Arbuscular mycorrhizal hyphopodia and germinated spore exudates trigger Ca2+ spiking in the legume and nonlegume root epidermis


Author for correspondence:
Paola Bonfante
Tel: +39 011 6705965


  • The aim of this study was to investigate Ca2+ responses to endosymbiotic arbuscular mycorrhizal (AM) fungi in the host root epidermis following pre-infection hyphopodium formation in both legumes and nonlegumes, and to determine to what extent these responses could be mimicked by germinated fungal spore exudate.
  • Root organ cultures of both Medicago truncatula and Daucus carota, expressing the nuclear-localized cameleon reporter NupYC2.1, were used to monitor AM-elicited Ca2+ responses in host root tissues.
  • Ca2+ spiking was observed in cells contacted by AM hyphopodia for both hosts, with highest frequencies correlating with the epidermal nucleus positioned facing the fungal contact site. Treatment with AM spore exudate also elicited Ca2+ spiking within the AM-responsive zone of the root and, in both cases, spiking was dependent on the M. truncatula common SYM genes DMI1/2, but not on the rhizobial Nod factor perception gene NFP.
  • These findings support the conclusion that AM fungal root penetration is preceded by a SYM pathway-dependent oscillatory Ca2+ response, whose evolutionary origin predates the divergence between asterid and rosid clades. Our results further show that fungal symbiotic signals are already generated during spore germination, and that cameleon-expressing root organ cultures represent a novel AM-specific bio-assay for such signals.


Arbuscular mycorrhizal (AM) fungi are ecologically important biotrophic plant symbionts, providing host roots with vital mineral nutrients in exchange for photosynthates and a privileged ecological niche essential for fungal development and propagation (Smith & Read, 2008). Following spore germination, AM fungi respond to host-secreted molecules, such as sesquiterpene lactones, by intense hyphal ramification within the vicinity of the root (Buée et al., 2000; Akiyama et al., 2005). This leads to root contact and the differentiation of fungal adhesion structures, known as hyphopodia, on nonroot hair epidermal cells referred to as atrichoblasts (Harrison, 2005; Parniske, 2008). In return, the plant responds to so far unidentified fungal signals with modifications in gene expression (Kosuta et al., 2003; Kuhn et al., 2010) and the organization of a polarized cytoplasmic assembly, termed the prepenetration apparatus (PPA), in hyphopodia-contacted cells (Genre et al., 2005). PPA-dependent intracellular infection is a conserved mechanism across plant species (Genre et al., 2008), and is associated with the construction of the specialized interface compartment surrounding the infection hypha, a hallmark of the AM association (Parniske, 2008).

Certain of these plant responses depend on a signal transduction pathway which is partly shared with the well-characterized pathway required for the initiation of the nitrogen-fixing root nodule symbiosis formed between legumes and soil rhizobia (Parniske, 2008). In the model legume Medicago truncatula, this so-called ‘common SYM’ pathway (Kistner & Parniske, 2002) includes a putative cation channel (MtDMI1) located in the nuclear membrane (Anéet al., 2004; Riely et al., 2007) and a leucine-rich repeat (LRR) receptor-like kinase (MtDMI2) of unknown function (Endre et al., 2002). The third shared component (MtDMI3) is a nuclear-localized calcium and calmodulin-dependent kinase (CCaMK) (Mitra et al., 2004; Smit et al., 2005). Specific rhizobial lipochito-oligosaccharide signals, known as nodulation factors (NFs), trigger oscillations (or spiking) of the intracellular Ca2+ concentration (Oldroyd & Downie, 2006). This response is dependent on DMI1 and DMI2, whereas the downstream DMI3 CCaMK is believed to decipher the oscillating calcium signal, leading to the activation of a regulatory cascade and specific gene transcription (Gleason et al., 2006).

The fact that hyphopodia are formed on roots of dmi mutants, but that these mutants are defective in both PPA formation and infection (Genre et al., 2005), indicates that activation of the common SYM pathway is necessary for the epidermal cell remodelling which precedes fungal colonization. This leads to the hypothesis that Ca2+ signalling is required for the initiation of PPA formation (Genre et al., 2005; Parniske, 2008). Evidence for a role for Ca2+ signalling during the early stages of the AM association has come from a study in M. truncatula showing that DMI1/2-dependent cytoplasmic Ca2+ spiking can be observed in a percentage of root hairs in the vicinity of ramifying AM fungal hyphae (Kosuta et al., 2008). Since the responding root hairs were not contacted by the fungus, this implies that the putative fungal signal triggering this host response is probably diffusible, in agreement with findings from other studies using a variety of biological assays to detect symbiotic AM signals (Kosuta et al., 2003; Olah et al., 2005; Navazio et al., 2007; Gutjahr et al., 2009; Kuhn et al., 2010).

As stated earlier, AM infection in the majority of legume and nonlegume hosts is initiated from hyphopodia formed on nonroot hair atrichoblasts. One of the principal objectives of this study was therefore to examine whether Ca2+ signalling responses could be observed in hyphopodia-contacted atrichoblasts before infection. For this purpose, we made use of a nuclear-targeted ‘cameleon’ calcium reporter (35S:NupYC2.1) expressed in root organ cultures (ROCs) of both M. truncatula and the nonlegume Daucus carota. ROCs are particularly interesting material for such studies since they can be successfully infected and colonized by AM fungi (Boisson-Dernier et al., 2001), but are unable to associate with rhizobia (Akashi et al., 2003). The NupYC2.1 reporter is a fusion between the nuclear protein nucleoplasmin and the cameleon YC2.1, and has recently been used to demonstrate rhizobial NF-elicited Ca2+ spiking within root hair nuclei of M. truncatula roots (Sieberer et al., 2009). This nuclear-localized cameleon is a convenient tool for monitoring Ca2+ signalling in atrichoblasts, since these epidermal cells lack the cytoplasm-rich apical region characteristic of growing root hairs.

We report, in this article, that sustained nuclear Ca2+ spiking can be detected in hyphopodium-contacted epidermal cells of both M. truncatula and carrot roots, thus demonstrating that fungal-activated Ca2+ spiking occurs in the cell type targeted by AM hyphopodia, and is most probably part of an ancient plant signalling pathway predating the divergence between the rosid and asteroid clades. In addition, we have been able to show that exudates of germinated AM spores (but not purified rhizobial NFs) are able to trigger nuclear Ca2+ spiking in the outer root tissues of ROCs, and that this response is limited to the AM-responsive root zone. Ca2+ spiking responses to both hyphopodia formation and AM exudate application are dependent on genes of the common SYM pathway in M. truncatula, but independent of the NFP gene encoding the lysine motif-receptor-like kinase (LysM RLK) which mediates NF perception (Arrighi et al., 2006). Together, these findings provide additional evidence that Ca2+ spiking is a key component of a highly conserved AM-activated signalling pathway required for intracellular fungal infection, and furthermore suggest that cameleon-expressing ROCs can provide valuable AM-specific bio-tests for the future characterization of fungal symbiotic signals.

Materials and Methods

Plant and fungal materials

Medicago truncatula genotype Jemalong A17 and the mutants nfp-2 (Arrighi et al., 2006), dmi1-1, dmi2-2 and dmi3-1 (Catoira et al., 2000; Wais et al., 2000) were used in this study, as well as the horticultural Daucus carota var. sativus. Agrobacterium rhizogenes-transformed ROCs expressing the 35S:NupYC2.1 construct (Sieberer et al., 2009) were obtained according to Boisson-Dernier et al. (2001) for M. truncatula and to Bécard & Fortin (1988) for carrot. Transformed roots with a high level of nuclear fluorescence were selected 21 days after inoculation, decontaminated and subcultured on M medium at 25°C in the dark for subsequent use as ROCs. The AM fungi used in this study were Gigaspora margarita isolate BEG 34 (International Bank for the Glomeromycota, University of Kent, UK) and G. gigantea isolate HC/F E30 (Herbarium Cryptogamicum Fungi, University of Torino, Italy). The latter is characterized by a strong cytoplasmic autofluorescence (Séjalon-Delmas et al., 1998).

For AM-targeting experiments, spores were pregerminated and placed in Petri dishes with fresh root cultures covered with Biofolie 25™ (Dutscher SAS, Brumath, France), as described in Genre et al. (2005). Medicago truncatula ROCs were grown in vertically oriented Petri dishes to favour the development of a regular fishbone-shaped root system (Chabaud et al., 2002). The nongravitropic carrot ROCs were initially grown horizontally and then switched to vertical growth following fungal inoculation in order to facilitate hyphal targeting of young lateral roots. The dishes were visually screened to detect highly ramifying hyphae and epidermal cells contacted by newly formed hyphopodia.

Production of the fungal exudate and root treatment

Batches of 100 surface-sterilized G. margarita spores were placed in 1 ml of sterile distilled water and incubated for 7 d at 30°C in the dark to induce germination (germination rate > 90%). The fungal germination medium was recovered by pipetting, concentrated 10-fold using a Lio5P lyophilizer (Cinquepascal, Milan, Italy) and then stored at −20°C.

Segments of primary roots carrying one or two young laterals (2–4 cm in length) were placed in a 2-mm-thick microchamber on a microscope slide containing sterile distilled water. After placing the root segments in the microchamber, they remained responsive to exudate for several hours. The water in the microchamber was rapidly (< 30 s) substituted by 100 μl of exudate before initiating confocal image acquisition. To prevent a cold-shock response, all solutions were warmed to 25°C before use.

For NF treatment, aqueous solutions at a final concentration of 10−8 M were prepared from a concentrated stock (10−3 M) of purified Sinorhizobium meliloti NF (kindly provided by Allan Downie, JIC, Norwich, UK), and root explants were treated as above. The activity of the NF stock was confirmed by eliciting epidermal-specific pMtENOD11-GUS reporter gene activity in seedlings of the transgenic M. truncatula line L416 (Charron et al., 2004).

Confocal microscopy and measurement of changes in nuclear Ca2+ levels

Confocal microscopy was used for all the fluorescence resonance energy transfer (FRET)-based cameleon experiments described in this work, thereby avoiding the background root fluorescence which would have been encountered using epifluorescence microscopy and which could have masked potential Ca2+ responses in atrichoblasts. FRET-based ratio imaging for the detection and plotting of relative changes in nuclear Ca2+ levels corresponding to yellow fluorescent protein (YFP) to cyan fluorescent protein (CFP) fluorescence intensity changes over time (Miyawaki et al., 1997, 1999) was performed according to Sieberer et al. (2009). The pinhole was set to 4–5 Airy Units in order to increase the thickness of the optical sectioning and thereby limit the likelihood of the nuclei moving out of focus during confocal scanning. A scanning resolution of 512 × 512 pixels was chosen to allow rapid imaging, and frames were collected every 5 s. Bright-field images were acquired simultaneously using the transmission detector of the microscope. For studies of hyphopodia-associated Ca2+ spiking, a minimum of 20 contacted epidermal cells from at least three independent roots (between 50% and 75% responding) was examined for each plant species and each mutant line, with the exception of M. truncatula nfp-2 roots for which only six cells were observed. For exudate treatments, a minimum of 10 reactive root epidermal cells (approximately 75% responding) was observed for each treatment and, in all cases, the ratio imaging presented in Figs 4 and 5 was initiated immediately following exudate addition.

Statistical analysis

The width of 50 peaks for either hyphopodium-contacted (seven nuclei from six independent roots) or exudate-treated (12 nuclei from four independent roots) epidermal cells of M. truncatula roots was measured manually. As the data for the exudate-elicited responses do not fit a normal distribution (Kolmogorov–Smirnov test, < 0.05), the respective median values were compared using the Kruskal–Wallis test.


Hyphopodium development triggers nuclear Ca2+ spiking in M. truncatula and D. carota root epidermal cells

By using the in vitro-targeted inoculation system developed by Chabaud et al. (2002), we were able to visually monitor the progressive growth of G. gigantea hyphae until contact with the host root and the subsequent development of hyphopodia on the epidermal surface of M. truncatula ROCs expressing the 35S:NupYC2.1 construct (Sieberer et al., 2009). Fungal contact sites were selected for FRET-based confocal microscopy (Fig. 1a,b), and a study of over 100 atrichoblast nuclei revealed significant nuclear Ca2+ spiking in almost one-half of these nonroot hair epidermal cells contacted by hyphopodia. The irregularity of the oscillatory signal is well illustrated in the 10-min-long recordings taken from three adjacent contacted epidermal cells shown in Fig. 1c–e. Fig. 2 shows that c. 45% of hyphopodium-contacted cells spiked between 3 and 10 times over the 10-min observation period. By comparison, the large majority of atrichoblasts from control noninoculated roots showed no spiking (e.g. Fig. 1k), and the remaining 30% a maximum of only one or two isolated spikes (Fig. 2). It is striking that the highest oscillation frequencies of 7–10 peaks per 10 min were only observed in cells for which the nucleus had moved to a position facing the site of hyphopodium contact (Fig. 2). This transcellular migration corresponds to the first visible response of the epidermal cell nucleus to the formation of the hyphopodium (Genre et al., 2005). Although only 15% of these cells showed high-frequency spiking during the 10-min observation period, statistical analysis confirms that this difference is significant with respect to cells for which the nucleus had not migrated.

Figure 1.

 Nuclear Ca2+ spiking in root epidermal cells contacted by arbuscular mycorrhizal (AM) hyphopodia. Top view of the epidermis of Medicago truncatula (a, b) and Daucus carota (f, g) roots contacted by Gigaspora hyphae. The bright-field images (a, f) show hyphopodia adhering to the root surface. The fungal outline is indicated by a dotted line in the corresponding pseudo-coloured fluorescence images (b, g) that show the NupYC2.1-labelled nuclei of the underlying epidermal cells. Circles identify the nuclei of hyphopodium-contacted cells (atrichoblasts) whose corresponding fluorescence resonance energy transfer (FRET) plots [yellow fluorescent protein (YFP) to cyan fluorescent protein (CFP) ratio] are presented in (c–e) and (h, i), and show repeated Ca2+ spiking over the 10-min acquisition period. (j) Nuclear spiking of a contacted epidermal cell of M. truncatula recorded continuously over 30 min. (k) Absence of Ca2+ spiking in an epidermal cell of a noninoculated M. truncatula root. Bars, 20 μm.

Figure 2.

 Distribution of Ca2+ spiking responses in nuclei of hyphopodium-contacted and neighbouring epidermal cells. The levels of Ca2+ spiking responses, represented by the total number of peaks recorded over a 10-min observation period, are presented for the following four categories of epidermal cells: atrichoblasts from control noninoculated roots (white bars, = 42); atrichoblasts not in direct contact with the arbuscular mycorrhizal (AM) fungus, but in the immediate vicinity (one to two cells distant) of hyphopodia (hatched bars, = 52); atrichoblasts directly contacted by hyphopodia but without nuclear positioning facing the site of fungal attachment (grey bars, = 47); and hyphopodia-contacted atrichoblasts with the nucleus facing the contact site (black bars, = 88). High-frequency spiking nuclei (7–10 spikes per 10 min) were only observed in cells for which the nucleus had migrated towards the site of hyphopodium contact. Statistical analysis shows that the two populations of cells represented by grey and black bars are significantly different (χ² = 10.26, < 0.01) with regard to the high-frequency spiking category of nuclei.

In addition to those atrichoblasts directly contacted by AM hyphopodia, we also observed that a small percentage (∼20%) of the surrounding epidermal cells (one to two cells distant from ramifying hyphae and hyphopodia) showed significant Ca2+ spiking with between three and six spikes over the 10-min period (Fig. 2). This suggests that diffusible fungal factors could be responsible for short-distance Ca2+ spiking responses in the epidermis, in line with the results of Kosuta et al. (2008). Finally, sustained nuclear Ca2+ spiking could be followed in certain hyphopodium-contacted epidermal cells for up to several hours (30 min continuous spiking is illustrated in Fig. 1j). In contrast with the variability in the spike periodicity, the width of the spikes in response to pre-infection AM contact was relatively homogeneous (27 ± 1.6 s). This spike duration lies within the 15–40 s range previously observed for NF-elicited nuclear spiking in root hairs (Sieberer et al., 2009).

Since AM fungi can associate with c. 80% of angiosperms, and since endosymbiosis-associated Ca2+ spiking has never been studied outside of the legume family, we decided to extend this study to the nonlegume host D. carota. We therefore introduced the NupYC2.1 cameleon into ROCs of D. carota and performed similar Gigaspora-targeting experiments as for Medicago. Fig. 1f–i shows that nuclear Ca2+ spiking is also triggered in carrot atrichoblasts in contact with AM hyphopodia, with a similar spiking pattern to that observed in M. truncatula. We therefore conclude that nuclear Ca2+ spiking is a prepenetration response to AM hyphopodium formation on the host root epidermis and is shared by both legume and nonlegume hosts.

The common SYM genes DMI1 and DMI2 are required for hyphopodium-elicited Ca2+ spiking in M. truncatula

In order to determine whether hyphopodium-induced Ca2+ spiking is dependent on the common SYM signal transduction pathway that is also implicated in rhizobial NF signal transduction (Parniske, 2008), we performed parallel experiments using NupYC2.1-expressing M. truncatula root cultures carrying mutations in the DMI1, DMI2 and DMI3 genes. Fig. 3 shows that nuclear Ca2+ spiking is absent in dmi1-1 and dmi2-2 mutant roots, whereas the dmi3-1 mutant displays pronounced spiking in hyphopodium-contacted cells. This indicates that the hyphopodium-dependent signals which trigger nuclear Ca2+ oscillations are transduced via the common SYM pathway, and is consistent with the downstream role of the nuclear-localized CCaMK as the decoder of Ca2+ spiking (Oldroyd & Downie, 2006). Finally, experiments performed with a mutant line (nfp-2) defective for NF perception in M. truncatula showed that the activation of hyphopodium-induced epidermal Ca2+ spiking is not dependent on the LysM RLK which mediates NF recognition (Arrighi et al., 2006) (Fig. 3d). This result is consistent with the fact that nfp mutants are not defective in AM colonization (Ben Amor et al., 2003).

Figure 3.

Medicago truncatula dmi1/2 mutants are defective in arbuscular mycorrhizal (AM) fungal-induced nuclear spiking. The Ca2+ spiking response induced by hyphopodium contact is absent in both dmi1-1 (a) and dmi2-2 (b) mutant lines. By contrast, hyphopodium-triggered Ca2+ spiking is normal for dmi3-1 (c) and nfp-2 (d) mutants.

AM fungal exudate elicits DMI1/2-dependent nuclear Ca2+ spiking in the root epidermis

Based on the results of Kosuta et al. (2008) and our observation that Ca2+ responses are not limited to hyphopodium-contacted atrichoblasts, we then asked whether exudates of germinated AM spores (G. margarita), known to contain bioactive molecules (Navazio et al., 2007; Gutjahr et al., 2009), could also induce Ca2+ spiking in the M. truncatula root epidermis. We found that the application of a 10× concentrated fungal exudate induced pronounced nuclear Ca2+ spiking in the majority (c. 75%, = 100) of the atrichoblasts located between 10 and 20 mm from the root tip (Figs 4b, 5a). This region corresponds to the developing and mature root hair zone, which is the primary target for AM root infection (Chabaud et al., 2002; Genre et al., 2005, 2008). Root hairs within this zone also responded to the spore exudate (Fig. 4c), although with poorly-defined and more irregular peaks. Water controls failed to elicit any Ca2+ spiking (Fig. 4d), and nonconcentrated (1×) fungal exudate only triggered low-frequency spiking (Fig. 4e). Finally, only very limited spiking was observed in the distal zone of the root (Fig. 4a), and no Ca2+ spiking was detected in the root elongation zone (Fig. 4f).

Figure 4.

 Epidermal Ca2+ spiking responses to fungal exudate vary according to position and cell type along the Medicago truncatula root. (a) Treatment with 10× concentrated fungal exudate elicited only limited nuclear Ca2+ oscillations within atrichoblasts located 40 mm from the root tip. (b) Atrichoblasts located 10–20 mm from the tip showed the strongest Ca2+ spiking responses to 10× fungal exudate. (c) Root hairs within the 10–20-mm zone responded to 10× fungal exudate with broader and more irregular peaks. (d) No spiking was observed following treatment with distilled H2O. (e) Spiking was less pronounced in atrichoblasts following 1× exudate application. (f) No response was observed to the fungal exudate within the root elongation zone.

Figure 5.

 Nuclear Ca2+ spiking in the Medicago truncatula and Daucus carota root epidermis treated with arbuscular mycorrhizal (AM) fungal exudate. (a–c) 10× fungal exudate induced nuclear Ca2+ spiking in atrichoblasts located in the responsive zone (10–20 mm from the tip) of M. truncatula (a) and carrot (b, c) root organ cultures. (d) No Ca2+ spiking was elicited in response to 10−8 M Sinorhizobium meliloti Nod factor treatment of M. truncatula roots. (e–h) Although fungal exudate failed to induce nuclear Ca2+ spiking in dmi1-1 (e) and dmi2-2 (f) mutants, Ca2+ responses were unaltered in dmi3-1 (g) and nfp-2 (h) mutants.

Similar results were obtained for the NupYC2.1-expressing carrot ROCs (Fig. 5b,c), further showing that the capacity of the host plant to respond to AM fungi via intracellular Ca2+ signalling is probably widespread among host plants. For both Medicago and carrot roots, Ca2+ spiking was generally initiated within the first 10 min following fungal exudate application and lasted for at least 30–40 min. As with the Ca2+ spiking responses to fungal hyphopodia, both plants showed a wide variability in the overall spiking pattern, with periodicities for M. truncatula ranging from two to five spikes per 10 min. Interestingly, the spike width median of 50 s (Fig. 6) was almost double that of the width median of the oscillations induced in response to hyphopodium contact. Finally, it should be underlined that treatment with purified S. meliloti NFs (10−8 M) failed to activate spiking in M. truncatula ROCs (20 cells analysed from four independent roots; Fig. 5d), in line with the fact that these roots cannot be nodulated by rhizobia (Akashi et al., 2003).

Figure 6.

 Variability of spike width for exudate-induced Ca2+ spiking compared with that associated with hyphopodium-activated Ca2+ spiking. Spike widths were measured manually for 50 peaks in each case (black bars for hyphopodium-elicited spiking, and grey bars for exudate-induced spiking). In the case of the exudate-induced spikes, the data could not be modelled by a normal distribution (Kolmogorvo–Smirnov test, < 0.05). The spike width medians (25 s for hyphopodium-elicited spikes vs 52 s for exudate-induced spikes) are statistically significantly different according to the Kruskal–Wallis test (< 0.05).

It is also possible to visualize the nuclear Ca2+ spiking variability elicited by germinated AM spore exudates by concomitant time-lapse imaging of the YFP to CFP ratio for a number of adjacent outer root cells (Supporting Information Video S1). Interestingly, both epidermal and outer cortical cell nuclei were included within the selected focal plane, thus revealing that low-frequency spiking is also elicited in the outer cortex in response to the fungal exudate. This time-lapse illustrates the cell autonomy of the Ca2+ spiking responses in adjacent cells, a feature previously described for NF-induced cytoplasmic and nuclear spiking in M. truncatula root hairs (Miwa et al., 2006; Sieberer et al., 2009).

Finally, we confirmed that Ca2+ spiking elicited by the fungal exudate is blocked in the AM-responsive zones of dmi1-1 (Fig. 5e) and dmi2-2 (Fig. 5f) roots, and unaltered in epidermal cells of dmi3-1 and nfp-2 mutants (Fig. 5g,h, respectively), as shown earlier for hyphopodium-induced Ca2+ spiking. In conclusion, bioactive molecules present in AM fungal exudates trigger robust Ca2+ spiking in root epidermal cells of both legume and nonlegume ROCs. In M. truncatula, this response is dependent on the activation of the common SYM pathway, and is most pronounced in atrichoblasts located in the root zone susceptible to AM infection.


Ca2+ spiking is part of an ancient signalling pathway required for AM infection

In vivo imaging of ROCs expressing the nuclear-specific NupYC2.1 cameleon construct has revealed that nuclear-localized Ca2+ spiking is induced in nonroot hair epidermal cells (atrichoblasts) in response to contact with AM fungal hyphopodia in both M. truncatula and D. carota roots. For M. truncatula, hyphopodium-elicited Ca2+ spiking is dependent on the DMI1 and DMI2 genes, but not on DMI3, consistent with the positioning of Ca2+ spiking downstream of DMI1 and DMI2 and the potential decoding of this secondary messenger by the nuclear CCaMK (Lévy et al., 2004). Since the activation of the common SYM pathway is required for the organization of the PPA (Genre et al., 2005), we hypothesize that the nuclear Ca2+ oscillatory response is a prerequisite for the initiation of this specific cellular remodelling which is essential for subsequent intracellular colonization by the AM fungus. In this context, it should be underlined that high-frequency nuclear spiking (7–10 peaks per 10 min) was only observed in atrichoblasts in which the nucleus had moved towards the hyphopodium contact site. This initial nuclear movement is the first observable cellular response to fungal contact and is independent of the DMI1/2 genes (Genre et al., 2005). Therefore, the physical proximity of the host nucleus to the adhered fungus may be a requisite for the establishment of the high-frequency Ca2+ spiking signature necessary for PPA formation. Since only 15% of cells with the nucleus positioned facing the hyphopodium show high-frequency spiking, we deduce that this particular signature is likely to be present during only a limited period following nuclear migration. Finally, the fact that Kosuta et al. (2008) were unable to observe Ca2+ spiking in root hairs several days after hyphae had contacted the epidermis presumably indicates that the fungal signal responsible for eliciting spiking is no longer being produced locally once the fungus has initiated infection. Indeed, our time lapse observations reveal that infection can be very rapid (within 24 h) following hyphopodium contact.

Notwithstanding the fact that common SYM genes (and in particular the gene encoding CCaMK) are highly conserved amongst all mycorrhizal plant species including bryophytes (Zhu et al., 2006; Markmann et al., 2008; Chen et al., 2009; Wang et al., 2010) and have been shown to be essential for the establishment of AM symbioses in monocots such as rice (Chen et al., 2007; Banba et al., 2008), this is the first time that a Ca2+ spiking response to symbiotic stimuli has been observed in a nonlegume. The fact that very similar nuclear Ca2+ spiking profiles can be detected in the carrot root epidermis in contact with AM hyphopodia argues that this response is a conserved element of AM fungal signal transduction predating the divergence between asterids and rosids, both belonging to the angiosperm eudicot clade (Wang et al., 2009). Since Ca2+ spiking is a key feature of rhizobial NF signal transduction in both temperate and tropical legumes (Capoen et al., 2010), it is therefore likely that the oscillatory intracellular Ca2+ signal is a hallmark of plant root endosymbioses across the plant kingdom.

Ca2+ spiking responses can be elicited by AM fungal exudates

Our results show that sustained nuclear Ca2+ spiking is not restricted to hyphopodium-contacted cells, but can also be induced in a small percentage of neighbouring cells which are not in direct contact with the fungus, albeit at lower spiking frequencies. This implies that Ca2+ spiking can be activated either by diffusible fungal factors and/or host cell-to-cell communication. The role of diffusible fungal factors would certainly be consistent with the observations of Kosuta et al. (2008) that cytosolic Ca2+ spiking responses can be elicited in noncontacted M. truncatula root hairs in the vicinity of ramifying AM hyphae. Furthermore, we have been able to show that nuclear Ca2+ spiking can also be activated by germinated spore exudates in both Medicago and carrot epidermal cells from the root zone susceptible to AM infection, thus indicating that AM signal molecules capable of activating the common SYM pathway are already secreted during AM spore germination in the absence of the host plant. Navazio et al. (2007) exploited similar fungal exudates to those used in this study to elicit a rapid Ca2+ transient in aequorin-expressing soybean cells. However it is difficult to compare these results with those obtained using cameleon-transformed roots since differentiated root epidermal cells are likely to respond differently to the cotyledon-derived soybean cell suspensions.

Previous studies in M. truncatula have provided evidence for the existence of diffusible AM signals based on a variety of host responses including the activation of gene expression (Kosuta et al., 2003; Kuhn et al., 2010), the stimulation of lateral root branching (Olah et al., 2005) and the activation of starch-related metabolic pathways (Gutjahr et al., 2009). Certain of these responses appear to be SYM pathway-dependent (Olah et al., 2005; Gutjahr et al., 2009; Kuhn et al., 2010), whereas others are not (Kosuta et al., 2003; Kuhn et al., 2010), thus suggesting that AM fungi generate a number of distinct biologically active molecules with different cellular targets and corresponding signalling pathways. We have found that Ca2+ spiking elicited by the fungal exudate in the responsive zone of the root is generally broader and more variable in width when compared with the relatively homogeneous spikes elicited by hyphopodium contact. This difference indicates that the bio-active molecules released by the presymbiotic mycelium may differ from the fungal signals generated following hyphopodium differentiation on the outer root surface.

During their co-evolution with symbiotic microbes, legumes may have needed to modulate Ca2+ spiking responses in order to discriminate between rhizobial and AM interactions. It is therefore intriguing that both cytoplasmic (Kosuta et al., 2008) and nuclear (this paper) pre-infection Ca2+ spiking responses elicited in M. truncatula epidermal tissues in response to AM fungi appear to be significantly less regular than those elicited by rhizobial NFs in root hairs (Kosuta et al., 2008; Sieberer et al., 2009). However, it remains to be shown whether such differences in the pattern of Ca2+ spiking represent key determinants in transducing the respective endosymbiotic signals into appropriate cellular responses. Finally, the observation that NFs are unable to induce Ca2+ spiking in ROCs is of particular interest, since this implies that it is now possible to distinguish between the activation of the two endosymbiotic SYM-dependent signal transduction pathways. We therefore anticipate that cameleon-expressing ROCs will provide useful AM-specific bio-assays for the future identification and characterization of fungal signals perceived through the common SYM pathway.


We thank David Douds for kindly providing an inoculum of G. gigantea and Allan Downie for supplying the S. meliloti NF. This research was funded by an Italian Converging Technologies project funded by Regione Piemonte (ref. CIPE-BIOBIT), an Italian project (ref. PRIN 2008), a French National Research Agency project (ref. ANR-08-BLAN-0029-01) and a joint project financed by the French National Centre for Scientific Research (ref. PICS 4267). Confocal microscopy carried out at the Laboratory of Plant–Microbe Interactions made use of the Microscopy-Imagery Platform facilities belonging to the Federated Research Institute (IFR40), Pole of Plant Biotechnology, Toulouse, France. Confocal microscopy in Torino was performed at the Laboratory of Advanced Microscopy, Department of Plant Biology, Torino, Italy.