expense of N-limited roots that stop growing. Together with the up regulation of the uptake capacities of these roots that was described above, this response leads to a very efficient compensation of N limitation, that allows the N-limited NO3− fed plant to maintain their growth at the same level as the control NO3− fed plants. Nevertheless, although less efficient as in the case of N fed plants, a long term response relying on long term plasticity of nodule development can be revealed in N₂ fed plants exposed to local N limitation. On the untreated roots, this response involved firstly the expansion of pre-existing nodule and secondly the initiation of new nodules demonstrating that both late and early stages of nodule development are targeted by systemic signals related to the N status of the plant. Interestingly the use of the Mt sunn hypernodulating mutant indicated that both responses may be mediated by distinct pathways, the first one being probably sunn independent whereas the second one sunn dependent. Although the mutant displays a higher number of nodules, they have a lower level of fixation activity (overall N₂ fixation capacity of the plant being at the level of the wild type). Surprisingly, unlike the wild type, when the mutant is exposed to N limitation it is able to up-regulate rapidly nodule N₂ fixation specific activity and compensate partially the N deficit suggesting that the presence of a large number of pre-existing nodule structures may allow this rapid response.

Based on our results indicating that nodule development was a major target for systemic N signaling a transcriptomics approach has been initiated in Mtr to identify: (i) genes associated with competence to nodulate (“competence program”), i.e. that are turned on or off in nitrogen-starved plants and allow the plant to activate symbiotic signaling with Rhizobia; and (ii) regulatory genes that control early stages of nodule development. The competence program is being analyzed with split root systems, using both wild type and sunn mutant, in which competence to nodulate is not affected by the plant N nutritional status. To identify genes controlling nodule development, we have exploited results of microarray analyses that allowed us to identify 185 regulators of gene expression up-regulated in Mtr nodules. To investigate their involvement in early nodulation steps we have set up a sensitive assay, based on quantitative RT-PCR on 384 plates, to study their expression in a series of wild type and early symbiotic mutants. We were thus able to identify regulatory genes associated with various symbiotic stages, from Nod factor signaling to Rhizobium infection and nodule primordium formation. The next step is to examine the impact of the nitrogen status on the activation of these genes.

As quoted above, N acquisition necessitates a large flow of carbon compounds both at the functional level (for respiration and amino acid synthesis) and for the development of root and nodules structures. As such, below ground structures can compete for assimilates with other plant sinks and carbon supply to nodules could become a limiting factor of N acquisition. In fact, in our experiment we have demonstrated that preferential carbon allocation to the untreated nodulated roots of N-limited plant is a very early event (detected after only 4 days of N deficiency) that may be revealed before any significant change either in N acquisition capacity or in nodule development. This may suggest that such event may be part of the early mechanisms leading to the developmental response.

The acquired knowledge has been aggregated in an ecophysiological model that allows simulating the relations between leaf area and N retrieval by roots. This model was developed to explain how nitrate availability in the nutritive solution and the quantity of incident photosynthetic active radiation, two main environmental factors affecting C and N acquisition, affect N retrieval by the plants. Because of the identification by the model of either variables or physiological parameters that could be associated to genetic variability, the model would furnish an analysis grid to characterize and analyze the genotypic diversity of plant N nutrition.

The model has been established using A17 line (cv. Jemalong), for the vegetative period, using data arising from two greenhouse experiments in controlled conditions concerning two environmental factors. The model takes into account symbiotic nitrogen fixation throughout the vegetative period and its modulation by nitrate availability. The proposed model [50] is structured around four integrative variables that characterize plant growth: plant biomass, below ground biomass, amount of plant N and projected leaf area. These integrative variables were decomposed in simpler intermediates variables allowing to hierarchy and quantify the impact of the environmental factors studied on plant growth. These intermediates variables are predicted by the model using response curves to the environmental factors.

The analytical approach described herein has allowed characterizing the adaptative strategy of the plants in response to both nitrogen limited nutrition and carbon limited nutrition. Under nitrogen limited conditions plants reduced their rate of structure set up (leaf area) while plant functioning was little affected. Interestingly a major impact of C nutrition was observed on model parameters driving both structure set up (leaf area and below ground biomass) and functioning (conversion of incident photosynthetic active radiation in biomass, specific N retrieval and efficiency of N conversion to projected surface), suggesting a strong adaptative response to carbon limited nutrition. This very significant result demonstrates clearly that N nutrition cannot be analyzed independently from the interaction between nitrogen and carbon nutrition [51]. At the plant level, when plants are N deficient, they still maintain their specific photosynthetic activity (i.e. carbon dioxide assimilation per unit of foliar surface) but they reduce their leaf expansion. This hence constitutes a relevant phenotypic target for detecting any contrasted N nutrition among various genotypes.

The model provides valuable support for assisting the detection of genetic variants affected for nitrogen uptake, including soil mineral nitrogen retrieval by the roots and symbiotic N₂ fixation by root nodules. The key variables (in particular leaf area) identified in the preceding steps were used as criteria for ranking genotypes according to their diversity in their capacity to retrieve N from different sources in a recombinant inbred lines population (LR4 Jemalong X DZA315.16 consisting of 175 inbred lines of Mtr), grown with contrasted nitrate availabilities in the nutritive solution. Plant leaf area was measured at a given date to discriminate genotypes according to their capacity to take up N while the estimation of their nitrogen nutrition index [52] allowed checking if the genotypes were effectively differentially taking up nitrogen. Leaf area at a given date was not a robust trait to highlight differences in N uptake among the genotypes. When measured dynamically, leaf area allowed calculating relative expansion rate which was found to be a good indicator for ranking genotypes according to their ability to uptake nitrogen in given environmental conditions affecting carbon and nitrogen acquisition [51] (Fig. 4).

This model was also a valuable tool for comparing the functional response to variations in N availability of a root architecture mutant (TR185) with its wild type Jemalong (WT). The mutant TR185, displaying highly branched roots and scarce small nodules phenotype, was selected after γ-ray mutagenesis on Jemalong [32]. In our study, whatever the NO3− regimes, the mutant TR185 displayed a higher number of lateral roots than WT but its leaf area was depressed (Fig. 5, A–B). Using the previous model, it was demonstrated that (i) although the mutant has a reduced growth (leaf area) as compared with WT, both genotypes had similar C assimilation and partitioning of biomass to roots; (ii) the mutant and WT had differential N acquisition per amount of root biomass (data not shown) but similar efficiency of N conversion to leaf area (Fig. 5C). As such, the model helped to conclude that the two genotypes contrasted for their root structure only differed functionally in their N acquisition capacity per amount of roots. In parallel with what was described for Sunn, a comparative transcriptomics study between TR185 and WT is underway to better target the genes to be used in reverse genetic strategies and create a new valuable genetic variability.

Therefore, the model developed furnishes a grid to analyse a spontaneous or induced genetic variability for plant N nutrition and to detect valuable candidate genes which may improve N acquisition and conversion efficiency, by dissecting very integrative phenotypic variables into more simple variables. In parallel with what was described for Sunn, a comparative transcriptomics study between TR185 and wild type Jemalong is underway to better target the genes to be used in reverse genetic strategies and create a new valuable genetic variability. The model proposed in this work could so constitute an innovative tool for diagnosing the differential origin between lines in their capacity to retrieve N [52].

This study was conducted using various expertise at different levels: nodule development, regulation of NO3− and NH4+ absorption, genetic of nodulation, acquisition of N and C and its modeling. Mechanistic molecular approaches were combined with genetic and ecophysiological approaches to characterize the physiological process involved at the level of the plant. A special attention was given to integrative mechanisms which allow adjusting nodulated root functions with the nitrogen demand of the plants.

In the literature, many studies analyzed the legume N₂ fixation reaction to elevated mineral nitrogen, whereas the plant adaptation to N source deprivation, a stress with probable occurrence in agronomic situations, has received much less attention. This knowledge is particularly needed for agronomic legumes because of its potential impact on their yield and stability.

Our work help to demonstrate that nitrogen acquisition capacity is adjusted by the N nutritional status of the plant through systemic signaling. We unraveled the components of the adaptative strategy of legume plants faced to modulation of N availability: it occurs on the short term in the case of nitrate fed plants (via both functional and developmental responses) and on the longer term (with a crucial role of nodule organogenesis as target of the plant N status) with N₂ fixing plants. It was shown that C metabolites allocated to “compensating” roots (i.e. those still exposed to the N resource upon N starvation on localized part) was very early and instrumental of the plant adaptative response, far before any functional or developmental regulation could be detected.

The use of ecophysiological and genetic tools allows characterizing the physiological processes involved and the underlying molecular determinants. Such tools allow to formalize through modeling the strategy of the plant response to its environment and to identify the molecular basis of genes involved in the functions described by the model. It can also lead to the formulation of modeling tools for high throughput phenotyping. Transcriptomics analysis helped identifying genes whose expression depend upon plant N status. Such genes could so constitute markers indicators of the plant N status and help in phenotyping on large plant population early adaptative responses to situations under which plant encounters N shortage.

The work here presented demonstrates that ecophysiological tools help to characterize a genetic diversity related to plant N nutrition. The first version of the model is based on trophic parameters related to the C and N acquisition and allocation. Root and shoot architecture variables should now be implemented in this model. Ultimately, the model may also integrate the effect of some major genes involved in N nutrition incorporating the contribution of particular alleles.

On the basis of germplasm collection genotyping, or RNAi or TILLING strategies, modelling work offers an innovative strategy to guide the enlargement of the genetic variability for particular candidate genes. It may finally help to propose target ideotypes and valuable alleles to breeders.