Phytohormonal interactions are essential to regulate plant organogenesis. In response to the presence of signals from symbiotic bacteria, the Nod factors, legume roots generate a new organ: the nitrogen-fixing nodule. Analysis of mutants in the Medicago truncatula CRE1 cytokinin receptor and of the MtRR4 cytokinin primary response gene expression pattern revealed that cytokinin acts in initial cortical cell divisions and later in the transition between meristematic and differentiation zones of the mature nodule. MtCRE1 signaling is required for activation of the downstream nodulation-related transcription factors MtERN1, MtNSP2 and MtNIN, as well as to regulate expression and accumulation of PIN auxin efflux carriers. Whereas the MtCRE1 pathway is required to allow the inhibition of polar auxin transport in response to rhizobia, nodulation is still negatively regulated by the MtEIN2/SICKLE-dependent ethylene pathway in cre1 mutants. Hence, MtCRE1 signaling acts as a regulatory knob, integrating positive plant and bacterial cues to control legume nodule organogenesis.
Legume plants have the ability to develop nitrogen-fixing nodules under starvation conditions and in the presence of a specific bacterial microsymbiont: Sinorhizobium meliloti in the case of the model legume Medicago truncatula. This symbiotic interaction depends on recognition by host-plants of bacterial Nod factors, which elicits root hair deformations allowing penetration of the bacteria into the epidermis (Oldroyd and Downie, 2008). Simultaneously, nodule organogenesis proceeds from the reactivation of differentiated root inner cortical and pericycle cells in front of protoxylem poles, which divide and lead to the formation of a primordium (Crespi and Frugier, 2008). Rhizobia then progress into infection threads towards the growing primordium, which will subsequently differentiate into a mature organ. The M. truncatula nodule has an indeterminate growth, generating a differentiation gradient initiated from a persistent apical meristem (zone I), and consisting of a region (zone II) where rhizobial infection occurs and cell differentiation is marked by the accumulation of amyloplasts, a functional zone where bacteria are differentiated into bacteroids fixing atmospheric nitrogen (zone III), and a senescence zone (zone IV; Vasse et al., 1990). Many plant mutants unable to nodulate were used to decipher the Nod factor signaling pathways (reviewed in Oldroyd and Downie, 2008), leading to the identification of several downstream transcription factors (TFs): the nodulation signaling pathway1 (nsp1) and nsp2 mutants affecting GRAS TFs (Kalo et al., 2005; Smit et al., 2005; Heckmann et al., 2006; Murakami et al., 2006); the bit1 (branched infection threads 1) mutant affecting the AP2/ERF TF MtERN1 (Middleton et al., 2007); and the nin (nodule inception) putative TF (Schauser et al., 1998; Marsh et al., 2007).
Appropriate spatio-temporal integration of various cues including phytohormones is essential to determine developmental outputs, as highlighted in the Arabidopsis root model (Benkova and Hejatko, 2009). For example, cytokinin effects on root elongation and lateral root formation involves regulation of ethylene biosynthesis and the downstream EIN2 signaling pathway (Alonso et al., 1999; Cary et al., 1995; Negi et al., 2008; Ruzicka et al., 2009). In addition, modifications of cytokinin levels or signaling led to changes in PIN expression and protein accumulation in root meristems and lateral root founder cells, likely impacting auxin distribution (Dello Ioio et al., 2008; Laplaze et al., 2007; Negi et al., 2008; Ruzicka et al., 2009). Several phytohormones have been also implicated in the regulation of nodule development. The M. truncatula sickle (skl) mutant is insensitive to ethylene and presents a hyperinfection phenotype leading to supernodulation (nod++; Penmetsa and Cook, 1997; Penmetsa et al., 2003). The SKL gene encodes an ortholog of the Arabidopsis Ethylene INsensitive 2 (EIN2) signaling protein (Penmetsa et al., 2008). In addition, an inhibition of polar auxin transport (PAT) followed by auxin accumulation in dividing pericycle and cortex cells is a prerequisite for indeterminate nodule initiation (Mathesius et al., 1998; Wasson et al., 2006). This inhibition may involve PIN proteins, as silencing of several of these auxin efflux carriers reduces nodulation (Huo et al., 2006).
Cytokinins play a major role in the control of nodulation (Frugier et al., 2008). Systematic RNAi targeting of the M. truncatula cytokinin receptors revealed that the MtCRE1 (Cytokinin Response 1) histidine kinase controls nodule formation (Gonzalez-Rizzo et al., 2006). Similarly, the Lotus japonicus hit1 (hyperinfected 1) mutant carrying a loss of function mutation in the LHK1 gene (orthologous to MtCRE1) showed strongly reduced nodulation (Murray et al., 2007). The L. japonicus phenotype was associated with increased proliferation of infection threads which remained however limited in M. truncatula CRE1 RNAi roots. In addition, the Lotus snf2 gain-of-function mutation affecting the same LHK1 receptor led to cytokinin hypersensitivity and spontaneous nodule formation in the absence of rhizobia (Tirichine et al., 2007). This result indicates that cytokinin signaling is necessary and sufficient to activate cortical cell divisions leading to nodule organogenesis. Genetic interactions showed that the LHK1 pathway acts downstream of Nod factor perception and upstream of the activation of early nodulation-related TFs (Tirichine et al., 2007). Cytokinin signaling relies on a histidine–aspartate multistep phosphorelay ultimately activating type-A response regulators (RRs), which are rapidly transcriptionally induced by cytokinins (Werner and Schmulling, 2009). In M. truncatula, the RR4 type-A RR is indeed transcriptionally induced by cytokinins and also upregulated upon rhizobial infection depending on the upstream Nod factor signaling (Gonzalez-Rizzo et al., 2006). Conversely, cytokinin regulation of the NIN early nodulation marker depends on MtCRE1/LHK1 (Gonzalez-Rizzo et al., 2006; Murray et al., 2007).
Characterization of TILLING mutants affecting the CRE1 cytokinin receptor in M. truncatula allowed us to analyze the integration of cytokinins with rhizobial and plant cues in nodule organogenesis. Despite an early inhibition of cortical cell divisions during nodule initiation, cre1 mutants developed few nodules showing perturbed differentiation and frequently multiple lobes. Accordingly, activation of cytokinin responses occurred both in nodule primordia and in the apical region of mature indeterminate nodules. Analysis of interactions between cytokinins and other cues regulating nodulation revealed that ethylene acts independently of the MtCRE1 cytokinin signaling, whereas this latter pathway acts upstream of auxin, regulating PIN protein expression, accumulation and PAT inhibition in response to rhizobia. Hence, MtCRE1-dependent cytokinin signaling integrates signaling from exogenous Nod factors as well as endogenous cues, such as auxin, to regulate indeterminate nodule organogenesis.
Nodulation phenotypes of a cre1 mutant allelic series
The kinase domain of the MtCRE1 cytokinin receptor was selected for TILLING (Le Signor et al., 2009). Three single nucleotide mutations introducing either a stop codon in the middle of the kinase domain (cre1-1) or single amino acid changes (cre1-2 and cre1-3) were characterized (Figure 1a,b). The cre1-2 and the cre1-3 mutations affected, respectively, a conserved residue in the G2 motif and a non-conserved region of the kinase domain. Analysis of root growth sensitivity to cytokinins revealed that both cre1-1 and cre1-2 were insensitive to BAP (benzylaminopurine) at 10−7m in contrast to the cre1-3 allele and wild-type controls (Jemalong A17 and a wild-type sibling line of the cre1-1 allele; Figure 1c,d). We then determined the ability of cre1 mutants to form symbiotic nodules in response to S. meliloti inoculation (Figure 2). In contrast to the cre1-3 allele, populations segregating cre1-1 or cre1-2 showed a strongly impaired nodulation in homozygous mutants (Figure 2a). Use of an S. meliloti strain carrying a ProHemA:LACZ fusion revealed that infections generally aborted in the epidermis or in the outer cortex, and did not lead to cortical cell division (Figure 2b). In some cases, infection was associated with cortical cell divisions and primordium formation, but infection threads showed many sac-like structures and ramifications despite their capacity to colonize the nodule primordia (Figure 2c). Both cre1-1 and cre1-2 alleles can develop a few nodules (Figure 2a), even though their globular shape, when compared with the elongated wild-type nodules, confirmed a delayed development (Figure 2d). Histological analysis of wild-type and cre1 nodules revealed structural differences even when nodules of equivalent sizes were compared (i.e. 20 d.p.i. cre1 nodules versus 7 d.p.i. wild-type globular nodules; Figure 2e,f). Amyloplast accumulation and autofluorescence associated with rhizobia-infected cells indicated an incomplete differentiation of cre1 nodules. Later on (2–3 months after inoculation), nodules frequently bare multiple lobes in contrast to the wild-type which were mainly unilobed (Figure S1), suggesting that several meristems where initiated in the mutant nodules.
These results indicate that cytokinins, apart from their crucial role in early steps of nodule organogenesis, may regulate the transition between meristematic and cell differentiation/elongation zones in indeterminate nodules.
Cytokinin responses are activated in dividing cortical cells and in the apical region of mature nodules
To determine the spatio-temporal regulation of cytokinin signaling during nodulation, the A-type Response Regulator primary cytokinin response gene MtRR4 was used to generate a transcriptional fusion with the GUS reporter (Figure 3a–f). Histological analysis of S. meliloti-inoculated roots showed that the ProMtRR4:GUS fusion was expressed in inner cortex and pericycle (Figure 3a–c), and then in dividing cortical cells and nodule primordia (Figure 3d,e). ProMtRR4:GUS expression was associated in mature nodules with vascular bundles and the apical region (Figure 3f). Use of amyloplasts lugol staining and infected cells autofluorescence as markers revealed that MtRR4 was expressed both in the meristematic zone I and infection zone II (Figure S2), in agreement with the expression pattern previously identified using in situ hybridization (Vernie et al., 2008). To further sustain a role of cytokinins in late nodulation, we analyzed expression patterns of other signaling components (Figure 3g–i). In situ hybridization revealed an overlapping expression domain of the MtCRE1 receptor, B-type RRs and A-type RRs in the apical region of differentiated nodules. Hence, late nodulation cre1 phenotypes correlate with the cytokinin signaling expression domain, suggesting a role of this phytohormone in the regulation of cell proliferation and/or differentiation in indeterminate nodules.
The MtCRE1-dependent cytokinin pathway is required to regulate Nod factor signaling genes
To analyze the interplay between Nod factors and cytokinins, we determined the cytokinin regulation of TFs acting downstream of this bacterial signal using real time RT-PCR (Figure 4). Similarly to MtRR4, a positive control for short-term cytokinin response, the early nodulation-related TFs MtNIN, MtERN1, and MtNSP2 were rapidly upregulated in roots exposed to cytokinins, even though various patterns could be identified (Figure 4a). MtNSP2 showed a transient induction after 1 h of BAP 10−7m treatment but was downregulated after 3 h, whereas a sustained induction was observed for MtRR4, MtNIN and MtERN1. In contrast to the other transcripts analyzed, MtNSP1 showed a much weaker regulation by cytokinins. Using a cycloheximide (CHX) pre-treatment, we determined whether these transcriptional regulations were directly linked to cytokinin signaling (Figure 4b). Besides the MtRR4 primary cytokinin response gene, none of the nodulation-related TFs tested showed a cytokinin response in presence of CHX indicating that de novo synthesis of other regulators was required. As the MtCRE1 cytokinin receptor is essential for nodule organogenesis, we analyzed transcriptional regulations in the cre1-1 mutant background (Figure 4c). After a short-term cytokinin treatment (BAP at 10−7m for 1 h), regulations of MtRR4, MtERN1, MtNSPs and MtNIN were lost in the mutant. Hence, transcriptional regulation of several nodulation-related TFs in response to cytokinins is mainly or even exclusively dependent on the MtCRE1 cytokinin receptor.
Interaction between MtEIN2/SKL-dependent ethylene and MtCRE1-dependent cytokinin pathways during nodulation
In M. truncatula, ethylene is a negative regulator of nodule formation (Penmetsa and Cook, 1997). Application of the broadly used ethylene biosynthesis inhibitor AVG (aminoethoxy vinyl glycine) simultaneously to rhizobial infection increased nodulation (Figure 5a). Similarly, cre1-1 and cre1-2 mutants showed increased nodulation in the presence of AVG, even though nodule numbers remained strongly reduced compared with the wild-type (Figure 5a). These results suggest that cre1 mutants are still sensitive to the inhibitory effect of ethylene at early stages of nodule formation.
To analyze more specifically the interactions between cytokinin and ethylene, the cre1 mutant was crossed to the sickle supernodulating mutant affected in the EIN2 ethylene signaling component (Penmetsa et al., 2008). The double mutant skl/cre1 had a reduced nodulation ability when compared with the wild-type, but showed enhanced nodulation when compared to cre1 (Figures 5b and 6a). This finding suggests that the inhibitory EIN2-dependent ethylene pathway acts in parallel to the CRE1-dependent cytokinin pathway to control nodule number. Histology of the few globular nodules formed on skl/cre1 roots, analyzed using lugol staining and autofluorescence as markers (Figure 6b–d), showed that nodule zonation in skl/cre1 is more similar to skl than to cre1. Indeed, lugol staining clearly identified the meristematic and infection regions (zones I and II) in contrast to cre1 nodules where the differentiation gradient was perturbed.
Overall, the epistatic interactions between skl and cre1 along nodulation point to different outputs of the cytokinin and ethylene interplay at various developmental stages.
MtCRE1 signaling is required to regulate specific PIN efflux carriers accumulation and Rhizobium inhibition of polar auxin transport
As PAT was shown to be crucial for indeterminate nodule formation, we analyzed by real-time RT-PCR expression of six PIN auxin efflux carriers detectable in M. truncatula roots and nodules (based on Schnabel and Frugoli, 2004; Table S1). In response to a short-term exogenous cytokinin treatment, no major change in PIN expression was observed in whole roots, except for MtPIN9 (a homolog of AtPIN5; Schnabel and Frugoli, 2004) which was strongly downregulated (Figure S3a–c). Steady state levels of PIN expression were determined in wild-type and cre1 roots, revealing a significant accumulation of MtPIN3 and MtPIN6 transcripts and reduced MtPIN9 levels in the mutant background (Figure 7a), mainly observed in root apices (Figure S4). We then developed an antibody (PIN62, see Experimental procedures) able to recognize most of the Medicago PIN auxin efflux carriers homologous to Arabidopsis proteins polarly localized in plasma membranes and proposed to be crucial for PAT (i.e. against the Medicago orthologs of AtPIN1 to 4, AtPIN6 and AtPIN7; Table S1; Petrasek and Friml, 2009). As expected, analysis of PIN localization pattern in wild-type Medicago roots revealed membrane polarly-localized signals (Figure 7b). Interestingly, PIN proteins were notably detected in inner cortical cell layers (Figure 7c), and stronger signals were observed in cre1 root apices (Figures 7d,e; a quantification of PIN-polar signals is provided in Figure S5). To further support this result, we analyzed PAT in wild-type and cre1 roots. Direct acropetal auxin transport measurements revealed that cre1 roots have an increased PAT rate (Figure 7f). In addition, the PAT inhibition observed in wild-type in response to S. meliloti inoculation was not detected in cre1 (Figure 7g). We then determined whether this deregulation was correlated with changes in PIN expression under symbiotic conditions. Whereas a high variability of PIN expression patterns was observed a few hours after rhizobia inoculation (data not shown), a significant accumulation of MtPIN4 and MtPIN10 transcripts and a decreased level of MtPIN9 expression were detected in response to a Nod factor treatment (10−8m, 1–6 h) in wild-type roots (Figure 7h). In the cre1-1 mutant, MtPIN9 downregulation was significantly retrieved, and similar trends were observed for MtPIN4 and MtPIN10, indicating that a regulation of PIN expression by Nod factors still occurred in cre1.
Hence, PAT regulation in roots depends on the MtCRE1 cytokinin pathway both under symbiotic and non-symbiotic conditions, as well as changes in the expression and accumulation of specific PIN auxin efflux carriers.
Our work showed various roles for cytokinins and their interactions with different cues during nodule organogenesis, from initiation of cortical cell divisions to regulation of nodule growth and differentiation. Two of the cre1 TILLING alleles identified in this study, one of those generating a nonsense mutation, showed similar cytokinin insensitivity and nodulation phenotypes. In both alleles, no obvious phenotype was detected in shoots, reinforcing that MtCRE1 has a major contribution in below-ground organs development. As previously shown in CRE1 RNAi roots infected by S. meliloti (Gonzalez-Rizzo et al., 2006), infection threads were mainly blocked in the cre1 mutant epidermis, and no extensive invasion of outer root cell layers was observed. In the few cases where primordia formed, infection threads developed many sac-like structures and ramifications, and infection was delayed. In L. japonicus, cytokinin pathways activation in the epidermis was proposed based on the use of the Arabidopsis ProARR5:GUS fusion (Lohar et al., 2004). Analysis of soybean root hair transcriptome in response to rhizobia did not reveal however any enrichment for endogenous transcripts related to cytokinin signaling (Libault et al., 2010). Using the endogenous MtRR4 primary cytokinin response gene, we showed that an activation of cytokinin pathways occurred in inner cortex and pericycle but not in outer cortex or epidermis where most of cre1 infections were blocked. Accordingly, expression of MtCRE1 was previously detected in cortical cells but not in the epidermis (Lohar et al., 2006). These results support that in Medicago, cytokinins primarily act in dividing inner cortical cells. Interestingly, the functionally related LHK1 pathway in L. japonicus was recently unambiguously associated to nodule organogenesis acting in the cortex, based on genetic interactions (Madsen et al., 2010). The contrasting cytokinin responses observed in Lotus and Medicago epidermis may reflect different functions for this phytohormone depending on host-plants and/or in feedback mechanisms generated by dividing cortical cells to control epidermal infection. Interestingly, hyperinfection is observed in M. truncatula skl ethylene-insensitive mutant but not in cre1, whereas a hyperinfection phenotype was reported in the L. japonicus cytokinin insensitive hit1 mutant (for hyper infection threads; Murray et al., 2007), but not in the recently identified ethylene insensitive enigma mutant affecting LjEIN2 (Gresshoff et al., 2009). In contrast, enigma mutants are defective in nodulation and infection thread formation. More generally, the variety of nodule organogenesis (i.e. cell divisions activated in outer versus inner cortex, or determinate versus indeterminate growth) are likely linked to different requirements for phytohormonal regulations.
In Lotus and Medicago, mutants affecting CRE1/LHK1 are similarly able to develop a few nodules. cre1 nodule organogenesis was delayed, leading to differentiation defects and formation of multiple lobes, which correlated with the expression of cytokinin signalling components in the nodule apical region. The increased cre1 nodule size may be however partly explained by a compensation effect of the low nodule number, as reported in other mutants (Libault et al., 2009; Magori et al., 2009). In Arabidopsis roots, cytokinins promote the exit of cells from meristems to elongate and differentiate (Dello Ioio et al., 2007). Our results suggest that cytokinins may act in mature nodules to regulate the balance between cell proliferation and differentiation.
Many early nodulation markers are transcriptionally regulated by cytokinins (Frugier et al., 2008), including MtNIN whose regulation depends on the MtCRE1/LHK1 pathway (Gonzalez-Rizzo et al., 2006; Murray et al., 2007). We now show that other TFs transcriptionally regulated by rhizobia and acting upstream of NIN are regulated by cytokinins through CRE1. In all cases however, de novo synthesis of other transcriptional components was required for this activation. Interestingly, MtNSP2, one of the most upstream TF acting in early Nod factors signaling both in cortical and epidermal cells (Oldroyd and Downie, 2008) shows a very dynamic regulation by cytokinins. The MtERN1 TF has been primarily linked to infection thread formation and ENOD11 expression in root hairs (Andriankaja et al., 2007; Middleton et al., 2007), but is however also expressed in spontaneous nodules formed in the absence of Rhizobium (Gleason et al., 2006). Similarly, NIN is associated with epidermal and cortical responses (Marsh et al., 2007). Based on our results, it is likely that the CRE1-dependent early induction of these nodulation-related TFs by cytokinins takes place in the cortex to activate nodule organogenesis.
Appropriate spatial and temporal regulation of signaling pathways and their interactions is essential to determine developmental outputs. In Arabidopsis roots, an interplay between auxin, ethylene and cytokinins occurs, ethylene controlling root elongation and lateral root formation through complex interactions with auxin biosynthesis and polar transport (Benkova and Hejatko, 2009; Negi et al., 2008). Our results show that, in cre1, nodulation is still repressed even if the ethylene pathway is inhibited through mutation of ein2/skl. Previously, the expression domain of an ACC oxidase enzyme, catalyzing ethylene biosynthesis, was localized between protoxylem poles, and negatively correlating with nodule initiation sites (Heidstra et al., 1997). The spatial regulation exerted by ethylene on nodule initiation may then restrict the domain where activation of cytokinins should occur, i.e. in cells in front of protoxylem poles. Our results additionally suggest that in mature indeterminate nodules, the balance between proliferation and differentiation may involve interactions between ethylene and cytokinins.
Auxin was recently shown to activate A-type ARR expression in the basal pole of Arabidopsis embryos corresponding to root stem cells, potentially leading to an inhibition of the cytokinin pathway (Muller and Sheen, 2008). Conversely, exogenously applied cytokinins regulate the spatiotemporal accumulation of PIN auxin efflux carriers in roots, suggesting that endogenous levels of cytokinins perturb auxin fluxes necessary for lateral root initiation (Laplaze et al., 2007) and to determine root meristem size and growth rate (Dello Ioio et al., 2008; Ruzicka et al., 2009). AtPIN1, PIN3 and PIN7 were accumulated in the root apex of the ahk3 (authentic histidine kinase 3) mutant, and their regulation by cytokinins was impaired in various cytokinin receptor mutant combinations. In Medicago cre1 root apices, a higher accumulation of PIN proteins was detected, notably in the inner cortex that is involved in nodule organogenesis. In addition, analysis of individual PIN gene expression revealed increased levels of MtPIN3 and MtPIN6 transcripts in cre1 roots. These expression patterns correlated with an increased PAT capacity of cre1 roots, likely leading to an increased accumulation of auxin in apices. The lack of PAT inhibition in response to rhizobia inoculation observed in cre1 further revealed that MtCRE1-dependent cytokinin signaling acts upstream of auxin-related pathways crucial for root and nodule organogenesis. The different PIN genes changing expression after a short-term Nod factor treatment do not however correspond to the ones whose transcripts accumulate in the cre1 mutant. Moreover, similar regulations were retrieved in the cre1 mutant, even though not significant for MtPIN4 and MtPIN10, indicating that response to Nod factors may be fainter or delayed in the mutant. In Arabidopsis, post-transcriptional modifications, such as phosphorylation, have been linked to PIN proteins function (Titapiwatanakun and Murphy, 2009), and these levels of regulation may be also involved in the response to rhizobia. In addition, other components of the PAT machinery or of the auxin response may be required for symbiotic nodule organogenesis depending on the CRE1 pathway. Finally, we cannot discard the hypothesis that cre1 roots may have became resistant to the PAT inhibition induced upon symbiotic bacteria inoculation indirectly, due to their increased auxin flux associated to the accumulation of PINs in the meristematic region.
Overall, our results show that MtCRE1 signaling integrates positive bacterial and plant cues (i.e. Nod factors and auxin) to temporally and spatially regulate nodule initiation. In addition, cytokinins and their interactions likely exert a continuous control on differentiation and growth of indeterminate nodules. Coordinated regulation of the cytokinin response by the host plant and its symbiont is therefore a crucial target to control various stages of nodulation, and may also be essential to determine the different types of nodule organogenesis and growth.
Medicago truncatula Jemalong A17 seeds were sterilized for 20 min in bleach (12% [v/v] sodium hypochlorite). After washing with sterilized water, seeds were sown on 1% agar plates, stored for 2 days at 4°C before incubating overnight at 24°C in the dark. Germinated seedlings were transferred to square plates containing appropriate medium (see below) and grown vertically in chambers at 24°C under long-day conditions (16 h light at 150 μE light intensity/8 h dark).
cre1 mutant alleles cre1-1, cre1-2, and cre1-3 were identified in this study by TILLING (Le Signor et al., 2009; primers indicated in Table S2). A wild-type sibling line of the cre1-1 allele was also used as control. In addition, the skl-1 mutant (Penmetsa and Cook, 1997; Penmetsa et al., 2008) was used to generate the skl-1/cre1-2 double mutant. The cre1-1 allele was backcrossed successively two times with Jemalong A17, and both cre1-1 and cre1-2 alleles where systematically used for phenotyping. cre1 genotyping was done using PCR with the same primers as the one used for TILLING and subsequent sequencing. Alternatively, based on these amplicons, a Bfa1 restriction was used to genotype cre1-1 (as the mutation generated such restriction site); a Kpn1 restriction to genotype cre1-2 (as the mutation deleted this restriction site); and a Mnl1 restriction to genotype cre1-3 (as the mutation deleted this restriction site). skl genotyping was done using PCR (primers indicated in Table S2) and subsequent sequencing.
For nodulation experiments, germinated seeds were grown in vitro on low-nitrogen liquid medium (‘i’; Blondon, 1964). Roots were inoculated with 10 ml of Sinorhizobium meliloti suspension (OD600 = 0.05) per plate for 1 h. Different bacterial strains were used: a wild-type Sm1021 strain and a derivative strain (GMI6526; Ardourel et al., 1994) carrying the pXLGD4 plasmid containing a ProHemA:LACZ transcriptional fusion. Nod factors were extracted from S. meliloti Sm2011 (GMI6390, pMH682) following the protocol described in Roche et al. (1991).
Hormonal and Nod factors treatments
Fifteen germinated seedlings were placed on a grid in a Magenta box with 30 ml of low-nitrogen ‘i’ liquid medium and grown in a shaking incubator (125 g) at 24°C under long-day conditions (16 h light/8 h dark). After 5 days, seedlings were treated with or without BAP (Sigma-Aldrich, http://www.sigmaaldrich.com/) at 10−7m and maintained under the same growth conditions for various incubation times (0, 1 and 3 h). Similarly, Nod factors purified from S. meliloti were used at 10−8m for 1, 3 or 6 h. A 1 h CHX (Sigma-Aldrich) 50 μm pre-treatment was also used in some experiments. Roots were collected at the indicated time points and immediately frozen in liquid nitrogen for RNA extraction. In all cases, three independent biological experiments were performed.
To test the sensitivity of the cre1 mutant roots to cytokinins and the effect of the ethylene biosynthesis inhibitor AVG (Sigma-Aldrich), roots were grown on growth papers (Mega International, http://www.mega-international.com/index.htm) placed on ‘i’ medium. After 3 days, plants were transferred on a fresh ‘i’ medium supplemented or not with BAP at 10−7m or with AVG at 10−6m. For cytokinin sensitivity, position of root tips was marked at the time of transfer, and root growth from this point was measured after 6 days using ImageJ software (http://rsbweb.nih.gov/ij/). Three independent biological replicates were performed for each experiment.
Gene expression analysis
Total RNA was extracted from frozen roots using the RNeasy plant mini kit (Qiagen, http://www.qiagen.com/). First-strand cDNA was synthesized from 1.5 μg of total RNA using the Superscript II first-strand synthesis system (Invitrogen, http://www.invitrogen.com/). Primer design was performed using Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/). Primer combinations showing a minimum amplification efficiency of 90% were used in real-time RT-PCR experiments (Table S2). Real-time RT-PCR reactions were performed using the LightCycler FastStart DNA Master SYBR Green I kit on a LightCycler apparatus according to manufacturer’s instructions (Roche, http://www.roche.com/). Cycling conditions were as follows: 95°C for 10 min, 50 cycles at 95°C for 5 sec, 58°C for 5 sec, and 72°C for 15 sec. PCR amplification specificity was verified using a dissociation curve (55–95°C). A negative control without cDNA template was always included for each primer combination. Technical replicates (on two independent syntheses of cDNA derived from the same RNA sample) and three independent biological experiments were performed in all cases. Ratios were done with constitutive controls for gene expression to normalize the data between different biological conditions. MtACTIN11, MtRBP1 and MtH3L were selected using Genorm software (http://medgen.ugent.be/~jvdesomp/genorm/; Vandesompele et al., 2002) as reference genes for experiments involving hormonal treatments in wild-type and cre1 mutant roots (primers shown in Table S2). MtRBP1 was chosen to calculate ratios, and the value of the experimental control condition was set up to 1 as a reference to determine relative expression or induction factors. When weak changes in expression were observed, a Mann–Whitney test (available in the Xlstat software; http://www.xlstat.com/) was used to assess significant differences based on the different biological replicates performed.
In situ hybridizations were performed as described in Valoczi et al. (2006) on 15 d.p.i. nodules using an Intavis InsituPro automat (http://www.intavis.com/en/). Sense and antisense RNA probes corresponding to a carbonic anhydrase gene (MtCa1) were included as negative and positive controls, respectively (Coba de la Pena et al., 1997; Boualem et al., 2008). A probe specifically directed against MtCRE1 was designed whereas, due to the extensive homology among B-type or among A-type RRs (Gonzalez-Rizzo et al., 2006), we could not find a discriminatory region for probing each individual gene. A nucleotide identity lower than 70% between A- and B-type RRs in the regions used as probes allowed however to discriminate each family. All probes used are listed in Table S2.
Agrobacterium rhizogenes root transformation
A 3030 bp sequence upstream of the MtRR4 start codon was amplified by PCR using a Pfx polymerase (Invitrogen) and primers MtRR4-5′ and MtRR4-3′ (Table S2). PCR product was subsequently cloned using Gateway technology (Invitrogen) into the pkGWFS7 vector (http://www.psb.ugent.be/gateway/index.php) carrying a GFP–GUS fusion downstream of the cloning recombination site. The resulting constructs were introduced into Agrobacterium rhizogenes ARqua1 (Quandt et al., 1993) and used for M. truncatula root transformation. The transgenic roots were obtained after kanamycin (25 mg/L) selection for 2 weeks as described by Boisson-Dernier et al. (2001). Composite plants were then transferred onto growth papers (Mega International) on Fahraeus medium without nitrogen (Truchet et al., 1985) for 4 days. Thereafter, the transgenic roots were inoculated with an S. meliloti 1021 strain. Roots or nodules were collected at 3, 5, 7, and 15 d.p.i. for histochemical GUS analysis. Three independent biological experiments were performed (n > 20).
Histochemical stainings and microscopic analyses
Histochemical staining for GUS activity was performed as previously described (Pichon et al., 1992; samples were incubated for up to 24 h at 37°C). After staining, nodules were included in 3% agarose and then sliced into 80-μm sections using a VT 1200S vibratome (Leica Microsystems, http://www.leica-microsystems.com/). A 5 min staining with lugol (Sigma-Aldrich) was used to detect amyloplasts. When necessary, samples or sections were cleared briefly with sodium hypochlorite, as described in Pichon et al. (1992), and observed in bright field using a DMI6000B microscope equipped with a DFC300 camera (Leica Microsystems).
Roots and nodules infected by the Sm1021 strain expressing the ProHemA:LACZ were used for β-galactosidase staining, as described in Ardourel et al. (1994). Stained samples were observed in bright field with a Reichert Polyvar microscope equipped with a QImaging Retiga2000 Camera (http://www.qimaging.com/).
For histological analysis, autofluorescence of vibratome sections mounted in water was observed using 450–490 nm excitation/500–550 nm emission filters (green fluorescence of nodule infected cells) or 340–380 nm excitation/450–490 nm emission filters (blue fluorescence of endodermis and flavonoids) on a DMI6000B microscope equipped with a DFC300 camera (Leica Microsystems).
Generation of an antibody recognizing MtPINs proteins and immunolocalization
An antibody able to recognize most of the predicted plasma-membrane located Medicago PIN proteins, referred to as ‘PIN62’ antibody, was generated based on immunization of two rabbits using the TPRASNLSNAEI peptide coupled to the KLH carrier, following a typical Eurogentec immunization protocol (http://www.eurogentec.com/). The sera obtained were purified against the peptide and subsequently tested in ELISA to determine their sensitivity in comparison to pre-immune sera.
One-day-old seedling roots were fixed in 4% PFA using a vacuum desiccator and after inclusion into 3% agarose, root tips were longitudinally sliced into 80-μm sections using a vibratome (Leica VT 1200S, Leica Microsystems). Immunodetection was performed using an Intavis InSituPro automat as described in Friml et al. (2003). PIN62 antibody was used at a dilution of 1:500, and anti-rabbit-Alexa488 or anti-rabbit-Alexa568 secondary antibodies (Invitrogen) were used at 1:1000. Pre-immune serum gave no detectable signal. After excitation at 488 nm or 568 nm, emission of the fluorochrome was detected respectively between 500 and 544 nm or 595 and 610 nm on a SP2 Inverted Confocal Microscope (Leica Microsystems). Quantification of the intensity of PIN polar signals was measured using ImageJ on images obtained using the same confocal configuration for wild-type and cre1-1 root vibratome sections (n = 4 sections/genotype from two independent biological replicates; and n = 30 cells/section).
Polar auxin transport measurements
Auxin transport experiments were done as described by Wasson et al. (2006) with the following modifications. Five-day-old in vitro grown seedlings on nitrogen-free Fahraeus medium were used. To test the effect of S. meliloti inoculation, 4-day-old seedlings were spot-inoculated with a 1 μl drop of S. meliloti culture at the zone of emerging root hairs. The inoculation site was marked on the plate. Roots were placed in the incubator for a further 24 h. Roots were excised 20 mm above the tip and the cut end was placed in contact with a small block of agar containing radio-labelled IAA for 6 h. In Rhizobium infected roots, the radiolabelled auxin was quantified in a 4 mm segment below the point of inoculation, and in an equivalent segment, 8-12 mm below the placement of the radio-labelled auxin block, in untreated roots, thus measuring acropetal (i.e. from root base to root tip) auxin transport. All experiments were repeated three times independently with n > 27.
We want to thank Richard Thomspon (INRA-Dijon, France) and Jonathan Clarke (JIC-Norwich, UK) for the identification of cre1 alleles by TILLING in the frame of the EEC GRAIN LEGUMES project, Marianne Brault-Hernandez for help in some real time RT-PCR experiments, as well as Jason Ng and Britta Winterberg for designing the PIN primers. Nod factors were provided by Fabienne Maillet and Jean Dénarié (LIPM, INRA-CNRS, Castanet-Tolosan). sickle-1 mutant seeds were provided by Douglas Cook (University of California-Davis, USA). Microscopy was done on the IMAGIF platform (CNRS, Gif sur Yvette, France; Marie Noelle Soler and Suzanne Bolte). Antibodies against M. truncatula PINs were developed in the frame of a French ACI program (‘cell polarity and organogenesis’) involving Béatrice Satiat-Jeunemaitre (CNRS-Gif sur Yvette, France) and Jan Traas (Ecole Normale Supérieure, Lyon, France). We thank the ECOS-Sud program A07B03, the ‘Ambassade de France’ and the ‘Banco Santa Fe Foundation’ for providing short-term fellowships to F.A. J.P. was the recipient of a doctoral grant from the ‘Ministère de la Recherche et de la Technologie’ (France). A.W. was supported by an Australian Postgraduate Award. U.M. was supported by funding from the Australian Research Council through an Australian Research Fellowship (DP0557692) and the Centre of Excellence for Integrative Legume Research (CE0348212). F.F. was supported by a French ANR project, LEGUROOT.