1. Plant Biology
Download icon

Epidermal LysM receptor ensures robust symbiotic signalling in Lotus japonicus

Research Article
  • Cited 0
  • Views 578
  • Annotations
Cite as: eLife 2018;7:e33506 doi: 10.7554/eLife.33506

Abstract

Recognition of Nod factors by LysM receptors is crucial for nitrogen-fixing symbiosis in most legumes. The large families of LysM receptors in legumes suggest concerted functions, yet only NFR1 and NFR5 and their closest homologs are known to be required. Here we show that an epidermal LysM receptor (NFRe), ensures robust signalling in L. japonicus. Mutants of Nfre react to Nod factors with increased calcium spiking interval, reduced transcriptional response and fewer nodules in the presence of rhizobia. NFRe has an active kinase capable of phosphorylating NFR5, which in turn, controls NFRe downstream signalling. Our findings provide evidence for a more complex Nod factor signalling mechanism than previously anticipated. The spatio-temporal interplay between Nfre and Nfr1, and their divergent signalling through distinct kinases suggests the presence of an NFRe-mediated idling state keeping the epidermal cells of the expanding root system attuned to rhizobia.

https://doi.org/10.7554/eLife.33506.001

eLife digest

Microbes – whether beneficial or harmful – play an important role in all organisms, including plants. The ability to monitor the surrounding microbes is therefore crucial for the survival of a species. For example, the roots of a soil-growing plant are surrounded by a microbial-rich environment and have therefore evolved sophisticated surveillance mechanisms.

Unlike most other plants, legumes, such as beans, peas or lentils, are capable of growing in nitrogen-poor soils with the help of microbes. In a mutually beneficial process called root nodule symbiosis, legumes form a new organ called the nodule, where specific soil bacteria called rhizobia are hosted. Inside the nodule, rhizobia convert atmospheric dinitrogen into ammonium and provide it to the plant, which in turn supplies the bacteria with carbon resources.

The interaction between the legume plants and rhizobia is very selective. Previous research has shown that plants are able to identify specific signaling molecules produced by these bacteria. One signal in particular, called the Nod factor, is crucial for establishing the relationship between these two organisms. To do so, the legumes use specific receptor proteins that can recognize the Nod factor molecules produced by bacteria. Two well-known Nod factor receptors, NFR1 and NFR5, belong to a large family of proteins, which suggests that other similar receptors may be involved in Nod factor signaling as well.

Now, Murakami et al. identified the role of another receptor called NRFe by studying the legume species Lotus japonicus. The results showed that NFRe and NFR1 share distinct biochemical and molecular properties. NRFe is primarily active in the cells located in a specific area on the surface of the roots. Unlike NFR1, however, NFRe has a restricted signaling capacity limited to the outer root cell layer. Murakami et al. found that when NRFe was mutated, the Nod factor signaling inside the root was less activated and fewer nodules formed, suggesting NRFe plays an important role in this symbiosis.

NFR1-type receptors have also been found in plants outside legumes, which do not form a symbiotic relationship with rhizobia. Identifying more receptors important for Nod-factor signaling could provide a basis for new biotechnological targets in non-symbiotic crops, to improve their growth in nutrient-poor conditions.

https://doi.org/10.7554/eLife.33506.002

Introduction

Perception of Nod factors by LysM receptor kinases, NFR1 and NFR5 in Lotus japonicus (Broghammer et al., 2012), triggers tightly coordinated events leading to root nodule symbiosis (Madsen et al., 2003; Radutoiu et al., 2003). Minutes after the activation of receptors, a signalling cascade (Stracke et al., 2002; Antolín-Llovera et al., 2014) leading to regular calcium oscillations in the root hair cells located in the susceptible zone is initiated (Miwa et al., 2006). These oscillations are interpreted by the Calcium Calmodulin Kinase (CCaMK) (Miller et al., 2013), which activates a set of regulators that launch transcription of symbiosis specific genes in the outer root cell layers (Yano et al., 2008; Hirsch et al., 2009). Progression of the symbiotic signalling events from epidermis into the cortex is necessary for nodule organogenesis and infection thread formation. NIN, a transcriptional regulator, and cytokinin signalling have been implicated in this epidermal to cortex signalling (Murray et al., 2007; Tirichine et al., 2007; Vernié et al., 2015).

Mutations in Nfr5 and its homologs in pea and M. truncatula eliminate all Nod factor-induced physiological, molecular and cellular responses (Madsen et al., 2003; Arrighi et al., 2006). However, some or several of these responses are retained in the Ljnfr1, Mtlyk3 and Pssym37 mutants (Radutoiu et al., 2003; Smit et al., 2007; Zhukov et al., 2008) raising the possibility that modular receptor complex formation regulated in a spatio-temporal manner might contribute to Nod factor signalling. The LysM receptor kinase family has greatly expanded in legumes through whole genome or tandem duplications (Zhang et al., 2009; Lohmann et al., 2010; Kelly et al., 2017). In L. japonicus, four genes, Lys1, Lys2, Lys6 and Lys7, are closely related to Nfr1. Lys1 and Lys2 are located in tandem and at ~10 kb distance from Nfr1 (Lohmann et al., 2010). Interestingly, a similar chromosomal organisation of NFR1-type receptors was reported in all studied legumes, as well as in genomes outside of Leguminosae clade raising the possibility that gene duplication leading to tandem NFR1-type receptors preceded the evolution of the legume family (De Mita et al., 2014). The precise role of these NFR1 paralogs and their homologs is unknown apart from the chitin receptors L. japonicus LYS6 (now CERK6) and Medicago truncatula LYK9 (Bozsoki et al., 2017). Nonetheless, key details about the signalling competencies of LYS2, LYS6 and LYS7 were obtained from functional complementation analyses in the nfr1 mutant using the LjUbiquitin promoter. Only the intracellular kinase regions of LYS6 and LYS7, but not that of LYS2, could restore nodulation and/or infection when coupled to the NFR1 Nod factor-binding domain (Nakagawa et al., 2011).

Here, we show that LYS1 is an epidermal LysM receptor contributing to the NFR1-NFR5 mediated signalling in a spatio-temporal manner. This gene is primarily expressed in epidermal cells of the susceptible zone where roots are competent for initiation of symbiosis, and has a restricted signalling capacity leading to Nin activation in the outer root cell layers. Our findings provide evidence for a complex Nod factor signalling where LYS1 activity in the outer root cell layers aids in maintaining a normal calcium spiking interval in the root hairs, integral transcript responses in the susceptible root zone, and initiation of nodule primordia on the expanding root system. The Lys1 gene is therefore renamed Nfre, in accordance with the identified role of this gene during Nod factor signalling in the epidermal layer.

Results

NFRe perceives Nod factor and has an active intracellular kinase

NFRe is predicted to encode a LysM receptor protein with a typical intracellular kinase domain (Figure 1A, Figure 1—figure supplement 1). Based on the close similarity to NFR1 we investigated its biochemical and molecular properties. For this, we analysed the binding capacity of NFRe towards purified pentameric M. loti R7A Nod factor ligand (Bek et al., 2010). The NFRe ectodomain was expressed in insect cells using a recombinant baculovirus induced-expression system (Kawaharada et al., 2015). Pure protein was obtained after four steps of purification and the homogeneity was confirmed by size exclusion chromatography (Figure 1—figure supplement 2). Biolayer interferometry (BLI) (Kawaharada et al., 2015) was used for receptor-ligand affinity measurements since this technique is well suited for handling sparingly soluble hydrophobic compounds like Nod factor. To enable ligand immobilization on streptavidin biosensors M. loti R7A Nod factor and chitin pentamer ((GlcNAc)5, CO5) were conjugated to a biotinylated linker using N-glycosyl oxyamine chemistry (Bohorov et al., 2006; Villadsen et al., 2017) (Figure 1—figure supplement 3). Affinity measurements showed that the ectodomain of NFRe has the capacity to bind M. loti Nod factor with an equilibrium dissociation constant (KD) of 29.1 ± 7.1 μM (Figure 1B,E). Next, we tested whether NFRe has the capacity to bind chitin but no signal was observed for CO5 ligands in this system (Figure 1B). To test our immobilised ligands, we performed the same binding experiment with purified NFR1 ectodomain (Figure 1—figure supplement 2), which gave a KD of 34 ± 6.3 μM to M. loti R7A Nod factor and no binding to CO5 (Figure 1C,E). As a positive control for our chitin ligand we additionally expressed and purified the Arabidopsis CERK1 ectodomain (Figure 1—figure supplement 2) and measured an affinity of 59 μM to the immobilized CO5 (Figure 1D,E), which is very similar to the previously reported KD of 66 μM measured by isothermal titration calorimetry (Liu et al., 2012). In short, our binding studies show that NFRe has the capacity to perceive Nod factor with comparable affinity as seen for the NFR1 and both receptor ectodomains distinguish Nod factor from pentameric chitin ligands in a BLI binding assay (Figure 1E). NFRe is a challenging and low expressed protein and further biochemical ligand competition studies are required to fully define the specificity and receptor capacity of NFRe.

Figure 1 with 3 supplements see all
NFRe perceives Nod factor and has an active intracellular kinase.

(A) The structure of Nfre gene (4663 bp) and predicted protein domains (600 aminoacids). The boxes indicate coding regions, lines are introns, and the location of mutations in the three alleles is indicated. The underlines indicate domains in NFRe; LysM domains, TM: transmembrane, JX: juxtamembrane, kinase. (B), (C) and (D) are binding curves obtained from the biolayer interferometry measurements of NFRe ectodomain, NFR1 ectodomain and AtCERK1 ectodomain interaction with two different ligands, R7A Nod factor and GlcNAc5. Both NFRe and NFR1 ectodomain do not bind to GlcNAc5 but show binding to R7A Nod factor. AtCERK1 ectodomain does not bind R7A Nod factor but binds GlcNAc5. (E) Binding constants of NFRe, NFR1 and AtCERK1 ectodomain to GlcNAc5 and R7A Nod factor obtained from biolayer interferometry steady state-analysis. (F) Nfre encodes an active kinase domain. Autophosphorylation and protein kinase activities of wild-type NFRe, T459A, D436A, D418A NFRe mutant versions, and bovine serum albumin as control are shown. Myelin basic protein was used as substrate for kinase activities. Autoradiogram (top), and SDS-PAGE gels (bottom) are shown. KD, Std. Error and n represent the dissociation constant, standard deviation and number of biological replicates used for the analysis. N.D. represents not detectable.

https://doi.org/10.7554/eLife.33506.003

Next, we assessed the activity of the intracellular kinase domain of NFRe (Figure 1—figure supplement 1). E. coli-produced NFRe kinase transphosphorylated the myelin basic protein (MBP) substrate and autophosphorylated (Figure 1F, lanes 1–3), showing that NFRe encodes a protein kinase with in vitro activity similar to NFR1 (Madsen et al., 2011). Alanine substitutions of three critical amino acids from the catalytic loop (D418), Mg2 +binding loop (D436), or P+1 loop (T459) abolished the phosphorylation activity of NFRe (Figure 1F, lanes 4–12) showing that conserved residues from NFRe kinase are critical for its biochemical activity. Together, our results from biochemical studies demonstrate that Nfre encodes a LysM receptor kinase that can perceive Nod factor and has an active kinase.

NFRe induces epidermal Nin expression

Since we now know that NFRe is an active LysM receptor with properties comparable to NFR1 in these in vitro assays, we next investigated the signalling capacities of NFRe compared to NFR1 in Lotus roots. We tested this by expressing NFRe in the nfr1-1 mutant line containing the symbiotic Nin:GUS reporter (Radutoiu et al., 2003). Activation of the Nin promoter in Nfre transformed roots of nfr1-1-Nin:GUS plants, or the development of nodule and/or infection threads would indicate activation of symbiotic signalling. While nfr1-1 roots transformed with the Nfr1 gene developed bona fide root nodules and induced Nin promoter, those transformed with the empty vector, and thus expressing the native Nfre gene did not show any responses to inoculation with rhizobia (Table 1 and Figure 2—figure supplement 1). These results indicate that Nfre, in its native status cannot replace the functions of NFR1. On the other hand, p35S-Nfre led to strong activation of the Nin promoter after inoculation with M. loti (Figure 2A and Figure 2—figure supplement 1). This symbiotic induction was however, only detected in the outer root layers (Figure 2A), and it was not followed by formation of nodules or infection threads even after 5 weeks of exposure to M. loti (Figure 2—figure supplement 1). This differed from the p35S-Nfr1-mediated signalling that induced Nin expression in both epidermal and cortical cells (Figure 2B) and led to formation of infected nodules (Figure 2—figure supplement 1). To understand whether this particular and cell layer specific activation of Nin by NFRe is a result of the expression of any LysM receptor of the NFR1-type, or specific to NFRe, Nin activation was assayed in nfr1-1-Nin:GUS plants transformed with the Lys paralogs of NFR1 (Lohmann et al., 2010). Under similar conditions, Lys2, Cerk6 or Lys7 driven by 35S promoter could not activate the Nin promoter, or induce nodule or infection thread formation in the nfr1-1 mutant. (Table 1 and Figure 2—figure supplement 1). These results demonstrate that in the presence of M. loti, NFRe, like NFR1, can initiate a symbiotic signalling cascade leading to Nin induction in Lotus roots, and that the cellular effects of this signalling are receptor-, and expression-dependent.

Figure 2 with 2 supplements see all
NFRe maintains a low, epidermal expression during root nodule symbiosis.

(A) Transversal root section of nfr1-1-Nin:GUS plants expressing p35S-Nfre shows activation of Nin promoter in the outer cell layer after M. loti inoculation. (B) Transversal root section of nfr1-1-Nin:GUS plants expressing p35S-Nfr1 shows activation of Nin promoter in all cell layers after M. loti inoculation. (C) The epidermal cells, primarily localized in the root susceptible zone, show activity of the Nfre promoter visualized by the nuclear localized triple YFP protein (tYFPnls). (D) Widespread activity of the Nfr1 promoter in the uninoculated root visualized by nuclear localized triple YFP protein (tYFPnls). (E) The expression of Nfre in the susceptible zone of the root, and in the root hairs is maintained after inoculation with M. loti (F). Scale bars, 40 μm (A, B), 0.5 mm (C, D), and 50 μm (E, F).

https://doi.org/10.7554/eLife.33506.007
Table 1
Nfre expression from p35S promoter activates Nin induction in the nfr1-1-Nin:GUS plants
https://doi.org/10.7554/eLife.33506.010
Construct
(plants analysed)
No. of plants
Nin positive
% of plants with
Nin induction
No. of nodulated
plants
% of nodulated
plants
Empty vector (28)0000
pNfr1:Nfre (21)0000
p35S:Nfre (58)284800
p35S:Nfre_T459A (21)0000
pNfr1:Nfr1 (19)1910019100
p35S:Nfr1 (26)25962596
p35S:Lys2 (14)0000
p35S:Lys6 (34)0000
p35S:Lys7 (34)0000

NFRe maintains a low, epidermal expression during root nodule symbiosis

Previous studies based on transcript measurement showed that Nfre is expressed in Lotus roots (Lohmann et al., 2010). However, our results from the nfr1-1 complementation studies revealed that expression of Nfre from p35S promoter is needed to induce observable Nin activation after rhizobia inoculation (Figure 2—figure supplement 1). To further understand the cause of this differential signalling we characterised the spatio-temporal regulation of Nfre in detail using GUS and tYFPnls (triple YFP-nuclear localised) reporter fusions, and measured the levels of Nfre transcript by quantitative RT-PCR. In uninoculated roots the Nfre promoter (2,6 kb) was preferentially active in root hair epidermal cells, in the susceptible zone of the root, and in the root tip (Figure 2C,E, and Figure 2—figure supplement 2). This differed from Nfr1 that is expressed in the whole uninoculated root (Radutoiu et al., 2003; Kawaharada et al., 2017) (Figure 2D). The expression pattern of Nfre did not change after inoculation with M. loti (Figure 2F and Figure 2—figure supplement 2), indicating that, unlike Nfr1 (Radutoiu et al., 2003; Kawaharada et al., 2017) and Figure 2—figure supplement 2), the expression of Nfre is not symbiotically regulated. Analyses of Nfre transcript levels in wild type roots either treated with Nod factor or inoculated with M. loti compared to control roots, further confirmed the unaltered expression observed from Nfre promoter studies (Figure 2—figure supplement 2). Direct comparison of Nfr1 and Nfre transcript levels in uninoculated wild type roots showed a 3-fold higher level for Nfr1. Interestingly this difference was reduced significantly after Nod factor treatment (8 hr post treatment) or M. loti inoculation (2 and 3 dpi post inoculation), where Nfr1 expression was down regulated, while Nfre maintained a low, but constant level (Figure 2—figure supplement 2). In summary, Nfre and Nfr1 differ in their expression level and pattern in uninoculated roots, and follow a differential regulation during root nodule symbiosis. These differences could therefore, at least in part, account for the differential signalling capacities of the two LysM receptors.

NFRe promotes nodule organogenesis

NFRe has the capacity to bind Nod factors in vitro (Figure 1B) and to induce a symbiotic signalling in planta when expressed in the nfr1-1 mutant from the 35S promoter (Figure 2A). This prompted us to ask whether NFRe plays a role in root nodule symbiosis. Homozygous mutant plants from three independent alleles with exonic insertion of LORE1 retroelement (Mun et al., 2016) (Figure 1A and Supplementary file 1) formed significantly fewer nodules compared to wild type when grown in a binary association with M. loti (Figure 3A). The contribution of NFRe to root nodule organogenesis became even more evident when wild type and nfre mutants were grown in soil and were exposed to the native bacterial community. After 9 weeks, nfre mutants developed only half the number of wild type nodules (Figure 3—figure supplement 1). The shoot biomass and the general plant fitness were significantly reduced (Figure 3—figure supplement 1). Wild type plants had well-developed green pods, while nfre mutants had only few open flowers (Figure 3—figure supplement 1). Analyses of the dynamics of nodule primordia formation on plate-grown plants, revealed that nfre mutants, besides a noticeable reduced nodulation at the early time point (two wpi), had a significantly lower ability to reinitiate nodule formation on the expanding root system (five wpi) (Figure 3B). Unlike nodule organogenesis, the formation of infection threads (IT) appeared not to be affected by mutations in the Nfre. A similar number of ITs were present in wild type and nfre root hairs at 9 or 14 dpi (Figure 3C). The mature nodules formed on nfre appeared normally infected (Figure 3—figure supplement 1), and the proportion of pink/total nodules formed by soil-grown wild type and nfre plants was similar (Figure 3—figure supplement 1), indicating a normal infection process in the nfre mutants.

Figure 3 with 4 supplements see all
NFRe promotes nodule organogenesis in Lotus japonicus.

(A) Greenhouse-grown nfre plants formed fewer root nodules compared to WT when analysed at eight wpi with M. loti. (B) Agar plate-grown nfre plants form fewer primordia than WT at 5 wpi with M. loti. (C) The nfre mutants and wild type plants form similar number of short and long root hair infection threads at 9 and 14 dpi. (D) Representative nuclear calcium oscillations (spiking) induced by R7A Nod factor (10–8 M) in wild type and nfre mutant root hairs. Ca2+ oscillations are presented as ratiometric values between YFP and CFP signals detected on the basis of the NLS-YC3.6 Ca2+sensor. (E) The inter-spike interval of nfre-1 mutant is significantly longer than that of WT. (F) Venn diagrams of Nod factor up- and down-regulated (parentheses) genes detected in the susceptible zone at 24 hr after treatment. The values given at the bottom of columns in (A) and (C) represents the number of plants analysed. Error bars represent standard error of the mean. *p<0.05 and **p<0.01, Student´s t-test compared to wild type.

https://doi.org/10.7554/eLife.33506.011

To further investigate the role of NFRe in Nod factor signalling we analysed its requirement for induction and maintenance of nuclear-associated calcium oscillations (spiking) after Nod factor treatment. Root hairs of wild type (n = 50) and nfre-1 (n = 46) stable transgenics expressing the nuclear localised YC3.6 (Yellow Cameleon) showed clear signs of calcium oscillations after M. loti Nod factor treatment (Figure 3D) (app. 80% of the analysed cells responded). Closer inspection of the spiking frequency revealed that the average inter-spike interval was significantly longer in the nfre cells (106 s) compared to wild type (86 s), indicating that NFRe contributes to a constant interval length of calcium oscillations (Figure 3—figure supplement 2).

Next, we used RNA sequencing to investigate the requirement for Nfre in the transcriptional changes induced by M. loti Nod factor in the susceptible zone of the root at 24 hr after treatment. Genes that were differentially expressed (DEGs adjusted p<0.05) (Materials and methods) in Nod factor treated roots compared to water control were identified in wild type, nfre-1 and nfre-2 mutants. A large proportion of these (44 out of 90), which includes Nin, expansins, nodulins, receptors, transporters and transcription factors, were regulated by Nod factor in wild type but not in the nfre roots, indicating that their appropriate regulation in the susceptible zone, at 24 hr after Nod factor treatment is dependent on an active NFRe (Figure 3—figure supplement 3 and Supplementary file 2). Other symbiosis-related genes like NFY-A, subtilase, and two genes encoding the cytokinin-induced message were found among the 13 DEGs in wild type and nfre mutants. Only one gene (an expansin) was found regulated by the Nod factor in both nfre mutants but not in wild type.

Our biochemical in vitro data based on BLI measurements shows that the NFRe ectodomain does not bind chitopentaose, suggesting that NFRe might not be involved in chitin signalling. To test this hypothesis in planta we measured the induction of reactive oxygen species (ROS) in response to CO8 or CO4 in the nfre mutants and wild type. We found that wild type, nfre-1 and nfre-2 mutants produced comparable levels of ROS, indicating that NFRe is unlikely to be involved in chitin signalling (Figure 3—figure supplement 4).

Together, these results show that NFRe represents an influential component of the epidermal Nod factor signalling in L. japonicus, promoting intracellular signalling that leads to optimal calcium spiking, activation of gene transcription and efficient nodule organogenesis on the expanding root system.

The activation segments of NFR1 and NFRe determine the signalling output

The clear difference observed between NFR1 and NFRe in their ability to induce Nod factor signalling and spatial activation of the Nin promoter in M. loti inoculated nfr1-1-Nin:GUS roots prompted us to identify the molecular determinants for this differential regulation. A chimeric receptor (NeK) containing the NFR1 extracellular domain followed by the transmembrane and intracellular kinase regions of NFRe (NFR1 ectodomain-NFRe kinase- NeK) (Figure 1—figure supplement 1) was constructed. This receptor was expressed in nfr1-1-Nin:GUS to test its capacity to induce activation of Nin promoter after M. loti inoculation. We observed that the signalling capacity of NeK receptor was similar to that of the NFRe, namely only epidermal induction of the Nin promoter (Table 2 and Figure 2—figure supplement 1). This provides evidence for the presence of a molecular determinant for specific Nin induction in the intracellular regions of NFR1 and NFRe receptors. Alignment of the two kinases identified several divergent regions (Figure 1—figure supplement 1), but clear differences were found in the activation segment (Figure 1—figure supplement 1). Based on these differences, and previous knowledge (Nakagawa et al., 2011) that this region is crucial for kinase signalling and substrate recognition, we hypothesised that a specific NFR1/NFRe activation segment determines the specificity of the downstream signalling. We tested this hypothesis by swapping the NFRe activation segment with the corresponding region of NFR1 in the NeK receptor (NFR1 ectodomain-NFRe kinase with the Activation segment of NFR1 -NeKA1) (Figure 1—figure supplement 1). In contrast to the NeK receptor that induced Nin in the outer root cell layers of the nfr1-1 mutant, NeKA1 led to cortical activation of Nin and nodule formation (Table 2 and Figure 2—figure supplement 1).

Table 2
The intracellular domains of NFRe and NFR1 kinases determine the signalling output
https://doi.org/10.7554/eLife.33506.016
Construct
(plants analysed)
No. of plants
Nin positive
% of plants with
Nin induction
No. of nodulated
plants
% of nodulated
plants
p35S:NeK (29)31000
pLjUbi:NeK (40)2500
pLjUbi:NeKA1 (40)26651435
pLjUbi:Nfr1 (22)221002195

These results show that the activation segment in NFR1 and NFRe determines the downstream signalling output in Lotus roots after M. loti inoculation.

NFRe signalling is dependent of NFR5

Genetic and molecular studies established that a concerted NFR1-NFR5 signalling induces the nitrogen-fixing symbiosis (Radutoiu et al., 2003; 2007). Here, we present evidence that in L. japonicus NFRe assists the development of root nodule symbiosis. Therefore, we hypothesised that NFRe-dependent signalling also involves NFR5. For this, we investigated the biochemical capacity of the NFRe kinase to transphosphorylate the intracellular NFR5 pseudokinase. The in vitro kinase assays showed that NFR5 is a substrate for the NFRe kinase. (Figure 4A, lanes 1–3) and that this transphosphorylation was dependent on the activation segment of the NFRe kinase. Mutation of T459 to A abolished the kinase activity of NFRe, while exchanging the native segment with the corresponding region of NFR1 maintained transphosphorylation (Figure 4A, lanes 4–6, 7–9). These results corroborate the observed nodule formation in the nfr1-1 expressing the NeKA1 receptor (Table 2 and Figure 2—figure supplement 1).

Figure 4 with 1 supplement see all
NFRe signalling is dependent of NFR5.

(A) The NFRe kinase phosphorylates NFR5 kinase, whereas the NFReT459A shows no phosphorylation activity. The kinase of NeKA1 receptor in which the activation segment of NFRe was swapped with the corresponding region of NFR1 also phosphorylates NFR5 kinase. NFR1 kinase serves as positive control for NFR5 kinase transphosphorylation. Bovine serum albumin and NFR5 kinase domain are negative controls. (B) Bimolecular fluorescence complementation (BiFC) of YFP signal indicates protein-protein interactions in tobacco leaves. NFRe forms homomers (NFRe-nY +NFRe cY), and heteromers with NFR1 (NFRe-nY +NFR1 cY), or NFR5 (NFRe-cY +NFR5 nY). Formation of heteromeric complexes with NFR5 is not dependent on an active NFRe kinase (NFRe T459A -nY + NFR5 cY). (C) Working model of Nod factor signalling (green line) in the susceptible zone ensuring an efficient nodulation on the expanding root system. NFRe (dark grey line) has a constant expression in the epidermal cells of the susceptible zone. NFR1 (light grey line) dominates the uninoculated root in terms of expression level and spatial distribution (1). Once the symbiotic process is initiated by the Nod factor (NF), the expression of NFR1 is rapidly downscaled (2). A sustained expression of NFRe in the epidermal cells ensures an idling signalling in the susceptible zone, keeping the expanding root system tuned in to rhizobia (3). NFR1 acts as a master switch triggering recurrent symbiotic events in a fast and efficient manner from NFRe-attuned epidermal cells (4).

https://doi.org/10.7554/eLife.33506.017

Next, we analysed the localisation and molecular properties of full-length NFRe and NFR5 using heterologous expression in Nicotiana benthamiana. The YFP tagged NFRe protein was found to localize to the plasma membrane and to co-localise with the plasma membrane marker, AtPIP2, like previously observed for NFR1 and NFR5 (Madsen et al., 2011) (Figure 4—figure supplement 1). Bimolecular fluorescence complementation (BiFC) analyses based on split YFP revealed that NFRe formed homomeric complexes alone and heteromeric complexes when co-expressed with either NFR1 or NFR5 (Figure 4B and Figure 4—figure supplement 1). The formation of heteromeric complex with NFR5 was not affected by kinase inactivation (Figure 4B). Like in the case of NFR1-NFR5 co-expression (Madsen et al., 2011), a signalling cascade leading to leaf cell death, dependent on an active NFRe kinase, was identified in N. benthamiana leaves co-expressing NFRe and NFR5 (Figure 4—figure supplement 1). Finally, we analysed whether the NFRe-dependent activation of Nin in L. japonicus roots was dependent on NFR5. Nfre driven by 35S promoter failed to induce Nin:GUS symbiotic reporter in the nfr5-2 mutant background (Figure 4—figure supplement 1 and Supplementary file 3). These results demonstrate that NFRe can interact with, and trans-phosphorylates NFR5 kinase, and induce a signalling cascade dependent on the NFR5 receptor.

Collectively, our results from biochemical studies of the extracellular and intracellular domains of NFRe, together with those obtained from mutant and functional analyses provide evidences for the involvement of NFRe ensuring a robust signalling for symbiosis with nitrogen-fixing rhizobia.

Discussion

Nod factor binding by NFR1-NFR5 LysM receptors is required to induce nodule organogenesis and infection thread formation in L. japonicus (Radutoiu et al., 2003; 2007). Here, we show that the NFR1-NFR5 signalling cascade operates on the framework provided by the epidermal LysM receptor NFRe. NFRe and NFR1 share biochemical and molecular properties that is similar Nod factor-binding affinity, and chitopentaose differentiating capacity when assessed by biolayer interferometry (Broghammer et al., 2012), functional kinases dependent on fully operative domains (Madsen et al., 2011), capacity to phosphorylate, and induce a signalling cascade dependent of NFR5 (Madsen et al., 2011). In spite of these similarities, NFR1 and NFRe have evolved distinct biological properties defined by specific spatio-temporal expression, and downstream signalling cascades controlled by diverged kinases.

The epidermis of the expanding root system is continuously exposed to Nod factors produced by rhizobia present in the rhizosphere. Nevertheless, the number and the location of primordia guiding the epidermal infection threads are precisely determined. Complex regulatory networks involving transcriptional regulators, hormones, shoot- and root-derived signals (Ferguson et al., 2010; Miyata et al., 2013; Sasaki et al., 2014; Miri et al., 2016; Roy et al., 2017), as well as tightly controlled receptor signalling (Mbengue et al., 2010; Kawaharada et al., 2017), collaborate to coordinate how many nodules the plant develops. With this framework in mind, a working model is emerging when considering our findings (Figure 4C). This model incorporates the interplay of the NFR1 and NFRe in the epidermis, ensuring efficient and robust signalling in the susceptible zone of the expanding root system. In the absence of the symbiont, Nfre has a low and constant expression in the susceptible zone, while Nfr1 outnumbers Nfre in terms of expression level and spatial distribution (Figure 4C-1). Once the symbiotic process is initiated, the expression of NFR1 is rapidly downscaled in the susceptible zone (Figure 4C-2). A sustained expression of NFRe in the epidermal cells of the susceptible zone could ensure an idling signalling, keeping the expanding root system tuned in to rhizobia (Figure 4C-3). NFR1 acts as a master switch triggering recurrent symbiotic events in a fast and efficient manner from NFRe-attuned epidermal cells (Figure 4C–4).

In general, protein-carbohydrate interactions are usually weak and low-affine (micromolar-millimolar range) (Holgersson et al., 2005) and signalling therefore, emerges as being controlled by ligand multivalency and/or by receptor multiplicity (Kiessling and Pohl, 1996; Rabinovich, 2002; Vasta et al., 2012). In line with this notion studies of receptors present at the plant and mammalian plasma membrane revealed a conserved strategy to ensure specific, instantaneous, switchable and evolvable downstream signalling; namely, increased responsiveness and specificity via combinatorial systems (Ostrom et al., 2001; Piñeyro, 2009; Bodmann et al., 2015; Bücherl et al., 2017). The signalling properties of NFRe remain to be determined, but our findings based on the properties of this LysM receptor kinase, together with the symbiotic phenotypes of nfre mutants unveil a more complex signalling operating in the epidermal cells of L. japonicus than anticipated from studies of the basic and essential receptor-components.

It is possible that multiple LysM receptors assemble into functional signalling complexes where signalling specificity is the result of the nature of the complex, rather than isolated LysM receptors alone. The mechanistic details of NFR1-NFRe signalling remain to be discovered, but we envision that differences might exist among legumes, since nfr1 in Lotus and lyk3 or sym37 in Medicago and pea have different symbiotic phenotypes (Radutoiu et al., 2003; Smit et al., 2007; Zhukov et al., 2008), indicating distinct evolutionary trajectories after separation of the IRLC (Inverted Repeat-lacking clade) legumes (Sprent, 2008). Tandem NFR1-type receptors are found in all legumes and in non-legume species as well (Liang et al., 2013De Mita et al., 2014). Ample comparative phylogenomics and trans-complementation studies targeting tandem duplicated LysM receptors will greatly help determining their evolutionary impact and their role in different plant species.

Materials and methods

Phylogenetic tree and alignment

Clustal Omega was used to prepare multiple sequence alignment for phylogenetic analysis. The region between 55 and 251 in this alignment was realignment to adjust the positions of CXC motif. This alignment was used for the phylogenetic analysis with Neighbor Joining. The distance was measured with Jukes-Cantor and the bootstrap was 1000 replicates. These alignment and phylogenetic analyses were performed in the CLC Main Workbench v7.9.1. The amino acid sequence of OsCERK1 (Os08g0538300-01) is available in The Rice Annotation Project Data Base (rap-db). The other sequences below are available in NCBI: AtCERK1 (NP_566689), NFR1 (CAE02590), NFRe (AB503681), LYS2 (AB503682), EPR3 (AB503683), LYS4 (AB503685), LYS5 (AB503686), LYS6 (AB503687), LYS7 (AB503688).

Expression and purification of NFRe, NFR1 and AtCERK1 ectodomains

NFRe and NFR1 ectodomain boundaries were defined by secondary structure prediction performed with PSIPRED (Buchan et al., 2013). Their signal peptides were predicted using the SignalP 4.1 server (Petersen et al., 2011). The AtCERK1 ectodomain boundaries were designed based on the reported crystal structure (Liu et al., 2012). The predicted ectodomain sequences were codon-optimized for insect cell expression and synthesized with an N-terminal gp67 secretion signal peptide and a c-terminal hexa-histidine tag (GenScript, Piscataway, USA) and inserted into the pOET4 transfer vector (Oxford Expression Technologies). Recombinant AcMNPV baculoviruses were produced in Sf9 cells cultured with SFX (Hyclone) or TNM-FH medium (Sigma-Aldrich) supplemented with 10% (v/v) FBS (Gibco), 1% (v/v) chemically defined lipid concentrate (Gibco) and 1% (v/v) Pen/Strep (10,000 U/ml, Life Technologies). The FlashBac Gold kit (Oxford Expression Technologies) was utilized according to the manufacturer’s instructions. Viruses were amplified until a third passage virus culture of 500 mL was obtained. For large scale protein expression Sf9 cells were infected with 5% (v/v) of the passage three virus solution and cultured in suspension with serum-free SFX insect cell medium (Hyclone) or BD BaculoGold MAX-XP medium (BD Biosciences, discontinued) supplemented with chemically defined lipid concentrate and Pen/Strep as described above. The culture was maintained in a shaking incubator at 26°C for five days, after which the medium was harvested by centrifugation in a Sorvall RC5plus centrifuge (SLA-1500 rotor) at 6000 rpm at room temperature for 25 min. Subsequently, the cleared medium was dialyzed against 10 volumes of buffer A (50 mM Tris-HCl pH 8, 200 mM NaCl) for one day at 4°C with the buffer being exchanged at least four times. The proteins were loaded on a HisTrap excel column (GE Healthcare) equilibrated with buffer A and recirculated over 3 days at 4°C using a peristaltic pump. After a washing step with buffer W (50 mM Tris-HCl pH 8, 500 mM NaCl and 20 mM imidazole) proteins were eluted with buffer B (50 mM Tris-HCl pH 8, 200 mM NaCl, 500 mM imidazole). Imidazole was removed by dialyzing against buffer A and the purity was improved by a second IMAC purification step using a HisTrap HP column (GE Healthcare). The NFRe ectodomain was dialyzed against a low salt buffer (50 mM Tris-HCl and 50 mM NaCl) before purification on a MonoQ column (GE Healthcare). The resulting flow-through containing NFRe was collected and concentrated in a Vivaspin column (10 kDa cut-off, Sartorius Stedim biotech). NFRe and NFR1 were finally purified by size exclusion chromatography using a Superdex 200 10/300 GL column (GE Healthcare) and AtCERK1 using a Superdex 75 10/300 GL size exclusion column (GE Healthcare) in SEC buffer (Phosphate buffered saline, pH 7.2, 500 mM NaCl). At each purification step, yield and purity were assayed by SDS-PAGE.

Synthesis of biotinylated R7A Nod factor and chitopentaose conjugates

Biotin conjugates were synthesized using a two-step procedure according to Figure 1—figure supplement 3. O-(2-Aminoethyl)-N-methyl hydroxylamine trifluoroacetic acid salt was prepared as described previously (Bohorov et al., 2006), and all other chemicals were purchased from Sigma-Aldrich and used without further purification. Nod factors from Mesorhizobium loti, strain R7A, NodMl-V(C18:1Δ11Z, Cb, Me, AcFuc), containing three main species (3-O-acetylated, 4-O-acetylated, or non-acetylated fucosyl unit) were purified as described previously (Bek et al., 2010). Purified R7A Nod factor (3.6 mg, 2.29 μmol, 5 mM) was dissolved in 0.62 M NaOAc buffer, pH 4.5, containing 50% acetonitrile, and O-(2-aminoethyl)-N-methyl hydroxylamine trifluoroacetic acid salt (150 mM, 30 equiv.) was added. The resulting mixture was allowed to react at room temperature for 16 hr, after which it was concentrated under a nitrogen flow. The intermediate product was purified by semipreparative HPLC on an UltiMate 3000 instrument fitted with a Waters 996 photodiode detector, using a Phenomenex Luna 5 μm, C18(2), 100 Å, 250 × 100 mm semi-preparative column. An isocratic elution at 40% acetonitrile in water, 5 mL/min for 30 min was used. The intermediate eluted at 9.5–11.5 min. Conjugate formation was confirmed by HR-MS (ES+): calcd for [M + H, 1Ac]+ = 1573.8081, found 1573.8185. The purified intermediate was dissolved at a concentration of 10 mM in 50 mM sodium tetraborate buffer, pH 8.5, containing 50% acetonitrile. NHS-dPEG4-biotin (15 mM, 1.5 equiv.) was added. The resulting mixture was allowed to react at room temperature for 16 hr. The biotin conjugate product was purified by semipreparative HPLC as described above. A gradient of 5–100% acetonitrile in water over 40 min, running at 5 mL/min, was used. The conjugate eluted after 21 min (68% acetonitrile). The chromatogram displayed a broad product peak due to the presence of the three species differing in substitution on the fucosyl residue. The biotin-R7A Nod factor conjugate (18% yield) was quantified using the HABA/avidin biotin quantification kit (Pierce). HR-MS (ES+): calcd for [M + 2 hr, 1Ac]2 += 1024.0175, found 1024.0184, and calcd for [M + 2 hr, 0Ac]2 += 1003.0122, found 1003.0129 (Figure 1—figure supplement 3). High-resolution mass spectra (HR-MS) were obtained using a Dionex Ultimate 3000 UHPLC instrument (Thermo) coupled to a Bruker Impact HDII QTOF mass spectrometer. The synthesis of a biotin-chitopentaose (GlcNAc)five conjugate was performed essentially as for the biotin-R7A Nod factor conjugate. The product was purified by HPLC using a Phenomenex Luna 5 μm, C18(2), 100 Å, 250 × 100 mm semi-preparative column, using a gradient of 5–100% acetonitrile in water, 5 mL/min for 40 min. The product eluted after 12.7 min (35% acetonitrile). The final yield of the biotin-(GlcNAc)five conjugate was determined to be 9%. HR-MS (ES+): calcd for [M + 2 hr]2 += 790.3552, found 790.3557 (Figure 1—figure supplement 3).

Biolayer interferometry

Binding of NFRe, NFR1 and AtCERK1 ectodomains to biotin-R7A Nod factor and biotin-(GlcNAc)5 (CO5) was measured on an Octet RED biolayer interferometer (Pall ForteBio). Biotinylated R7A Nod factor and (GlcNAc)5, were immobilized on streptavidin biosensors (for kinetics, Pall ForteBio) at a concentration of 250 nM for 5 min. Immobilization levels of biotin-R7A and biotin-(GlcNAc)five were followed during immobilization and amounted to approximately 2.4 nm and 0.4 nm of saturation, respectively. Interaction with NFRe, NFR1 and AtCERK1 ectodomains was measured in dilution series at protein concentrations ranging from 0.5 to 64 μM (NFRe), 0.78–100 μM (NFR1) or 0.93–160 µM (AtCERK1) for 10 min. Subsequently, dissociation was recorded for 5 min. All steps were conducted in phosphate-buffered saline, pH 7.4, 500 mM NaCl, 0.01% Tween20. Parallel background measurements using biosensors immobilized with free biotin were subtracted from R7A Nod factor and (GlcNAc)five curves to correct for unspecific binding. Sensorgrams were processed using ForteBio Data Analysis 7.0 (Pall ForteBio). Equilibrium dissociation constants from steady-state analysis were calculated in GraphPad Prism 6 (GraphPad Software) by nonlinear regression using the response at equilibrium (Req) plotted against protein concentration.

NFRe kinase domain expression and purification

The NFRe, NFR1 and NFR5 kinase domains were predicted using TMHMM Server v. 2.0 and PSIPRED secondary structure prediction. NFRe kinase domain was cloned into pET-30 Ek/LIC vector (Novagen), NFR1 and NFR5 kinase domains were cloned into pET-32 Ek/LIC vectors (Novagen). Three NFRe kinase domain mutants were created using the Quikchange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies) according to the manufacturers’ instructions. For the chimeric kinase of NeKA1, cDNA fragment was assembled by PCR as described previously (Heckman and Pease, 2007). NFRe kinase was expressed into E. coli BL21-CodonPlus, NFR1 and NFR5 kinase into E. coli Rosetta 2. Cultures were grown until OD600 = 0.8 and cold-shocked for 30 min in an ice bath. Protein expression was subsequently induced with 1 mM IPTG and left overnight to shake at 20°C. Cells were harvested by centrifugation at 3300 rpm in a Sigma swing-out rotor 13855 and afterwards resuspended in 100 ml lysis buffer (50 mM Tris-HCl pH8, 400 mM NaCl, 1 mM Benzamidine, 20 mM Imidazole, 5 mMβ-mercaptoethanol and 10% (v/v) glycerol). Resuspended pellets were broken by sonication and cell debris removed by centrifugation at 12000 rpm (F21S-8 × 50 y rotor, ThermoFisher). The resulting supernatant was loaded on a Ni-NTA IMAC affinity column (ThermoFisher) equilibrated with lysis buffer at 4°C using a peristaltic pump. After a wash step with buffer W-kinase (50 mM Tris-HCI pH 8, 1 M NaCl, 1 mM Benzamidine, 50 mM Imidazole, 5 mM β-mercaptoethanol and 10% Glycerol) to remove contaminants, kinases were eluted with buffer B-kinase (50 mM Tris-HCl pH 8, 300 mM NaCl, 500 mM imidazole, 5 mM β-mercaptoethanol and 10% Glycerol). His-tagged TEV protease (homemade) was added to the eluted proteins at a 1:100 (w:w) ratio and dialysed against a dialysis buffer (50 mM Tris-HCl pH 8, 300 mM NaCl, 5 mM β-mercaptoethanol and 10% Glycerol) overnight at 4°C. The cleaved kinase domain proteins were subjected to a second round of IMAC affinity column purification and collected in the flow-through. The kinase domain proteins were concentrated in a Vivaspin column (10 kDa cut-off) before being subjected to size exclusion chromatography using either a Superdex 75 or Superdex 200 10/300 GL columns on ÄKTA Purifier system (both GE Healthcare). Purification was performed by isocratic elution in Buffer GF (50 mM Tris-HCl pH 8, 300 mM NaCl and 5 mM β-mercaptoethanol). After each purification step, yield and purity were assayed using SDS-PAGE.

In vitro kinase assay

4 µg of purified kinase domain proteins were incubated in Kinase Activity Buffer (2 mM MnCl2, 2 mM NaCl, 2 mM MgCl2, 1 mM ZnCl2, 50 mM HEPES pH 7 and 100 µM ATP) and 2mCi ATP, [gamma-32P] (PerkinElmer) in a total reaction volume of 20 µL. 2 µg Myelin Basic Protein (Sigma Aldrich) and 4 µg NFR5 kinase domain were added to the appropriate reactions. Additionally, controls without Myelin Basic Protein, ATP [gamma-32P] were made. The reactions were left to incubate for 1 hr at room temperature before loading and running on a 15% SDS-PAGE Gel. After staining with Coomassie Brilliant Blue, the gel was transferred on a phosphor plate and exposed overnight before scanning on a Typhoon Scanner (GE Healthcare).

Plant material

Lotus japonicus, ecotype B–129 Gifu (Handberg and Stougaard, 1992) is the wild type used for all experiments. Homozygous nfre (previously called lys1) mutants were identified in the LORE1 collection (Fukai et al., 2012; Urbański et al., 2012) and the primers used for genotyping are listed in Supplementary file 1. Seeds were sterilized and germinated and the 3 days old seedlings were transferred to the corresponding conditions below.

Bacterial strains and constructs

Mesorhizobium loti, strain R7A labelled with GFP or dsRed, and NZP2235 were used for phenotypic analyses. An inoculum density of OD600 = 0.02 was used for all studies. Agrobacterium rhizogenes AR1193 (Stougaard et al., 1987) was used for hairy root transformation. A. tumefasiens AGL1 was used for the transient expression in leaves of N. benthamiana.

The various constructs used for L. japonicus transformation were assembled using Golden Gate Cloning (Engler et al., 2014), and constructs for N. benthamiana were assembled using Gateway system with 35S promoter driving the expression. The details of each construct and primers for cloning are presented in Supplementary file 1. All constructs were confirmed by sequencing.

Complementation and promoter analysis using hairy root

The seedlings for hairy root transformation were moved to half-strength B5 media and transformed as described previously (Stougaard, 1995). The composite plants were transferred to Magenta boxes containing sterile clay granule substrate or to sterile agar plates supplemented with ¼ B and D media and inoculated.

GUS staining and cross section

Transformed roots were incubated in GUS staining buffer [0.5 mg/mL X-Gluc, 50 mM phosphate buffer (pH7.0), 5% methanol, 1 mM K4(Fe(CN)6), 1 mM K3(Fe(CN)6), 0.05% Triton X-100] at 37°C for 18 hr in dark. The samples were washed with 50 mM phosphate buffer (pH7.0) and stored in 70% ethanol at 4°C. GUS stained roots and nodules were observed using a Leica M165FC stereomicroscope and Leica DFC 310 FX camera system. Three to five representative samples were used for generating transversal sections, as described previously (Gavrilovic et al., 2016).

Microscopic observations for promoter analysis using tYFPnls

For promoter activity using tYFPnls, transformed roots were fixed with paraformaldehyde and cleared as described previously (Warner et al., 2014). The samples were analysed on a ZEISS confocal microscope LSM780. The whole root images were obtained using Z-stack and tail scan tools, the images of root surface were obtained using Z-stack tool. Final images were generated by Maximum Intensity Projection in ZEN software (ZEISS) or ImageJ.

Quantitative RT-PCR

For transcript measurement by quantitative RT-PCR, 3 days seedlings were moved to agar plates supplemented with 1/4 B and D media and whole roots of 12 days-old (Figure 2—figure supplement 2) or 14 days old (Figure 3—figure supplement 3) plants were harvested after specific treatment as specified in each experiment. The mRNA was isolated from whole roots (Figure 2—figure supplement 2) or the susceptible zone (Figure 3—figure supplement 3) using Dynabeads mRNA DIRECT TM kit (Invitrogen). RevertAid Reverse Transcriptase (Fermentas) was used for cDNA synthesis. The quantitative RT-PCR was performed with LightCycler 480 II and LightCycler 480 SYBR Green I Master mix (Roche). ATP-synthase (ATP), Ubiquitin-conjugating enzyme (UBC) and Protein phosphatase 2A (PP2A) were used as reference genes. The three biological (each consisting of 40 plants in Figure 2—figure supplement 2 or 30 plants in Figure 3—figure supplement 3) and three technical repetitions were used to calculate the geometric mean of the relative transcript levels and the corresponding upper and lower 95% confidence.

Plant phenotyping

Sterile agar plates or clay granule substrate supplemented with ¼ B and D media was used for phenotypic analysis in laboratory and greenhouse conditions (Figure 3A,B). Cologne soil (Zgadzaj et al., 2016) with no additional nutrients or inoculum was used for plant phenotyping in Figure 3—figure supplement 1.

Observation of infection threads (IT)

The 3 day old seedlings were transferred to agar plates supplemented with ¼ B and D media. After 3 days growth, the plants were inoculated with M. loti R7A labelled with dsRed. The infection threads were counted at 9 and 14 days after inoculation.

Nuclear calcium oscillation in root hairs

Seedlings were grown on agar plates supplemented with 1/4 B and D with 12.5 µg/mL AVG for 1–2 weeks. One seedling was transferred to a glass slide and Nod factor treatment was performed using 10–8 M M. loti R7A Nod factor solution. The samples were analysed on a confocal microscope LSM780 (ZEISS) and a water lens (W plan-Apochromat 40x/1.0 DIC M27, ZEISS). YC3.6 was excited at 458 nm, and emissions from ECFP and cpVenus were split into different detectors and collected at 463 to 509 and 519 to 621 nm. Calcium spiking was monitored for up to 3 hr after the Nod factor treatment on each root. Several regions of the same root were monitored for 10 to 30 min, and minimum five nuclei were monitored on each root. In total, 50 nuclei from wild type and 46 nuclei from nfre-1 root hairs were monitored. The fluorescence intensity data collected in the first 10 min for each nucleus was analysed by CaSA software (Russo et al., 2013). For calculation to the mean time between Ca2+ spikes (inter spike interval, ISI) for each genotype, the mean of ISI for one cell was used.

RNA sequencing

For RNA sequencing, 3 days seedlings were moved to agar plates supplemented with 1/4 B and D media and susceptible zone of 14 days-old plants was harvested after specific treatment as specified in Figure 3F. The total RNA was isolated from the susceptible zone (15 mm root pieces) using Nucleo spin RNA plant (Macherey-Nagel). Total RNA (>0.8 µg) from two biological replicas per sample was used by GATC Biotech (Germany) to prepare random primed cDNA library and for sequencing with Illumina HiSeq: read length 1 × 50 bp. For the analysis of the RNA sequencing data the read trimming and mapping were performed by CLC genomics workbench 9.5.3 using Lotus japonicus v3.0 at Lotus base (https://lotus.au.dk/) (Mun et al., 2016), as reference. Differentially expressed genes (log2 fold change >0 or<0, adjusted p value < 0,05) were determined using the DESeq2 R package, with the ‘fittype’ parameter set to ‘local’ and the ‘betaprior’ parameter to ‘true’. The HTS filter R package was integrated in the DESeq2 pipeline before calling for differentially expressed genes, in order to remove from the analysis the genes with low read counts. Venn diagrams were generated with the VennDiagram R package (Chen and Boutros, 2011).

Reactive oxygen species (ROS) measurements

Seedlings were germinated and grown on a stack of wet filter paper in upright position at 21°C under 16/8 hr light/dark conditions. Roots of 7 day old seedlings were cut to 0.5 cm pieces, collected to white 96 well flat bottom polystyrene plates (Greiner Bio-One) and kept overnight in sterile water in darkness at room temperature to recover from stress before the treatment. ROS measurements were conducted in a Varioskan Flash Multimode Reader (Thermo Scientific) in luminometric measurement mode. The reaction mixture consisted of the respective elicitor, 20 µM luminol (Sigma) and 5 µg/ml horseradish peroxidase (Sigma). As elicitor 1 µM tetra-N-acetyl-chitotetraose, CO4 (Megazyme) or octa-N-acetyl-chitooctaose, CO8 (IsoSep) was used. In the negative control wells water was replacing the elicitor. In one measuring well six roots (10 mg root material) was used. In one repetition three wells were measured for every treatment for every genotype. At least two repetitions were conducted with similar results.

Protein localization and BIFC studies in N. benthamiana leaves

N. benthamiana, infiltrated leaves were analysed after 3 days using a Zeiss LSM510 MetaConfocal microscope. The leaves were infiltrated with 0.8 M mannitol to induce plasmolysis. The samples were mounted in 30% glycerol on the slide. For cell death in N. benthamiana, infiltrated leaves were observed after 4 days.

Data availability

RNA-seq reads were deposited at ArrayExpress (accession: E-MTAB-5855).

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
    Agobacterium rhizogenes as a Vector for Transforming Higher Plants
    1. J Stougaard
    (1995)
    In: H Jones, editors. Plant Gene Transfer and Expression Protocols. Totowa, NJ: Springer New York. pp. 49–61.
    https://doi.org/10.1385/0-89603-321-X:49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60

Decision letter

  1. Thorsten Nürnberger
    Reviewing Editor; University of Tübingen, Germany

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

[Editors’ note: this article was originally rejected after discussions between the reviewers, but the authors were invited to resubmit after an appeal against the decision.]

Thank you for submitting your work entitled "Epidermal Nod factor receptor NFRe ensures robust symbiotic signalling in Lotus japonicus roots" for consideration by eLife. Your article has been reviewed bythree3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

The identification of eNFR as an 'idling state' epidermis-specific Nod factor receptor in Lotus japonicus is an important and novel finding. In particular, eNFR-mediated differential spatio-temporal signaling in partnership with NFR5 bears the potential of re-shaping the field's view on Nod factor perception and signaling. However, as substantial concerns about the ligand specificity of this newly identified receptor and its predominant epidermal localization remain we are unable to recommend publication of your manuscript in its present form.

Reviewer #1:

Simona Radutoiu and Jens Stougaard have shown over the last decades that the LysM receptor kinase gene family has a crucial role in the perception of rhizobial signaling and exopolysaccharides. The gene family is large and rapidly evolving, making the analysis of the function of individual members challenging, further impeded by partial redundancy and limited functional conservation between model legume species. The manuscript by Murakami Eiichi et al., describes another LysM-domain containing receptor-like kinase (NFRe) that is closely related to the Nod factor receptor NFR1. Like previously shown for NFR1, the presently described NFRe binds Nod-factor and plays a role in the symbiotic signalling in Lotus japonicus. The authors confirm the predicted Nod-factor binding of the NFRe ectodomain by bio-layer interferometry with a biotinylated Nod factor and demonstrate that the kinase kinase domain is active with an in vitro kinase assay. Because of the biochemical properties and the sequence conservation with the known RLK NFR1, the authors conclude a close relationship of the two proteins. The expression of NFRe is restricted to epidermal cells and the expression pattern does not change after Nod-factor treatment. They show this with a promoter-GUS assay and promoter-fluorophore reporter fusion in microscopic pictures. The nfre mutant exhibits a much weaker nodulation phenotype (reduced nodule number) compared to the nfr1 or nfr5 mutants, which are completely unable to nodulate. The authors also claim that nfre mutants shows altered calcium spiking after inoculation which they measured using a calcium sensitive FRET reporter. According to the authors the transcriptome determined via RNA sequencing is also altered after inoculation with rhizobia. In addition, they suggest but do not demonstrate, an interaction of NFRe with NFR5. In the discussion, they propose that NFRe plays a role in attuning epidermal cells for rhizobial infection. Overall the work employs a wide variety of experimental techniques.

There are a number of major issues that must be resolved in order to properly support some of their major claims (see below).

Importantly, the authors fail to make clear what aspects of the identification and characterization of NFRe constitute a level of scientific novelty sufficient to justify publication in eLife.

1) The authors claim that nfre mutants show altered calcium spiking. However, in figure 3—figure supplement 3B,C,D it is not apparent how many samples were taken. If I understood correctly from the material and methods, 5 nuclei were measured. This is sample size is too small for a student’s t-test. The nodulation specific pattern of calcium spiking may differ between cells inside and outside the susceptible zone. The selection of the nuclei may therefore be especially relevant and must be comparable between the genotypes analyzed. Also, the potential influence of background mutations must be ruled out. It is perfectly fine if the authors drop this claim in case they cannot provide sufficient statistical support for it.

2) The authors claim that NFRe specifically binds Nod-factors and with a similar Kd as NFR1. This is supported by a bio-layer interferometry binding assay. However, considering that the Radutoiu/Stougaard laboratories in Aarhus have already claimed differential binding competence of individual LysM receptors for at least three different ligands, the specificity of NFRe must be supported with a necessary array of controls available in their laboratories. Here, they only included CO5 as a negative control. Because of the potential binding competence of LysM domains for exopolysaccharide fragments as shown for the putative exopolysaccharide LysM receptor EPR3 (Kawaharada et al., 2015) the inclusion of this ligand appears essential. Ideally, the authors also include as a negative control the EPR3 ectodomain to demonstrate the specificity of their assay and of the NFRe ectodomain.

3) The authors claim that NFRe is a paralog of NFR1 but shows a distinct downstream signalling based on differences in the kinase activation segment. However, they switch between three different promoters and use inconsistent sets of control and swap constructs. In particular the key domain swap called N1_E.1 lacks the NFR1 and NFRe control constructs. This makes it impossible to reach conclusions from the quantitative complementation results. The NFR1 promoter constructs lack the N1_E.1 swap construct, the ubiquitin promoter constructs lack the NFR1 and NFRe controls and the 35S constructs do not contain the N1_E.1 swap. The experimental strategy underlying this construct arrangement is obscure.

4) The authors claim that nfre mutants show reduced nodulation. Mutants were analysed two, five and eight weeks post inoculation and all the mutants are carrying 'significantly' lower nodules only eight weeks post inoculation. However, in the relevant figure 3—figure supplement 1, the statistical test applied (student’s t-test) is inappropriate for the small subset of individual plants investigated.

5) The authors claim that NFRe does not undergo expression changes after inoculation with Nod factor or rhizobia. However, the expression level of NFRe and NFR1 after treatment with Nod-factor or Rhizobia were measured by qRT-PCR shown in figure 2—figure supplement 3. It compares the expression levels to a mock treated plant. The mock treatment is not specified also not in the Material and methods section. The mock treatment is only shown for one of the three timepoints measured, therefore the variation in expression without treatment is not clear.

6) The authors claim that NFRe interacts with NFR5, based on a BiFC assay in N. benthamiana leaves. The limited validity of this assay for membrane bound proteins is well recognized in the community. One reason is the inability to provide negative controls because of the high non-specific interaction frequency resulting from overexpression and the diffusion of membrane proteins in only two dimensions. Consistent with this problem, the authors of this work also fail to provide a proper negative control.

7) The authors claim that NFRe phosphorylates NFR5. However, the kinase assay lacks controls demonstrating that the phosphorylation of NFR5 by NFRe is specific. Because plant receptor kinase kinase domains are highly promiscuous with regards to their phosphorylation substrate in in vitro kinase assays, the results do not allow the conclusion that the kinase interacts with NFR5 in vivo.

8) The authors claim that the presence of NFRe in the epidermal cells keeps them tuned to rhizobia also after downregulation of NFR1 during symbiotic signaling. This is the conclusion of the claims mentioned above for which some need further investigation.

Reviewer #2:

Nod factor signaling is initiated by LysM domain receptor kinases and in Lotus NFR1 and NFR5 play essential roles. As a result of their strong phenotypes, there is general acceptance in the symbiosis field that the nod factor receptors are known (NFR1/NFR5) and therefore the other LysM domain kinases must be involved in other aspects and not in nod factor signaling. As shown here, this is not true and complete nod factor signaling is achieved through different receptor complexes. In this careful and comprehensive study of NFRe, the authors reveal the contribution of a second receptor kinase and show that it is necessary for the complete signaling required to attain wildtype levels of nodulation. NFRe contributes by providing different spatio-temporal signaling and does so through partnership with NFR5. These findings are unexpected but that data are convincing (with one caveat mentioned below) and demonstrate the contribution of NFRe which, in Lotus japonicus, plays a minor role relative to the dominant receptor NFR1. Not only does the study illustrate complexities in nod factor signaling, it is very likely that it will provide the basis for answers to differences observed in nodulation signaling among the legume where currently significant differences in the phenotypes of the receptor mutants are not understood. The NFR1/NFRe duplication occurs in all legumes and different levels of contribution from the two paralogs may explain many of the differences between legumes. These data will cause the symbiosis field to reconsider nod factor signaling.

One of the major claims is a difference in spatial and temporal expression patterns of the two receptors. While the Q-RT-PCR and promoter-GUS data are convincing, I found it difficult to see differences between NFR1 and NFRe promoter-YFP-NLS (Figure 2). Maybe additional panels could be added to illustrate the expression differences in the susceptible zone.

Reviewer #3:

In this paper, the authors report that an epidermal LysM receptor kinase (LYS1/NFRe) perceives Nod factors and contributes to rhizobial symbiosis, probably by keeping the epidermal cells at an "idling" state. Although the involvement of NFRe in the symbiotic signaling is evident from the results with the corresponding mutants but the detailed mechanism by which NFRe contributes to the symbiosis is still largely obscure. Importantly, biochemical characterization of NFRe contains several problems that raise serious questions on the "receptor" function of NFRe.

The authors analyzed the interaction of the heterologously expressed ectodomains of NFRe and NFR1 using bilayer interferometry, where the biotinylated derivatives of a Nod factor and (GlcNAc)5 were immobilized onto the sensor tip. The results are quite strange. Potato lectin, which is known to bind chitin oligosaccharides but nothing to do with symbiosis, showed a very high affinity (Kd, 30 nM) to the Nod factor, while both NFRe and NFR1 showed thousand times lower (!) affinities to the Nod factor. The affinities of NFRe and NFR1 to the Nod factor (Kd, approximately 30 μM) are also far low compared to the biological activity of Nod factors. In general, the instrument used here seems to give a rather high estimate of the affinity, because the affinity of potato lectin to (GlcNAc)5 obtained here (Kd, 5.2 nM) seems to be much higher than the traditional estimate of the affinity of this lectin to chitin oligosaccharides (The lectins, Academic Press). Thus, it is quite questionable whether such a low affinity of NFRe and NFR1 to the Nod factor observed in this work reflects a biologically significant interaction. I am afraid that such a very low affinity simply reflects some non-specific interaction of these proteins with the hydrophobic region of Nod factor, fatty acid moiety.

To avoid such a possibility, the authors should incorporate inhibition type experiments in their binding studies. If the observed binding of NFRe and NFR1 to the Nod factor reflects a biologically significant interaction, it should be inhibited by the biologically active Nod factor but not by inactive Nod factors or chitin oligosaccharides.

There are some problems for the characterization of the expressed proteins too. Although the authors described "Proteins of high purity were obtained after four steps of purification, as confirmed by the homogeneous and sharp peaks observed after size exclusion chromatography" (Subsection “NFRe is a Nod factor receptor with an active intracellular kinase”), the results in Figure 1—figure supplement 2B shows the presence of multiple peaks for NFR1-ECD. If the authors used only the peak indicated as NFR1-ECD, they should describe it in the figure legend. It is better to show the elution profiles of final preparations of NFRe/NFR1-ECD by using the same column. The authors should also explain why these preparations showed multiple bands in SDS-PAGE (Figure 1—figure supplement 2C). It is also questionable why the sizes of the smallest bands for NFR1-ECD and NFRe-ECD are very different, while the expected sizes of these proteins are quite similar. I am afraid that some limited proteolysis happened for NFR1-ECD.

Characterization of the nfre mutants clearly showed that NFRe contributes to symbiotic signaling and nodule formation. However, the results of the overexpression experiments seem to be difficult to understand. In the nfr1 background, slight increase of the copy number of Nfre gene resulted nothing but the overexpression of the same gene induced a limited symbiotic response induced by M. loti, i.e., activation of Nin promotor in the epidermal cell layer. This Nin activation was not associated with the formation of nodules. Is it safe to connect the results only observed with the overexpression experiments to physiological function? Does the result fit to the hypothesis of "idling" function of NFRe?

The Venn diagram shown in Figure 3F indicates the presence of significant differences between nfre-1 and nfre-2. Is it OK?

Would it be better to confirm the results of nfre mutants by complementation?

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for choosing to send your work entitled "Epidermal Nod factor receptor NFRe ensures robust symbiotic signalling in Lotus japonicus roots" for consideration at eLife. Your article and your letter of appeal have been considered by a Senior Editor and the Reviewing Editor of your manuscript, and we regret to inform you that we are upholding our original decision.

Although we acknowledge the technical issue, it is difficult that a method hardly used in the field should explain obvious discrepancies noted by the referees. In this context, it also is somewhat problematic that if the reviewers do not catch it, the majority of readers won't either. Thus, the binding studies should be complemented with additional assays, using a more frequently used reliable standard method, such as SPR or MST. We would be ready to evaluate a resubmission of a manuscript that has been amended accordingly.

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Epidermal Nod factor receptor NFRe ensures robust symbiotic signalling in Lotus japonicus roots" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Christian Hardtke as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Nod factor signaling is initiated by LysM domain receptor kinases and in Lotus NFR1 and NFR5 play essential roles. Your finding of NFRe, a receptor kinase that it is necessary for the complete signaling required to attain wildtype levels of nodulation, is intriguing and novel. In addition, the finding that NFRe contributes by providing different spatio-temporal signaling and does so through partnership with NFR5, is exciting and will impact future research on how plant rhizobia symbiosis is established.

While all referees appreciate much of the information reported they still request additional experimentation and clarification to be provided. These points are listed below:

1) The highly volatile nature of the spiking frequency makes it very difficult to determine fundamental differences. Thus, more detailed dissection of the patterns is required to support your conclusions. Likewise, strong statements in the text and the abstract should be softened in order to reflect the strength of the phenotype observed in the nfre1-1 mutant.

2) The referees believe that (transient) complementation of the nfre1-1 mutant is required to overcome problems associated with quantitatively differing phenotypes observed in the three nfre1-1 mutant alleles.

3) The use of inhibition type experiments is mandatory in addition to the differential binding experiments. This is important in order to demonstrate specificity of binding, which in addition to ligand affinity is a major criterion for biologically meaningful receptor-ligand interaction. Use of Nod factor, chitin oligosaccharide or exopolysaccharide as competitors in biolayer interferometry are likely to strengthen your claim about the receptor function of NFRe.

You may also wish to show that the biotinylated R7A Nod factor retains the biological activity.

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Epidermal Nod factor receptor NFRe ensures robust symbiotic signalling in Lotus japonicus roots" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Thorsten Nürnberger as the Reviewing Editor and Christian Hardtke as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

The referees acknowledge to a great extent the changes introduced into your manuscript. They also realize that you have tried hard to tackle the problem of demonstrating nod factor binding specificity to NFRe, which unfortunately was unsuccessful for technical reasons. Nevertheless, the identification of NFRe and genetic evidence for its role in nodulation remains a very valuable finding. We would therefore recommend publication of your study given you re-write those chapters of your manuscript that describe NFRe as a nod factor binding site (receptor). Instead, you may refer to eNFR as a protein involved in nodulation for which receptor function is yet to be shown. You may also briefly refer to technical problems you had with handling this protein in order to demonstrate its receptor function, which may be useful information for our readership. Likewise, you may state what type of experiment would be needed to demonstrate receptor specificity.

https://doi.org/10.7554/eLife.33506.026

Author response

[Editors’ note: the author responses to the first round of peer review follow.]

The identification of eNFR as an 'idling state' epidermis-specific Nod factor receptor in Lotus japonicus is an important and novel finding. In particular, eNFR-mediated differential spatio-temporal signaling in partnership with NFR5 bears the potential of re-shaping the field's view on Nod factor perception and signaling. However, as substantial concerns about the ligand specificity of this newly identified receptor and its predominant epidermal localization remain we are unable to recommend publication of your manuscript in its present form.

We are happy to see that the reviewers and the editors appreciate the novelty and the high relevance of our findings. We have identified and addressed the concerns on NFRe capacity to bind Nod factor, and of chosen controls, have performed additional experiments and addressed reviewer’s comments in detail.

Reviewer #1:

Simona Radutoiu and Jens Stougaard have shown over the last decades that the LysM receptor kinase gene family has a crucial role in the perception of rhizobial signaling and exopolysaccharides. The gene family is large and rapidly evolving, making the analysis of the function of individual members challenging, further impeded by partial redundancy and limited functional conservation between model legume species. The manuscript by Murakami Eiichi et al. describes another LysM-domain containing receptor-like kinase (NFRe) that is closely related to the Nod factor receptor NFR1. Like previously shown for NFR1, the presently described NFRe binds Nod-factor and plays a role in the symbiotic signalling in Lotus japonicus. The authors confirm the predicted Nod-factor binding of the NFRe ectodomain by bio-layer interferometry with a biotinylated Nod factor and demonstrate that the kinase kinase domain is active with an in vitro kinase assay. Because of the biochemical properties and the sequence conservation with the known RLK NFR1, the authors conclude a close relationship of the two proteins. The expression of NFRe is restricted to epidermal cells and the expression pattern does not change after Nod-factor treatment. They show this with a promoter-GUS assay and promoter-fluorophore reporter fusion in microscopic pictures. The nfre mutant exhibits a much weaker nodulation phenotype (reduced nodule number) compared to the nfr1 or nfr5 mutants, which are completely unable to nodulate. The authors also claim that nfre mutants shows altered calcium spiking after inoculation which they measured using a calcium sensitive FRET reporter. According to the authors the transcriptome determined via RNA sequencing is also altered after inoculation with rhizobia. In addition, they suggest but do not demonstrate, an interaction of NFRe with NFR5. In the discussion, they propose that NFRe plays a role in attuning epidermal cells for rhizobial infection. Overall the work employs a wide variety of experimental techniques.

We thank reviewer #1 for the appreciative words on our work and efforts in defining the role of LysM receptors despite the inherent difficulties.

There are a number of major issues that must be resolved in order to properly support some of their major claims (see below).

Importantly, the authors fail to make clear what aspects of the identification and characterization of NFRe constitute a level of scientific novelty sufficient to justify publication in eLife.

Since their discovery in 2003, NFR1/NFR5 have been considered the sole Nod factor receptors in L. japonicus. This is primarily because of their critical role in the initiation of nodule development and rhizobial infection. Our detailed characterisation of NFRe demonstrates that this capacity to perceive and signal after Nod factor perception is not solely mediated through the known NFR1/NFR5 receptors – and robust signalling is dependent on NFRe. Our results shear light on the complexity in Nod factor perception, that unlike previously thought, seems to require receptor multiplicity to ensure switchable and evolvable signalling.

To accommodate this major comment, we have adjusted our Discussion section that reads now: “In general, protein-carbohydrate interactionsare characterised by a low affinity36. Signalling therefore, emerges as being controlled by ligand multi-valency and/or by receptor multiplicity37-39. In line with this notion studies of receptors present at the plant and mammalian plasma membrane revealed a conserved strategy to ensure specific, instantaneous, switchable and evolvable downstream signalling; namely, increased responsiveness and specificity via combinatorial systems40-43. Our findings on the biochemical and molecular properties of NFRe, together with the symbiotic phenotypes of nfre mutants unveil an example of a more complex signalling operating in the epidermal cells of L. japonicus than anticipated from studies of the basic and essential receptor-components.”

1) The authors claim that nfre mutants show altered calcium spiking. However, in figure 3—figure supplement 3B,C,D it is not apparent how many samples were taken. If I understood correctly from the material and methods, 5 nuclei were measured. This is sample size is too small for a student’s t-test. The nodulation specific pattern of calcium spiking may differ between cells inside and outside the susceptible zone. The selection of the nuclei may therefore be especially relevant and must be comparable between the genotypes analyzed. Also, the potential influence of background mutations must be ruled out. It is perfectly fine if the authors drop this claim in case they cannot provide sufficient statistical support for it.

We would like to assure reviewer #1 that we did take into consideration the different regions of the root and that we mentioned this in the Materials and methods section: “Several regions of the same root were monitored for 10 to 30 min, …” Furthermore, in the main Figure 3 we present the actual number of nuclei analysed for calcium spiking in wild-type (n=50) and in nfre (n=46). To make this point clearer we have now added the numbers to the main text and to the methods section that reads now: “In total, 50 nuclei from wild type and 46 nuclei from nfre-1 root hairs were monitored.” For clarity, in the Figure 3—figure supplement 2 we present the pattern of calcium spiking in representative nuclei.

2) The authors claim that NFRe specifically binds Nod-factors and with a similar Kd as NFR1. This is supported by a bio-layer interferometry binding assay. However, considering that the Radutoiu/Stougaard laboratories in Aarhus have already claimed differential binding competence of individual LysM receptors for at least three different ligands, the specificity of NFRe must be supported with a necessary array of controls available in their laboratories. Here, they only included CO5 as a negative control. Because of the potential binding competence of LysM domains for exopolysaccharide fragments as shown for the putative exopolysaccharide LysM receptor EPR3 (Kawaharada et al., 2015) the inclusion of this ligand appears essential. Ideally, the authors also include as a negative control the EPR3 ectodomain to demonstrate the specificity of their assay and of the NFRe ectodomain.

Previous work to which we have contributed significantly demonstrated the capacity of LysM ectodomains to bind and signal after perception of specific carbohydrate ligands; Nod factor, COs, peptidoglycan, exopolysaccharide. This is not unusual for receptors as previously shown for the LRR ectodomains that are able to perceive e.g. a diverse panel of peptides. We agree that performing large comparative studies based on single ectodomains binding capacities might be informative, but these need to be interpreted in the context of their biological function. We now included AtCERK1 as control and demonstrate that NFRe binds Nod factor and discriminates COs in a similar way as NFR1, while AtCERK1 shows the opposite binding specificities. Furthermore, none of our results obtained from nfre mutant characterisation pointed towards a role in exopolysaccharide perception, which were shown to be important for infection thread formation and elongation. The nfre mutants have a normal rate of bacterial infection in the epidermis, and inside the nodules. Additionally, inoculation with the M.loti exoU mutant resulted in reduction of the number of nodule primordia like observed for M.loti wild-type, rather than rescue of the infection phenotype as observed for the epr3 mutants (Murakami et al., personal communication).

3) The authors claim that NFRe is a paralog of NFR1 but shows a distinct downstream signalling based on differences in the kinase activation segment. However, they switch between three different promoters and use inconsistent sets of control and swap constructs. In particular the key domain swap called N1_E.1 lacks the NFR1 and NFRe control constructs. This makes it impossible to reach conclusions from the quantitative complementation results. The NFR1 promoter constructs lack the N1_E.1 swap construct, the ubiquitin promoter constructs lack the NFR1 and NFRe controls and the 35S constructs do not contain the N1_E.1 swap. The experimental strategy underlying this construct arrangement is obscure.

We agree that our presentation did not distinguish the different aims of these experiments. The choice of promoter was driven by the scientific question to be addressed. First, we investigated the role of NFRe expression for its functionality by driving Nfre from Nfr1 and 35S promoters. This revealed that a high level of expression from 35S promoter was needed to drive Nin induction after Rhizobium inoculation. Next, we use this promoter and a highly expressed promoter of Lotus (Lj-ubiquitin) to drive the expression of chimeric constructs defining the role of different intracellular regions of NFR1 and NFRe. We have revised the manuscript (Results section) for clarity.

4) The authors claim that nfre mutants show reduced nodulation. Mutants were analysed two, five and eight weeks post inoculation and all the mutants are carrying 'significantly' lower nodules only eight weeks post inoculation. However, in the relevant figure 3—figure supplement 1, the statistical test applied (student’s t-test) is inappropriate for the small subset of individual plants investigated.

The reviewer #1 is right, the nfre mutants show a significant reduction in the number of nodules at 5 weeks when using the plate system or when analysed after 8 weeks in greenhouse or soil conditions. This apparent “late” but significant and robust response across three growth conditions demonstrates its role in ensuring a robust symbiotic signalling in the epidermis necessary for NFR1/NFR5 master switch that trigger recurrent symbiotic events on the growing root system. The statistical test for the Figure 3—figure supplement 1 was adapted to the number of analysed plants (Kruskal-Wallis) and a significant difference between wildtype and mutant plants was observed.

5) The authors claim that NFRe does not undergo expression changes after inoculation with Nod factor or rhizobia. However, the expression level of NFRe and NFR1 after treatment with Nod-factor or Rhizobia were measured by qRT-PCR shown in figure 2—figure supplement 3. It compares the expression levels to a mock treated plant. The mock treatment is not specified also not in the Material and methods section. The mock treatment is only shown for one of the three timepoints measured, therefore the variation in expression without treatment is not clear.

Numerous observations of uninoculated transgenic roots expressing Nfr1/Nfr5/Nfre-GUS or tYFP fusions revealed no detectable changes in the expression pattern or intensity, consequently we used their expression levels in roots at 8h after water treatment (Mock) as baseline.

6) The authors claim that NFRe interacts with NFR5, based on a BiFC assay in N. benthamiana leaves. The limited validity of this assay for membrane bound proteins is well recognized in the community. One reason is the inability to provide negative controls because of the high non-specific interaction frequency resulting from overexpression and the diffusion of membrane proteins in only two dimensions. Consistent with this problem, the authors of this work also fail to provide a proper negative control.

We show that NFRe can interact with NFR5 in N. benthamiana, that its kinase can phosphorylate NFR5 kinase and that Nin induction by 35S-NFRe in Lotus roots is dependent of NFR5. All together, we use three different experimental systems (in vitro, N. benthamiana leaves and Lotus roots) and show that NFRe can interact with NFR5 like it was previously shown for NFR1.

We do not understand what the reviewer means by lack of “proper negative control for this work”. We provide evidences for the lack of BiFC in the absence of the respective tested partners (Figure 4—figure supplement 1).

7) The authors claim that NFRe phosphorylates NFR5. However, the kinase assay lacks controls demonstrating that the phosphorylation of NFR5 by NFRe is specific. Because plant receptor kinase kinase domains are highly promiscuous with regards to their phosphorylation substrate in in vitro kinase assays, the results do not allow the conclusion that the kinase interacts with NFR5 in vivo.

We agree with the reviewer #1 that kinases tend to be highly promiscuous and considering the number of potential targets it is not possible to demonstrate specificity. We have therefore not claimed specificity for NFRe-NFR5 interaction. We show that NFRe kinase can phosphorylate NFR5 and we do not draw the conclusion of NFRe interaction with NFR5 on the basis of this trans-phosphorylation capacity alone. We have however, shown that the activation loops of NFR1 and NFRe are major determinants of the downstream signalling from the two kinases.

We refer the reviewer to the following sentence in subsection “NFRe signalling is dependent of NFR5”, “These results demonstrate that NFRe can interact with, and trans-phosphorylates NFR5 kinase, and induce a signalling cascade dependent on the NFR5 receptor. Collectively, our results from biochemical studies of the NFRe extracellular and intracellular domains, together with those obtained from mutant and functional analyses provide evidences for a novel epidermal Nod factor receptor ensuring a robust signalling for symbiosis with nitrogen-fixing rhizobia.” that clearly state the combined experimental evidence for NFRe function.

8) The authors claim that the presence of NFRe in the epidermal cells keeps them tuned to rhizobia also after downregulation of NFR1 during symbiotic signaling. This is the conclusion of the claims mentioned above for which some need further investigation.

We have addressed the above comments and revised the manuscript. The results obtained from nfre mutant analyses, functional studies of NFRe in the nfr1 mutant and biochemical evidences for Nod factor binding suggest a model for the interplay between NFR1 and NFRe on the expanding root system. We believe that identification of a new Nod factor receptor in legume roots with a differential function opens up for novel studies for fine dissection of the signalling initiated at the plasma membrane.

Reviewer #2:

Nod factor signaling is initiated by LysM domain receptor kinases and in Lotus NFR1 and NFR5 play essential roles. As a result of their strong phenotypes, there is general acceptance in the symbiosis field that the nod factor receptors are known (NFR1/NFR5) and therefore the other LysM domain kinases must be involved in other aspects and not in nod factor signaling. As shown here, this is not true and complete nod factor signaling is achieved through different receptor complexes. In this careful and comprehensive study of NFRe, the authors reveal the contribution of a second receptor kinase and show that it is necessary for the complete signaling required to attain wildtype levels of nodulation. NFRe contributes by providing different spatio-temporal signaling and does so through partnership with NFR5. These findings are unexpected but that data are convincing (with one caveat mentioned below) and demonstrate the contribution of NFRe which, in Lotus japonicus, plays a minor role relative to the dominant receptor NFR1. Not only does the study illustrate complexities in nod factor signaling, it is very likely that it will provide the basis for answers to differences observed in nodulation signaling among the legume where currently significant differences in the phenotypes of the receptor mutants are not understood. The NFR1/NFRe duplication occurs in all legumes and different levels of contribution from the two paralogs may explain many of the differences between legumes. These data will cause the symbiosis field to reconsider nod factor signaling.

We thank reviewer #2 for pointing out the current status of understanding in the field and the relevance of our new findings.

One of the major claims is a difference in spatial and temporal expression patterns of the two receptors. While the Q-RT-PCR and promoter-GUS data are convincing, I found it difficult to see differences between NFR1 and NFRe promoter-YFP-NLS (Figure 2). Maybe additional panels could be added to illustrate the expression differences in the susceptible zone.

We have revised the manuscript and included additional results on the NFR1/NFRe expression pattern. Figure 2 contains now the expression pattern of the two receptors as observed from one optical section. This illustrates the epidermal expression of NFRe and the whole root expression of NFR1. Supplementary Figure 5G, h illustrates the differences in the expression pattern of the two receptors in the root hairs at 3 days post inoculation with M. loti. The lower level of expression of NFR1 in the root hairs is clearly visible in spite of the high stability of the triple YFP protein.

Reviewer #3:

In this paper, the authors report that an epidermal LysM receptor kinase (LYS1/NFRe) perceives Nod factors and contributes to rhizobial symbiosis, probably by keeping the epidermal cells at an "idling" state. Although the involvement of NFRe in the symbiotic signaling is evident from the results with the corresponding mutants but the detailed mechanism by which NFRe contributes to the symbiosis is still largely obscure.

We thank reviewer #3 for pointing out that our study based on mutant analyses shows clearly the role of NFRe in the symbiosis signalling. We show that Nod factor perception by NFRe is important for regular calcium spiking in the root hairs, for early gene activation after Nod factor treatment and nodule organogenesis. We also show that NFRe and NFR1 kinases have a differential downstream signalling and that the activation loop of the two receptors is the major determinant for this distinct activity. Our findings contribute with important pieces for our understanding of Nod factor signalling and as mentioned by reviewer #2, these will likely change the view of the field on this matter.

Importantly, biochemical characterization of NFRe contains several problems that raise serious questions on the "receptor" function of NFRe.

We present below a detailed response to the comments on the biochemical characterisation of NFRe binding capacities.

The authors analyzed the interaction of the heterologously expressed ectodomains of NFRe and NFR1 using bilayer interferometry, where the biotinylated derivatives of a Nod factor and (GlcNAc)5 were immobilized onto the sensor tip. The results are quite strange. Potato lectin, which is known to bind chitin oligosaccharides but nothing to do with symbiosis, showed a very high affinity (Kd, 30 nM) to the Nod factor, while both NFRe and NFR1 showed thousand times lower (!) affinities to the Nod factor. The affinities of NFRe and NFR1 to the Nod factor (Kd, approximately 30 μM) are also far low compared to the biological activity of Nod factors. In general, the instrument used here seems to give a rather high estimate of the affinity, because the affinity of potato lectin to (GlcNAc)5 obtained here (Kd, 5.2 nM) seems to be much higher than the traditional estimate of the affinity of this lectin to chitin oligosaccharides (The lectins, Academic Press). Thus, it is quite questionable whether such a low affinity of NFRe and NFR1 to the Nod factor observed in this work reflects a biologically significant interaction. I am afraid that such a very low affinity simply reflects some non-specific interaction of these proteins with the hydrophobic region of Nod factor, fatty acid moiety.

To avoid such a possibility, the authors should incorporate inhibition type experiments in their binding studies. If the observed binding of NFRe and NFR1 to the Nod factor reflects a biologically significant interaction, it should be inhibited by the biologically active Nod factor but not by inactive Nod factors or chitin oligosaccharides.

We acknowledge and agree with the reviewer that the audience can be misled by the lectin control. The sole purpose of the potato lectin control was to demonstrate correct immobilization and presence of the carbohydrate conjugates on the BLI biosensors, as it is a known binder for N-acetylglucosamine moieties, although its dissociation constant might be affected by avidity. We have exchanged it for a well-studied control, which is the ectodomain of the well-characterized AtCERK1 receptor. The dissociation constant for CO5 to AtCERK1-ECD on BLI was Kd = 59 µM. This value is equal to the previously published dissociation constant Kd = 66 µM for this interaction published by Liu et al., (2012), who used a similar expression and purification protocol and obtained a crystal structure of the AtCERK1-ECD – CO5 complex.

Dissociation constants in the µM concentration range have been observed for multiple LysM receptor kinase ectodomains to various ligands (Liu et al., 2012, Kawaharada et al., 2015, Liu et al., 2016, Bozsoki et al., 2017). Nanomolar affinities have been published for full-length receptors (Iizasa et al., 2010, Broghammer et al., 2012), where avidity between membrane-bound proteins influences the binding affinity. In summary our binding studies are comparable to previous results with LysM ectodomains of plant receptor kinases and the results from nfre mutant studies demonstrate a role in Nod factor signalling.

There are some problems for the characterization of the expressed proteins too. Although the authors described "Proteins of high purity were obtained after four steps of purification, as confirmed by the homogeneous and sharp peaks observed after size exclusion chromatography" (Subsection “NFRe is a Nod factor receptor with an active intracellular kinase”), the results in Figure 1—figure supplement 2B shows the presence of multiple peaks for NFR1-ECD. If the authors used only the peak indicated as NFR1-ECD, they should describe it in the figure legend. It is better to show the elution profiles of final preparations of NFRe/NFR1-ECD by using the same column. The authors should also explain why these preparations showed multiple bands in SDS-PAGE (Figure 1—figure supplement 2C). It is also questionable why the sizes of the smallest bands for NFR1-ECD and NFRe-ECD are very different, while the expected sizes of these proteins are quite similar. I am afraid that some limited proteolysis happened for NFR1-ECD.

We understand the concerns of the reviewer, as the presence of a single band is usually a quality indicator for a protein preparation. Indeed, as correctly noted by the reviewer, the SDS-PAGE shows multiple bands. However, these are not due to limited proteolysis, but instead to different N-glycosylation patterns. LysM receptor kinase ectodomains have multiple disulphide bridges in close proximity and several Nglycosylation sites (see crystal structures in Liu et al., 2012, Liu et al., 2016, Boszoki et al., 2017 as well as Mulder et al. 2006 shows heavy glycosylation for Medicago NFP) and therefore recombinant production requires a eukaryotic expression system capable of providing these modifications to ensure correct folding and good protein quality. This can be seen on the size exclusion profiles, where the ectodomains elute as single peaks, while the void volume peaks contain contaminants. When the recombinant proteins are treated with an endoglycosidase, e.g. PNGase F, the bands collapse to one single, sharp band, clearly demonstrating that the different bands are due to the presence of N-glycans and not to partial degradation. We have followed the reviewer’s suggestions and expanded the figure captions to address N-glycosylations and peaks on the size exclusion profile. We also show an elution profile of NFR1-ECD on the same column as NFRe-ECD.

Characterization of the nfre mutants clearly showed that NFRe contributes to symbiotic signaling and nodule formation. However, the results of the overexpression experiments seem to be difficult to understand. In the nfr1 background, slight increase of the copy number of Nfre gene resulted nothing but the overexpression of the same gene induced a limited symbiotic response induced by M. loti, i.e., activation of Nin promotor in the epidermal cell layer. This Nin activation was not associated with the formation of nodules. Is it safe to connect the results only observed with the overexpression experiments to physiological function? Does the result fit to the hypothesis of "idling" function of NFRe?

We agree with reviewer #3 that caution needs to be taken when interpreting the results from overexpression experiments. In the context of our study, we believe it is important to understand whether NFRe has the capacity to induce Nin induction by Rhizobium when constrains related to the expression level (own or NFR1 promoters) are not present. Furthermore, in order to understand, what is the mechanism behind NFR1 and NFRe differential signalling we have performed direct comparisons between receptors driven by the same promoter. All together the results from biochemical studies of the ectodomain and intracellular kinase, the expression pattern of the Nfre and Nfr1, the results from nfr1 and nfr5 complementation and from the characterisation of the nfre mutant point towards this model where NFRe, most likely together with NFR5, maintains the epidermal layer of the growing root attuned but not committed to symbiosis, state that ensures a rapid switch by NFR1/NFR5 master regulators.

The Venn diagram shown in Figure 3F indicates the presence of significant differences between nfre-1 and nfre-2. Is it OK?

The nfre-1 and nfre-2 are independent alleles that differ in the number and location of LORE1 retroelement insertions, consequently, a difference in the expression pattern is expected. We considered the common differentially expressed genes for comparisons to wild-type, to be NFRe dependent.

Would it be better to confirm the results of nfre mutants by complementation?

We agree that genetic complementation is indeed a powerful tool and based on the nfre phenotype envision that stable genetic lines would be needed. We provide compelling evidence for the role of NFRe in symbiosis based on analyses of three independent alleles which is in accordance with the current accepted standards for genetic analysis.

[Editors’ note: the author responses to the re-review follow.]

Although we acknowledge the technical issue, it is difficult that a method hardly used in the field should explain obvious discrepancies noted by the referees. In this context, it also is somewhat problematic that if the reviewers do not catch it, the majority of readers won't either. Thus, the binding studies should be complemented with additional assays, using a more frequently used reliable.

The main objection towards our manuscript seemed to be based on doubts about the validity of the results from bio-layered interferometry (BLI). The scientific reasoning behind employing BLI approach is that Nod factors, because of the hydrophobic fatty-acid chain and well-known sticky properties, are difficult to work with in solution and need to be immobilized. One major concern from the reviewers was based on the assumption that BLI does not produce binding affinities that are comparable to other methods. To this end, we have now expressed and purified the ectodomain of AtCERK1, which is one of the best studied LysM receptors, and performed BLI measurements to assess binding affinities to chitopentaose and R7A Nod factor, thus providing a previously characterized, published and broadly accepted control for LjNFRe ectodomain binding. We show that AtCERK1 ectodomain has a dissociation constant of KD = 59 µM to chitopentaose and therefore clearly demonstrate that our BLI measurements indeed are robust and comparable to independently published affinity measurements using ITC (Liu et al., 2012; KD = 66 µM)). We would like to stress that the BLI method was already established in 2005 and has now been used in more the 650 published studies including 3 publications in eLife.

A revised manuscript has been submitted for your consideration. It contains the aforementioned results on AtCERK1 ectodomain, the activities of Nfr1 and Nfre promoters in the susceptible zone of the root, together with additional revisions according to reviewer’s suggestions.

[Editors’ note: the author responses to the re-review follow.]

1) The highly volatile nature of the spiking frequency makes it very difficult to determine fundamental differences. Thus, more detailed dissection of the patterns is required to support your conclusions. Likewise, strong statements in the text and the abstract should be softened in order to reflect the strength of the phenotype observed in the nfre1-1 mutant.

We agree that analysis of calcium spiking pattern requires careful analysis and interpretation, which we believe it has been done in this manuscript and revealed a quantitative difference between nfre-1 and wild-type. To this end we have i) investigated the effect of Nfre mutation on calcium spiking in stably transformed plants avoiding the interference from a disturbed hormonal balance and growth patterns in hairy roots. ii) investigated a large number of nuclei (50 for wild-type and 46 for the nfre-1) to capture the existing diversity in patterns. This is a much larger number than previously used for such detailed microscopy investigations. iii) used an automated software for unbiased analysis of the raw data and interpretation. This type of analysis has been used previously to highlight differences in spiking regularity between MYC- and NOD-induced signals. (Russo et al., 2013)

We have incorporated the reviewer comments into the text that reads now:

Abstract: “Mutants of Nfre react to Nod factors with significantly increased calcium spiking interval, reduced transcriptional response and fewer nodules in the presence of rhizobia.” subsection “NFRe perceives Nod factor and has an active intracellular kinase”: “Our findings provide evidence for a complex Nod factor signalling where NFRe activity in the outer root cell layers aids in maintaining a normal calcium spiking interval in the root hairs, integral transcript responses in the susceptible root zone, and initiation of nodule primordia on the expanding root system.”

Subsection “NFRe promotes nodule organogenesis”: “Closer inspection of the spiking frequency revealed that the average inter-spike interval was significantly longer in the nfre cells (106s) compared to wild type (86s), indicating that NFRe contributes to a constant interval length of calcium oscillations (Figure 3—figure supplement 2).”

Subsection “The activation segments of NFR1 and NFRe determine the signalling output”: “Together, these results show that NFRe represents an influential component of the epidermal Nod factor signalling in L. japonicus, promoting intracellular signalling that leads to optimal calcium spiking, activation of gene transcription and efficient nodule organogenesis on the expanding root system.”

2) The referees believe that (transient) complementation of the nfre1-1 mutant is required to overcome problems associated with quantitatively differing phenotypes observed in the three nfre1-1 mutant alleles.

We have performed the genetic complementation analysis of the nfre-1 mutant and the results show clearly that the symbiotic phenotype of nfre-1 is caused by mutation in the Nfre gene. These results are now integrated in the revised manuscript Figure 3—figure supplement 1A.

3) The use of inhibition type experiments is mandatory in addition to the differential binding experiments. This is important in order to demonstrate specificity of binding, which in addition to ligand affinity is a major criterion for biologically meaningful receptor-ligand interaction. Use of Nod factor, chitin oligosaccharide or exopolysaccharide as competitors in biolayer interferometry are likely to strengthen your claim about the receptor function of NFRe.

You may also wish to show that the biotinylated R7A Nod factor retains the biological activity.

Analyses of receptor-ligand biochemistry in competition assays have been reported so far for successful studies where specificity, or lack of it, could be shown. As correctly pointed by the reviewers, this is an analysis that can provide important information on receptor specificity. We have therefore set out to investigate this in the revision time. We have performed additional expression and purification experiments, but the production of the required amount of protein for competition studies on BLI was not achievable. Our experience based on long term experimentation (J. Cheng- whole PhD study) is that NFR1 and NFRe ectodomains are among the most challenging LysM proteins in terms of expression and purification. We have, nevertheless used the very limited amount of reliable NFRe and NFR1 proteins to perform competition assays based on R7ANod factor/CO5 -coated streptavidin beads in comparison with CERK6, the Lotus japonicus chitin receptor (Bozsoki et al., 2017). Unfortunately, the results obtained from two independent experiments are totally inconclusive. We provide the details of our results in the tables below together with representative WB figures as information for editorial and reviewer’s consideration. Our experience is in line with the current literature where no competition binding assays have been reported for any of the ten LysM receptors that have been biochemically and functionally characterised so far in rice, Arabidopsis and Lotus, indicating that such classical competition studies for carbohydrate LysM receptor studies might not resolve an intricate specificity issue as originally thought.

We present here, the most refined binding study for 3 LysM receptors to date, with direct binding data to pure and well characterized Nod factor and chitin ligands. These results show that NFRe and NFR1 ectodomains can bind M.loti Nod factor but not CO5 in a biolayer interferometry biochemical assay which is in line with the plant mutant phenotypes showing an impairment in Nod factor but not CO5 signalling. Furthermore, the additional results from the complementation analyses strengthen our conclusion that NFRe is important for Nod factor signalling. The inconclusive results from the competition assays using single receptors point as well towards our proposed model of how complex Nod factor signalling is in planta. It is very likely that multiple LysM receptor assemble into functional signalling complexes and that signalling specificity is the result of the nature of the complex, rather than isolated LysM receptors alone. We have incorporated this view more clearly now in the Discussion section of the revised manuscript (Discussion section) to emphasise the complexity of LysM receptor-mediated signalling and revised the following sections as below:

Subsection “NFRe perceives Nod factor and has an active intracellular kinase”. “In short, our study shows that NFRe can bind Nod factor with comparable affinity as seen for the bona fide Nod factor receptor NFR1 and both receptor ectodomains distinguish Nod factor from pentameric chitin ligands in a biolayer interferometry assay (Figure 1E).”

Discussion section: “Here, we show that the NFR1-NFR5 signalling cascade operates on the framework provided by the epidermal Nod factor receptor NFRe. NFRe and NFR1 share biochemical and molecular properties i.e similar Nod factor-binding affinity, and chitopentaose differentiating capacity when assessed by biolayer interferometry …”

https://doi.org/10.7554/eLife.33506.027

Article and author information

Author details

  1. Eiichi Murakami

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon 0000-0001-6414-0254
  2. Jeryl Cheng

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  3. Kira Gysel

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Data curation, Formal analysis, Supervision, Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Zoltan Bozsoki

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Data curation, Formal analysis, Investigation, Writing—original draft
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-4267-9969
  5. Yasuyuki Kawaharada

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Present address
    Department of Plant Bio Sciences, Faculty of Agriculture, Iwate University, Morioka, Japan
    Contribution
    Data curation, Investigation, Methodology
    Competing interests
    No competing interests declared
  6. Christian Toftegaard Hjuler

    Department of Chemistry, University of Copenhagen, Frederiksberg, Denmark
    Contribution
    Data curation, Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon 0000-0001-6055-6391
  7. Kasper Kildegaard Sørensen

    Department of Chemistry, University of Copenhagen, Frederiksberg, Denmark
    Contribution
    Data curation, Investigation, Methodology
    Competing interests
    No competing interests declared
  8. Ke Tao

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  9. Simon Kelly

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Data curation, Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  10. Francesco Venice

    Department of Life Sciences and Systems Biology, University of Torino, Torino, Italy
    Contribution
    Data curation, Validation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
  11. Andrea Genre

    Department of Life Sciences and Systems Biology, University of Torino, Torino, Italy
    Contribution
    Formal analysis, Validation, Visualization, Methodology
    Competing interests
    No competing interests declared
    ORCID icon 0000-0001-5029-6194
  12. Mikkel Boas Thygesen

    Department of Chemistry, University of Copenhagen, Frederiksberg, Denmark
    Contribution
    Formal analysis, Validation, Visualization, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-0158-2802
  13. Noor de Jong

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  14. Maria Vinther

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  15. Dorthe Bødker Jensen

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  16. Knud Jørgen Jensen

    Department of Chemistry, University of Copenhagen, Frederiksberg, Denmark
    Contribution
    Formal analysis, Supervision, Methodology
    Competing interests
    No competing interests declared
  17. Michael Blaise

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Present address
    IRIM-UMR 9004, Research Institute in Infectiology of Montpellier, University of Montpellier, CNRS, Montpellier, France
    Contribution
    Formal analysis, Supervision, Methodology
    Competing interests
    No competing interests declared
  18. Lene Heegaard Madsen

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Formal analysis, Supervision, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
  19. Kasper Røjkjær Andersen

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Data curation, Formal analysis, Supervision, Validation, Visualization, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-4415-8067
  20. Jens Stougaard

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Funding acquisition, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-9312-2685
  21. Simona Radutoiu

    Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark
    Contribution
    Conceptualization, Supervision, Funding acquisition, Visualization, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    radutoiu@mbg.au.dk
    Competing interests
    No competing interests declared
    ORCID icon 0000-0002-8841-1415

Funding

Danmarks Grundforskningsfond (DNRF 79)

  • Jeryl Cheng
  • Kira Gysel
  • Zoltan Bozsoki
  • Yasuyuki Kawaharada
  • Christian Toftegaard Hjuler
  • Kasper Kildegaard Sørensen
  • Simon Kelly
  • Mikkel Boas Thygesen
  • Maria Vinther
  • Knud Jørgen Jensen
  • Michael Blaise
  • Lene Heegaard Madsen
  • Kasper Røjkjær Andersen

Engineering Nitrogen Symbiosis for Africa (Engineering Nitrogen Symbiosis for Africa)

  • Eiichi Murakami
  • Noor de Jong
  • Jens Stougaard
  • Simona Radutoiu

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to thank Rebecca Fitchett for proofreading the manuscript, Terry Mun for computational analysis support, Mette Hofmann Asmussen for technical assistance and Finn Pedersen for taking care of the plants. This work was supported by the Danish National Research Foundation grant no. DNRF79, and by The Bill and Melinda Gates Foundation, as part of Engineering the Nitrogen Symbiosis for Africa.

Reviewing Editor

  1. Thorsten Nürnberger, Reviewing Editor, University of Tübingen, Germany

Publication history

  1. Received: November 19, 2017
  2. Accepted: June 5, 2018
  3. Version of Record published: June 29, 2018 (version 1)

Copyright

© 2018, Murakami et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 578
    Page views
  • 128
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)