NLP1 binds the CEP1 signalling peptide promoter to repress its expression in response to nitrate
Nitrogen (N) is an essential macronutrient for plant growth and development, which is often growth-limiting in agricultural systems. Nitrate not only is the main source of N for most plants but also acts as a molecular signal to regulate gene expression and root development (Liu et al., 2015). As immobile organisms, plants cannot choose but can adapt to their surrounding environment. Therefore, plants have evolved sophisticated strategies for N acquisition in relation to source availability, to coordinate plant growth and development in response to a fluctuating external supply (Oldroyd & Leyser, 2020). Regulatory mechanisms for nitrate acquisition comprise both local and systemic long-distance signalling pathways that inform on the intrinsic nutritional status across the whole plant (Alvarez et al., 2012; Bellegarde et al., 2017; Gautrat et al., 2021).
Legume plants can symbiotically associate with rhizobium soil bacteria to form root nodules and fix atmospheric N2 for the benefit of the host plant. This is, however, an energy-consuming process that is thus inhibited by the presence of mineral N in soils, including nitrate and ammonium (Streeter & Wong, 1988; Carroll & Mathews, 1990). Studies in the model legume Medicago truncatula and Lotus japonicus revealed that the nodule inception (NIN)-like proteins, NLP1 and NLP4, are critical for the nitrate inhibition of nodulation (Lin et al., 2018; Nishida et al., 2018, 2021). Nitrate triggers MtNLP1 translocation from the cytosol to the nucleus, where MtNLP1 interacts with MtNIN to repress target gene expression (Lin et al., 2018). In L. japonicus, the related LjNLP4 transcription factor binds the promoter of the LjCLE-RS2 (CLAVATA-Like Root Signal 2) signalling peptide-encoding gene on nitrate-responsive elements (NREs) to promote its expression, leading to a systemic repression of nodulation through the shoot LjHAR1 (hypernodulation and aberrant root) receptor (Nishida et al., 2018). A similar mechanism is conserved in M. truncatula, where MtNLP1 promotes the expression of the MtCLE35-related gene, through a direct binding to its promoter on evolutionarily conserved NREs (Luo et al., 2021; Mens et al., 2021; Moreau et al., 2021).
In response to nitrate deficiency, the expression of another family of signalling peptides is induced, the C-terminally encoded peptides (CEPs), which in Arabidopsis act systemically from roots to shoots through the CEPR1 and CEPR2 receptors (Tabata et al., 2014). In M. truncatula, the overexpression of MtCEP1 or exogenous applications of MtCEP1 peptides both negatively regulate lateral root formation and promote nodulation through the MtCRA2 (Compact Root Architecture 2) receptor, closely related to AtCEPR1/2 (Imin et al., 2013; Huault et al., 2014; Mohd-Radzman et al., 2016; Laffont et al., 2019; Gautrat et al., 2020). Knowing that the expression of most MtCEP genes is promoted by a N deficiency, we hypothesized that, in addition to potential yet unknown N deficiency regulations that would promote CEP gene expression, MtNLP1 may be involved in the suppression of their expression in response to nitrate. In this study, we indeed demonstrate using different complementary approaches that MtNLP1 binds to the MtCEP1 promoter through specific half NRE (hNRE) motifs to repress its expression in the presence of nitrate.
As a first step, we analysed whether the MtCEP1-promoting effect on nodulation was affected in the nlp1 mutant. To this end, we applied synthetic MtCEP1 peptides in wild-type and nlp1 mutant plants and quantified nodule number in the absence or presence of nitrate (Supporting Information Methods S1). Whereas the nlp1 mutant nodulation was less sensitive to the nitrate inhibition than the wild-type nodulation, MtCEP1 peptides similarly increased nodule number in wild-type and nlp1 mutant roots either in the absence or in the presence of nitrate (Fig. 1a,b). This suggests that MtCEP1 may act downstream of MtNLP1 to modulate nodule number according to nitrate availability.
To investigate how MtCEP1 expression is repressed by nitrate (Imin et al., 2013), we tested whether the nitrate-induced suppression of MtCEP1 expression was dependent on MtNLP1. RT-qPCR analysis of MtCEP1 transcript levels in nlp1 mutants grown with or without nitrate first revealed slightly lower levels in nlp1 than in wild-type roots in the absence of nitrate (Fig. 1c). Interestingly, the nitrate repression of MtCEP1 in wild-type roots was not observed in nlp1 mutant roots (Fig. 1c). Similar results were obtained for the nitrate repression of other CEP genes, with MtCEP2, MtCEP6 and MtCEP7 being fully dependent on MtNLP1, but MtCEP4 being only partially, and MtCEP5 expression being even higher in nlp1 (Fig. S1). This result suggests that MtNLP1 is required for the nitrate-induced suppression of the expression of different CEP genes and that other NLPs (e.g. NLP4) (Lin et al., 2018), or even other unknown pathways, are involved. To further validate that nitrate represses CEP1 gene expression through MtNLP1, CEP1 transcript levels were analysed in roots overexpressing MtNLP1 (pUb-NLP1) and treated with KCl, as a control, or KNO3, 10 mM each. First, the expression of MtCEP1 was increased in pUb-NLP1 control roots in the absence of nitrate; and second, the repression of MtCEP1 expression by nitrate was stronger in MtNLP1-overexpressing roots than in control roots (Fig. 1d). To independently validate the link between MtNLP1 and MtCEP1 expression, a pCEP1-GUS transcriptional fusion was expressed in wild-type or nlp1 mutant roots. Nitrate suppressed the GUS signal in wild-type roots, whereas it was barely acting in nlp1 mutant roots (Fig. 1e). These results independently demonstrated that MtNLP1 is required for the nitrate repression of at least MtCEP1 expression.
As a next step, we wanted to determine whether MtCEP1 could be directly repressed by the MtNLP1 transcription factor. We first used a dexamethasone (DEX)-inducible system, where MtNLP1 is fused to the glucocorticoid receptor (NLP1-GR) and constitutively expressed thanks to a ubiquitin promoter, to transiently induce MtNLP1 expression in M. truncatula roots. The expression of MtCEP1 was reduced within 4 h of DEX treatment, followed by a steady reduction in transcript levels after a 24-h treatment, indicating that MtNLP1 expression is sufficient on its own to repress MtCEP1 expression. Combining DEX together with the cycloheximide (CHX) protein synthesis inhibitor did not suppress the DEX-inducible expression (Fig. 2a), demonstrating that de novo protein synthesis was not required for the repression of MtCEP1 expression. This result suggests that MtNLP1 might indeed directly target the MtCEP1 promoter to repress its expression. A dual-luciferase reporter assay was then used in Nicotiana benthamiana leaves to further analyse whether NLP1 was sufficient to repress the MtCEP1 transcriptional activity. The co-expression of a pCEP1-LUC transcriptional fusion with a p35S-NLP1 or an EV (empty vector), as a control, revealed that MtNLP1 can repress the activity of the MtCEP1 promoter in response to nitrate (Fig. 2b). A CEP1 promoter deletion analysis was then subsequently carried out using this dual-LUC assay in N. benthamiana leaves. The MtNLP1 repression of MtCEP1 expression was abolished when a 220-bp fragment (−1674 to −1426) of the promoter was deleted (Fig. 2c,d).
As NLPs were shown to bind nitrate-responsive cis-elements in the promoter of different nitrate-inducible genes (Konishi & Yanagisawa, 2013; Soyano et al., 2015; Nishida et al., 2018), we then searched NRE motifs in this 220-bp region (−1674 to −1426 bp) of the MtCEP1 promoter. A NRE motif was identified (Fig. 2e,f), and when comparing its sequence to those previously identified in promoters activated by NLPs (Nishida et al., 2018; Moreau et al., 2021; and this study), we observed that this CEP1 promoter NRE motif associated with a MtNLP1-repressive activity was shorter than for upregulated genes (e.g. NREs in LjCLE-RS2 and MtCLE35 promoters; Nishida et al., 2018; Luo et al., 2021), corresponding to only half of the initially described palindromic motif (Fig. 2e). Similar hNRE motifs were present in promoters of other M. truncatula CEP genes repressed by MtNLP1, and even in CEP gene promoters from Arabidopsis (Fig. S2), suggesting an evolutionary conservation. We then performed a chromatin immunoprecipitation (ChIP) assay to test whether this region is enriched for MtNLP1 binding. A MtNLP1-GFP translational fusion was expressed in M. truncatula roots, and using a GFP antibody, an enrichment for MtNLP1 binding was detected in the region of the MtCEP1 promoter containing this hNRE motif (Fig. 2g). We additionally showed that MtNLP1 binds the MtCLE35 promoter in two regions containing full-size palindromic NREs (Fig. 2e–g), in agreement with previous results obtained in L. japonicus, knowing that MtCLE35 is most closely related to LjCLE-RS2 (Luo et al., 2021; Mens et al., 2021; Moreau et al., 2021). This MtCLE35 result can thus be considered as a positive control for the MtNLP1 ChIP assay. The binding of MtNLP1 to the hNRE within the MtCEP1 promoter was additionally examined using a yeast one-hybrid assay. In this heterologous system, MtNLP1 indeed binds to the region containing the hNRE1 motif, and its deletion (∆hNRE, −8 bp) is sufficient to reduce this MtNLP1 binding activity (Fig. 2h). Finally, a deletion of this hNRE cis-element in the pCEP1 : GUS transcriptional fusion (pCEP1∆hNRE-GUS) revealed that in M. truncatula roots that the nitrate repression of the pCEP1:GUS activity was barely detectable (Fig. 2i,j).
Overall, we can now propose a refined model (Fig. 2k) by combining these new results with previous studies, showing that MtNLP1 not only activates MtCLE35 expression to mediate the nitrate inhibition of nodulation (Luo et al., 2021; Moreau et al., 2021; and this study) but also acts as a repressor by binding the promoter of CEP genes to inhibit their expression in the presence of nitrate, at least partially explaining the higher expression levels of CEP genes in N deficiency. MtNLP1 thus acts as a bifunctional transcription factor, and interestingly, instead of the classical full-size NREs found in nitrate/MtNLP1-activated genes, hNRE motifs were retrieved in the promoter of MtCEP genes that are repressed by MtNLP1. Some other transcription factors with such a bifunctional transcriptional activity exist, such as WUSCHEL that acts both as a transcriptional repressor and as an activator according to the tissues considered (Ikeda et al., 2009). Within the NLP family, AtNLP7 was proposed to act both as a transcriptional activator and as a repressor, potentially through the recruitment of different partners (Marchive et al., 2013). Accordingly, LjNLP1/4 could repress the expression of symbiotic-specific genes such as NF-YA/B, EPR3 and RinRK1 (Nishida et al., 2021). The NIN symbiosis-specific transcription factor has both positive and negative functions in nodulation and is active in different root tissues required for nodulation (epidermis, cortex and pericycle), which opens the possibility of a differential transcriptional activity according to cell types and recruited partners (Marsh et al., 2007; Liu et al., 2019).
Half NRE motifs were identified in different CEP gene promoters both in M. truncatula and in Arabidopsis, suggesting an evolutionary conservation of the NLP-dependent repression of these signalling peptide-encoding genes. Recently, preferential binding motifs were proposed for nitrate-related NLPs vs the symbiosis-specific NIN family member (Nishida et al., 2021). We now highlight here that binding motifs may be different according to the active vs repressive action of the NLP transcriptional complex. Besides such differential activating or repressing NLP binding motifs, the recruitment of partners and/or of diverse NLP post-translational modifications is likely involved. Interestingly, the CEP7 promoter does not contain any hNRE motif, but has instead a NIN binding site, which can be activated by NIN, as previously characterized in Laffont et al. (2020). Indeed, CEP7 is a very specific member of the CEP gene family in which expression is induced by rhizobium through the NIN transcription factor. This suggests that the CEP7 repression by nitrate is similar to the regulation of other rhizobium-inducible NIN target genes, for example EPR3, NF-YA and NF-YB, which are repressed by nitrate in an NLP1-dependent manner through an interaction with the rhizobium-induced NIN transcription factor (Lin et al., 2018; Nishida et al., 2021). In Arabidopsis, it was recently shown that an HBI1 (homolog of brassinosteroid enhanced expression2 interacting with IBH1)–TCP20 (Teosinte Branched 1, Cycloidea, PCF (TCP)-domain family protein 20) transcription factor complex positively regulates CEP gene expression (Chu et al., 2021), and interestingly, TCP20 can also interact with NLP6/7 to mediate N availability responses (Guan et al., 2014). Thus, it will be interesting in the future to identify such additional NLP partners compared with NIN, as well as potential post-translational modifications, ideally at a cell-type-specific level, to have a detailed mechanistic understanding of their regulatory functions in legume nitrate responses, including the regulation of symbiotic nodulation.
Acknowledgements
We thank Yongrui Wu (CAS Center for Excellence in Molecular Plant Sciences, China) for providing the dual-LUC system vector, and Jeremy Murray (CAS Center for Excellence in Molecular Plant Sciences, China) for providing yeast one-hybrid constructs and for helpful discussions. The FX team thanks the CAS Project for Young Scientists in Basic Research (YSBR-011) and National Natural Science Foundation of China (NSFC 31670242) agencies for funding. The FF team thanks the ‘Agence Nationale de la Recherche’ (ANR) project ‘PSYCHE’ and the ‘Ecole Universitaire de Recherche’ (EUR) Saclay Plant Sciences (SPS) for funding.
Author contributions
FX, ZL and FF designed the work, analysed the data and drafted the manuscript. ZL and JW performed most of the experiments and analysed the data. CM performed the ChIP-qPCR experiments, contributed to Figs S1, S2 and analysed the corresponding data.
