Mastoparan activates calcium spiking analogous to Nod factor-induced responses in Medicago truncatula root hair cells.

Mas7 induces calcium spiking similar to Nod factor responses The rhizobial derived signalling molecule Nod factor is essential for the establishment of the Medicago truncatula / Sinorhizobium meliloti symbiosis. Nod factor perception and signal transduction in the plant involves calcium spiking and leads to the induction of nodulation gene expression. It has previously been shown that the heterotrimeric G-protein agonist mastoparan can activate nodulation gene expression in a manner analogous to Nod factor activation of these genes and this requires DMI3 , a calcium and calmodulin dependent protein kinase (CCaMK) that is required for Nod factor signalling. Here we show that Mastoparan activates oscillations in cytosolic calcium, similar but not identical to Nod factor induced calcium spiking. Mastoparan induced calcium changes occur throughout the cell, whereas Nod factor induced changes are restricted to the region associated with the nucleus. Mastoparan induced calcium spiking occurs in plants mutated in the receptor-like kinases NFP and DMI2 and in the putative cation channel DMI1 , that are all required for Nod factor induction of calcium spiking, indicating either that Mastoparan functions downstream of these components or that it uses an alternative mechanism to Nod factor for activation of calcium spiking. However, both Mastoparan and Nod factor induced calcium spiking are inhibited by cyclopiazonic acid and n-butanol, suggesting some common mechanisms underpinning these two calcium agonists. The fact that Mastoparan and Nod factor both activate calcium spiking and can induce nodulation gene expression in a DMI3 dependent manner strongly implicates CCaMK in the perception and transduction of the calcium signal.


INTRODUCTION
Calcium is a common secondary messenger that functions in a diverse array of signalling pathways. The maintenance of specificity for such a ubiquitous signal is likely to be a feature of the calcium signature, which is defined by both spatial and temporal components of the calcium response (Sanders et al., 2002). One of the most complex calcium signatures is repetitive oscillations in calcium and these have been identified in a number of plant and animal signalling cascades. In plants, oscillatory calcium has been seen in guard cells in response to ABA, cold and external calcium (McAinsh et al., 1995), in growing pollen tubes (Iwano et al., 2004) and in legume root hair cells in response to the rhizobial signalling molecule Nod factor (Ehrhardt et al., 1996). In guard cells the frequency and number of oscillations dictates the long term closure of the stomate (Allen et al., 2001) and this is consistent with calcium oscillation frequency dictating the level and spectrum of gene induction in animal cell lines (Dolmetsch et al., 1997).
indicates lysosome like vesicles as a target of NAADP action (Churchill et al., 2002). The vacuole acts as a major internal store for calcium in plant cells (Sanders et al., 2002) and both IP 3 and cADPR calcium-mobilising activity has been shown to exist on the tonoplast membrane (Allen et al., 1995). However, there is also evidence for IP 3 , cADPR and NAADP activatable channels on plant ER membranes (Martinec et al., 2000;Navazio et al., 2000;Navazio et al., 2001). cADPR, IP 3 and NAADP can act in concert or individually and the diverse combinatorial nature of these three messengers partly explains the diverse calcium signatures that are produced.
Proliferation of cytosolic IP 3 is a function of phospholipase C (PLC) that degrades plasma membrane phosphatidylinositol 4,5-bisphosphate to generate IP 3 and diacylglycerol. PLC is activated by heterotrimeric G-proteins that in turn are induced by G-protein coupled receptors (Singer et al., 1996;Assmann, 2002). In this way ligand perception at the plasma membrane can be linked to calcium changes in the cytosol by the mobilisation of IP 3 . In animal systems G-protein agonists such as the wasp venom peptide Mastoparan can activate PLC independent of ligand perception via activation of the heterotrimeric G-proteins and thus induce calcium changes (Ross and Higashijima, 1994;Sukumar et al., 1997). Mastoparan has been shown to activate a number of responses in plant cells including the mobilisation of calcium (Tucker and Boss, 1996;Pingret et al., 1998). However, recent evidence indicates that Mastoparan effects in plants can occur independently of the heterotrimeric G-proteins (Miles et al., 2004) and to date the target for Mastoparan in plant cells has not been defined.
Mastoparan has previously been shown to activate nodulation gene expression in a manner analogous to Nod factor activation of these genes (Pingret et al., 1998). This induction requires the activity of DMI3 (Charron et al., 2004) a calcium and calmodulin dependent protein kinase (CCaMK; (Levy et al., 2004;Mitra et al., 2004), but does not require DMI1 or DMI2, two components of the Nod factor signalling pathway upstream of calcium spiking. This activity of Mastoparan was taken to indicate the presence of heterotrimeric G-proteins in the Nod factor signalling pathway (Pingret et al., 1998).
Here we show that the synthetic analog of Mastoparan, Mas7, activates oscillations in calcium similar to Nod factor induced calcium spiking. Analogous to the nodulation gene expression studies Mas7 activation of calcium spiking does not require NFP, DMI1 or DMI2 that are necessary for Nod factor activation of calcium spiking. The fact that CCaMK is required to transduce the signal from two independent activators of calcium oscillations, strongly implicates this protein in decoding the oscillatory calcium signal.

RESULTS
The Mastoparan synthetic analog Mas7 activates calcium oscillations in M.
truncatula root hair cells.
The requirement for DMI3/CCaMK in Mastoparan and Mas7 (Mastoparan synthetic analog) activation of the Nod factor reporter ENOD11- GUS (Charron et al., 2004) suggested that Mastoparan and Mas7 were able to activate calcium responses in root hair cells of Medicago truncatula. It has previously been shown that the concentration of Mas7 (0.5μM) required for induction of ENOD11-GUS (Fig. S1) induces lethality in approximately 50% of M. truncatula root hair cells (Charron et al., 2004). We therefore chose to assess calcium responses using an M. truncatula line stably transformed with the calcium reporter cameleon YC2.1 (Miwa et al., 2006), which allows a large number of cells to be assayed. Cameleon also has the advantage of allowing ratiometric measurements of calcium changes within individual cells (Miyawaki et al., 1997;Miyawaki et al., 1999). Similar to previous reports we also saw a high degree of lethality of root hair cells treated with 0.5 μM Mas7 as indicated by loss of cytoplasmic streaming.
However, among the cells that survived this treatment (indicated by maintenance of cytoplasmic streaming) a significant proportion, 27%, showed calcium oscillations ( Fig.   1; Table 1). The inactive analog Mas17 that contains a single amino acid change in the peptide showed no calcium oscillations (Table 1). Treatment with 0.2 μM Mas7 activated calcium oscillations, but in only 13.5% of root hair cells analysed, while treatment with 2 μM Mas7 was predominantly lethal. The calcium oscillations induced by Mas7 appear broadly similar to Nod factor induced calcium spiking, but the period between oscillations is longer ( Fig. 1; Fig. 2a) and the lag from treatment to induction of oscillations is also longer ( Fig. 2b) with Mas7 induced calcium oscillations. However, for both period and lag Mas7 induced oscillations are much more variable than Nod factor induced calcium spiking. The variability in period is mostly the result of variability between cells with Mas7 induced calcium oscillations showing a much broader range of periods (Fig. 2c). Indeed Mas7 induced calcium oscillations can show a much longer period, 10-20 minutes, than has ever been observed in Nod factor responses. These cells showing very long periods between oscillations still maintain a continuity of period ( Fig.   1), indicating a mechanism for rhythmicity even over these long periods. We have shown that Nod factor induced calcium spiking is not restricted to those cells that activate ENOD11 (Miwa et al., 2006), and in an analogous manner we observed Mas7 induced calcium oscillations in both growing and mature root hair cells.

The structure of the Mas7 induced calcium oscillations
Nod factor induced calcium spiking shows a very rapid increase in calcium followed by a more gradual decline (Fig. 3;(Ehrhardt et al., 1996). This can be interpreted to indicate the opening of calcium channels on internal stores and the rapid movement of calcium down its concentration gradient, followed by the slower active uptake of calcium into calcium stores. This uptake would depend on calcium ATPases and this is consistent with pharmacological studies (Engstrom et al., 2002). A careful comparison between Nod factor induced calcium spiking and Mas7 induced calcium oscillations indicates subtle differences between these two responses. The overall time from initiation of the calcium increase to a return to basal calcium levels is equivalent between Nod factor and Mas7 induced responses (Fig. 3). However, Mas7 shows a slower initial increase in calcium levels such that the first phase of the spike represents 34% of the overall spike time for Mas7 compared with 17% in Nod factor induced spiking. The initial phase of the spike will be dependent on the activation dynamics of a population of calcium channels. We interpret the Mas7 spike structure to indicate that either the dynamics of calcium channel opening differ at the individual channel level compared with Nod factor activation of these channels or that Mas7 is less effective at activating the population of channels at an equivalent time. In order to ensure that the differences we observed in Mas7 induced calcium oscillations were not a function of unrelated actions of Mas7 we treated cells undergoing Nod factor induced calcium spiking with different concentrations of Mas7.
Treatment with Mas7 did not affect the structure of Nod factor induced calcium spiking, until the application of higher concentrations of Mas7 caused cell death (Fig. S2).

The spatial resolution of Mas7 induced calcium oscillations
Nod factor induced calcium spiking is mostly restricted to the cytosol associated with the nucleus (Fig. 4b), with minor changes occurring at the root hair tip (Shaw and Long, 2003). Because of the interference from the florescence of neighbouring cells, we can only assess the hair region of root hair cells in the cameleon transformed plants.
Therefore, in order to assess the spatial nature of the Mas7 induced calcium changes, we micro-injected root hair cells with the calcium responsive dye Oregon green dextran (10, 000 MW) and the unresponsive dye Texas red dextran (10,000 MW). Injection of the two dyes allows pseudo-ratio imaging of calcium changes (Shaw and Long, 2003), with Texas red providing a control for cytoplasmic content within the zone being analysed.
We found that in contrast to the nuclear restricted nature of Nod factor induced calcium spiking ( Fig. 4d- where Nod factor calcium changes are not observed (Fig. 4c). This suggests that Mas7 is able to activate calcium channels in a region of the cell that Nod factor cannot and so these data reveal some differences between Nod factor and Mas7 induced calcium oscillations.

Mas7 induction of calcium oscillations does not require NFP, DMI1 and DMI2
To assess the relationship between Nod factor and Mas7 induced calcium oscillations we analysed the ability of Mas7 to induce calcium responses in plants mutated in components of the Nod factor signalling pathway. Nod factor signalling is initiated by receptor-like kinases with sugar-binding motifs that are strong candidates for the Nod factor receptor and are represented by NFP in M. truncatula (Amor et al., 2003;Madsen et al., 2003;Radutoiu et al., 2003;Arrighi et al., 2006). Acting downstream of NFP is a second receptor-like kinase, DMI2 (Endre et al., 2002;Stracke et al., 2002) and a putative cation channel, DMI1 (Ane et al., 2004) and all three genes are required for Nod factor activation of calcium spiking (Wais et al., 2000). The calcium/calmodulin dependent protein kinase DMI3 (Levy et al., 2004;Mitra et al., 2004) is also required for Nod factor signal transduction but is not required for the activation of calcium spiking (Wais et al., 2000), indicating a function downstream of this calcium response. At the time of these experiments we did not have these mutant lines containing the cameleon reporter and therefore we chose to microinject root hair cells of nfp, dmi1-1, dmi2-1 and dmi3-1 with Oregon green dextran and Texas red dextran. We found that all mutant lines were able to initiate calcium oscillations following treatment with Mas7 ( Fig. 5; Table 2). The mutant alleles chosen are presumed to be null alleles since they contain premature stop codons (Endre et al., 2002;Ane et al., 2004;Levy et al., 2004;Mitra et al., 2004). This work indicates that unlike Nod factor Mas7 does not require NFP, DMI1 or DMI2 to activate calcium responses.

Inhibitors of Mas7 induced calcium oscillations
To further assess the relationship between Mas7 and Nod factor induced calcium oscillations we analyzed the effects of a number of inhibitors known to abolish Nod factor induced calcium spiking. Both cyclopiazonic acid (CPA), that inhibits type IIA calcium ATPases, and n-butanol, that inhibits phospholipase D (PLD), have been shown to inhibit Mas7 induction of ENOD11-GUS (Charron et al., 2004). Both of these inhibitors block Nod factor induced calcium spiking without obvious detrimental effects on cellular viability (Engstrom et al. 2002;Charron et al. 2004;H. W. and A. D. unpublished data). In order to assess the effect of these inhibitors on Mas7 responses we treated cells that showed robust and highly repetitive Mas7 induced calcium oscillations.
Three cells on three independent plants showed inhibition of Mas7 induced calcium oscillations following treatment with 0.5% n-butanol (Fig. 6b). Treatment with 10 μM CPA did cause inhibition of Mas7 induced calcium oscillations (Fig. 6c), but was less severe than has been observed with Nod factor induced calcium spiking (Fig. 6a). We  (Dolmetsch et al., 1998;Allen et al., 2001). This work has revealed that calcium oscillations can activate downstream responses and the period between oscillations dictates the nature of the downstream response. We have attempted to recapitulate this work in M. truncatula root hair cells (J. S. and G. O. unpublished data), but we have been unsuccessful in activating 40 calcium spikes, that have been shown to be required for nodulation gene expression (Miwa et al., 2006). Here we show that in addition to Nod factor we can activate calcium oscillations in M. truncatula root hair cells using the Gprotein agonist Mas7. This provides a valuable tool for assessing the relevance of calcium oscillations for the activation of nodulation responses.
Mas7 activates calcium oscillations broadly similar to Nod factor induced calcium spiking, but with a number of differences: a slower initial release of calcium in the Mas7 response; greater variability in the lag and period in Mas7 induced calcium oscillations and an expansion in the cellular location of Mas7 induced calcium changes. Despite the differences observed between Mas7 and Nod factor induced calcium spiking, both calcium agonists induce ENOD11 and ENOD12 expression with a similar spatial pattern to Nod factor activation of these genes (Pingret et al., 1998). Both Nod factor and Mas7 induction of ENOD11 is dependent on DMI3 (Charron et al., 2004), whose mutant phenotypes suggest a specific role in symbiosis signalling (Catoira et al., 2000). These data indicate that Mas7 induced calcium oscillations mimic Nod factor induced calcium spiking, and this highlights the importance of an oscillatory calcium signal for the induction of CCaMK and its activation of nodulation gene expression. In the absence of a forced calcium oscillatory system, the fact that two calcium agonists, Mas7 and Nod factor, induce calcium oscillations and both activate appropriate nodulation gene expression represents the best evidence we have to show the importance of calcium oscillations in this symbiosis signalling pathway.
The Mas7 induction of calcium oscillations and ENOD11 expression (Charron et al., 2004) do not require NFP, DMI1 or DMI2, M truncatula Nod factor signalling genes required for Nod factor induced calcium spiking (Wais et al., 2000). This can be interpreted in two ways: either Mas7 induced calcium oscillations are mechanistically unrelated to Nod factor induced calcium spiking; or Mas7 activates a component of the Nod factor signalling pathway downstream of NFP, DMI1 and DMI2. If we presume that Mastoparan and Nod factor induced calcium oscillations are mechanistically related, then the fact that Mastoparan can activate calcium oscillations in DMI1 mutants indicates that this putative cation channel is not the calcium channel responsible for Nod factor induced calcium spiking. Two inhibitors that are known to abolish Nod factor induced calcium spiking, the calcium ATPase inhibitor CPA and the PLD inhibitor n-butanol, inhibit Mas7 induced calcium oscillations, indicating common mechanisms between these two responses. However, this is not definitive proof that the calcium channel(s) that are involved in Nod factor and Mas7 induced calcium spiking are equivalent. Indeed, the differences observed between the structure and cellular location of these two calcium responses indicates that Mas7 and Nod factor act on different sets of calcium channel(s) or differentially activate equivalent calcium channel(s).
Mastoparan/Mas7 has been shown to activate G-proteins by catalysing GDP/GTP exchange similar to the action of G protein coupled receptors (Sukumar et al., 1997).
However, Mastoparan has also been shown to have additional sites of action in animal studies including inhibition of calcium ATPases (Longland et al., 1999) and modifying glycogen phosphorylase that modulates calcium release from the ryanodine receptor (Hirata et al., 2000;Hirata et al., 2003). As such Mastoparan can activate calcium release independent of G protein activation. This is consistent with observations in G protein mutants of Arabidopsis that show Mastoparan activation of mitogen-activated protein kinases (Miles et al., 2004) indicating a Mastoparan site of action independent of plant Gproteins. Mastoparan activates phospholipase C and D in legume roots (den Hartog et al., 2001) and based on pharmacological analysis it appears that PLD is required for Mas7 induction of ENOD11 (Charron et al., 2004). This indicates that Mastoparan leads to the activation of PLD that in turn induces calcium oscillations and such a model is supported by the inhibition of Mas7 induced calcium oscillations by n-butanol, a known inhibitor of plant PLD. However, it is also possible that the site of Mastoparan action requires the activity of PLD, for instance the protein that is modulated by Mastoparan may require phosphatidic acid for activity. Hence it is currently unclear whether Mastoparan modulates phospholipases directly or indirectly to generate a secondary signal that activates the calcium channel, or whether Mastoparan modulates other components of the calcium signalling machinery that require the activity of phospholipases. Defining the mode of action of Mastoparan could provide insights into mechanisms of the Nod factor signalling pathway and the induction of calcium spiking.

Plant material and growth conditions
M. truncatula seeds were scarified for 5 minutes in concentrated sulfuric acid, washed twice in sterile water, surface sterilized for 3 minutes in undiluted Clorox bleach, then washed 5 times in sterile water. Seeds were imbibed for 1-3 hours, transferred to Petri plates with damp filter papers and germinated overnight in the dark at room temperature.
Germinated seedlings were transferred to buffered nodulation media (BNM) (Ehrhardt et al., 1992) agar at pH 6.5 containing the ethylene inhibitor 0.1μM L-α-(2aminoethoxyvinyl)-glycine and grown overnight at 25 o C, with the roots shaded by wrapping the plates in aluminium foil. The root tips of the seedlings were removed to avoid disturbance during imaging and the seedlings were then transferred to a liquid BNM bath contained on a large coverslip. 0.5 μM Mas7 (Sigma), 0.5 μM Mas17 (Sigma) or 1nM Sinorhizobium meliloti Nod factor were added to the bath during imaging. For the calcium measurements using cameleon we used an M. truncatula R108 line that was stably transformed with cameleon YC2.1 regulated by the 35S promoter (Miwa et al., 2006). We used mutant lines that had at minimum been backcrossed once. Nod factor preparations were isolated as described by Ehrhardt et al. (1996).

Calcium imaging
For the analysis of mutants we microinjected root hair cells with calcium responsive dyes as described by Ehrhardt et al (1996), with slight modifications as described by Wais et al. (2000). Micropipettes were pulled from filamented capillaries on a pipette puller (model 773, Campden Instruments). These were loaded with Oregon green dextran (10,000 MW, Molecular Probes Inc) and Texas red dextran (10,000 MW, Molecular Probes Inc). Injections were performed using iontophoresis with currents generated from a cell ampliflier (model Intra 767, World Precision Instruments) and a stimulus generator made to our specifications (World Precision Instruments). Cells were analyzed on an inverted epiflourescence microscope (model TE2000, Nikon) using a monochrometer (model optoscan, Cairns Research) to generate specific wavelengths of light. During image capture the image was split using the Optosplit (Cairn Research), and each image passed through a filter for either Oregon green or Texas red emissions prior to exposure on the ccd chip (model ORCA-ER, Hamamatsu). The data was analyzed using Metaflor (Universal Imaging).
Lines expressing the calcium reporter cameleon YC2.1 were generated as described (Miwa et al., 2006). These lines were analyzed on the same inverted epiflourescent microscope (model TE2000, Nikon). During image capture the image was split using the Optosplit (Cairn Research), and each image passed through a filter for either CFP or YFP emissions prior to exposure on the ccd chip. The ratio of CFP:YFP was assayed using Metaflor (Universal Imaging).      following treatment with 0.5% (68 mM) n-butanol. (c). 10 μM CPA treatment inhibits 0.5 μM Mas7 induced calcium oscillations, but this inhibition is less severe than CPA inhibition of Nod factor induced calcium spiking (a). All traces are representative of three independent experiments. The y axis is the ratio of YFP:CFP in arbitrary units.