Here we describe use of a mitochondrial targeted Cameleon to produce stably transformed Arabidopsis plants that enable analyses of mitochondrial Ca2+ dynamics in planta and allow monitoring of the intra-mitochondrial Ca2+ concentration in response to physiological or environmental stimuli. Transgenic plants co-expressing nuclear and mitochondrial targeted Cameleons were also generated and analyzed. Here we show that mitochondrial Ca2+ accumulation is strictly related to the intensity of the cytoplasmic Ca2+ increase, demonstrating a tight association between mitochondrial and cytoplasmic Ca2+ dynamics. However, under all experimental conditions, mitochondrial Ca2+ dynamics were substantially different from those monitored in the cytoplasm, demonstrating that mitochondria do not passively sense cytosolic Ca2+, but actively modulate the intra-mitochondrial level of the cation. In particular, our analyses show that the kinetics of Ca2+ release from mitochondria are much slower than in the cytoplasm and nucleus. The mechanisms and functional implications of these differences are discussed.
Mitochondria are key organelles that are involved in many aspects of cell functions from plants to humans, ranging from cell metabolism to stress responses and programmed cell death regulation (McAinsh and Pittman, 2009; Contreras et al., 2010).
Despite the pronounced metabolic connection between mitochondria and other cellular compartments in plants, little is known regarding the signals that mediate these inter-communications or coordination of the activities by signaling molecules. Many factors have been suggested to be involved in retrograde signaling control, including reactive oxygen species, cellular carbohydrate status, mitochondrial electron transfer chain reduction state (Yoshida et al., 2010) and calcium (Ca2+) (Subbaiah et al., 1998).
As far as Ca2+ is concerned, it is one of the most versatile second messengers in all eukaryotic organisms. In plants, Ca2+ is involved in nearly all aspects of development and participates in many regulatory processes (Dodd et al., 2010; Kudla et al., 2010). Ca2+ signals take the form of transient increases in the cytosolic free Ca2+ concentration ([Ca2+]c) arising in the cytosol as a result of fluxes from the external medium and subcellular compartments (Hetherington and Brownlee, 2004; Dodd et al., 2010). Cytoplasmic Ca2+ signal specificity is guaranteed by a decoding mechanism based on the amplitude and/or frequency of the Ca2+ transients: series of repetitive Ca2+ oscillations are commonly referred to as ‘Ca2+ signatures’ (McAinsh and Pittman, 2009).
In plants, in vivo analysis of the intra-mitochondrial Ca2+ concentration, [Ca2+]m, is still in its infancy (Logan and Knight, 2003), mainly due to technical difficulties in measuring changes in the mitochondrial free Ca2+ concentration. For a long time, it was only possible to measure [Ca2+]m in isolated mitochondria loaded with Ca2+-specific indicators (Zottini and Zannoni, 1993). The introduction of genetically encoded Ca2+ indicators (e.g. aequorin and GFP-based Ca2+ probes), and their detection either by photon emission measurements or by fluorescence microscopy, have revolutionized the detection and visualization of intracellular Ca2+ dynamics in living plants. The first tools developed for in vivo intracellular analyses of Ca2+ dynamics in plants were based on use of aequorin targeted to the cytoplasm (Knight et al., 1991), nucleus (van Der Luit et al., 1999), chloroplasts (Johnson et al., 1995), tonoplast (Knight et al., 1996) and mitochondria (Logan and Knight, 2003). In particular, in vivo analyses of mitochondrial Ca2+ dynamics showed that several stress stimuli, including osmotic stress, induced [Ca2+]m increases with dynamics that were different from those in the cytoplasm (Logan and Knight, 2003). Experiments performed using aequorin produce very reliable data, but reflect the response of a population of cells or plants, and intercellular heterogeneity could not be investigated.
In order to improve cellular and subcellular resolution, use of other Ca2+ sensors has been pursued in recent years, and, among the genetically encoded Ca2+ indicators, the most frequently used in plants are Cameleons. Cameleons are fluorescence resonance energy transfer (FRET)-based indicators in which two GFP variants, CFP and YFP (or circularly permuted variants of YFP) are linked together by the Ca2+-binding protein calmodulin and the calmodulin-binding peptide M13. Binding of Ca2+ to the Ca2+-responsive elements alters the efficiency of FRET, allowing a quantitative measurement of Ca2+ dynamics. In recent years, Cameleon variants have been developed (Palmer and Tsien, 2006; Palmer et al., 2011) with improved fluorescence, greater changes in FRET upon Ca2+ binding, and a broad range of Ca2+ affinities. Cameleons have been widely used to monitor [Ca2+] in the cytoplasm of plant cells (Allen et al., 1999; Swanson et al., 2011). They have also been targeted to the cytoplasm/nucleus (Allen et al., 1999; Monshausen et al., 2008; Yang et al., 2008; Krebs et al., 2012), nucleus (Sieberer et al., 2009; Krebs et al., 2012), endoplasmic reticulum (Iwano et al., 2009), plasma membrane (Krebs et al., 2012) and peroxisomes (Costa et al., 2010).
Here we report data on in vivo analyses of Arabidopsis transgenic plants harboring the genetically encoded Ca2+ probe Cameleon YC3.6 specifically targeted to the mitochondrial matrix. By use of external stimuli, it was possible to monitor mitochondrial Ca2+ dynamics in guard and root tip cells. Moreover, by crossing plants expressing Cameleons targeted to the nucleus or mitochondria, we were able to monitor Ca2+ dynamics simultaneously in these two subcellular compartments in vivo.
Results and Discussion
Construction of the YC3.6 Cameleon targeted to mitochondria
In order to analyze Ca2+ dynamics in mitochondria, two main issues must be addressed: (i) choice of the most appropriate Ca2+ sensor, and (ii) precise targeting of the sensor to the mitochondrial matrix.
In order to target YC3.6 to plant mitochondria matrix, we fused the mitochondrial targeting sequence from subunit VIII of human cytochrome c oxidase to the N-terminal end of the probe. This sequence has been shown to be highly efficient in targeting other Cameleons to mammalian mitochondria, especially when repeated four times (4mt) (Filippin et al., 2005; Palmer et al., 2006). To ensure ubiquitous expression of the probe, YC3.6 was placed under the control of a single CaMV 35S promoter (Figure S1), and the entire cassette was inserted into the pGreen 0179 binary vector. This vector confers hygromycin resistance to transgenic plants, but the Cameleon cassette can be easily transferred to other pGreen backbones (e.g. pGreen 0029 or pGreen 0229) (Hellens et al., 2000) by a one-step sub-cloning procedure (see Experimental Procedures and Figure S2) in order to change the antibiotic used for the plant selection. The construct generated was renamed 4mt-YC3.6 (Figure S1).
Generation of Arabidopsis transgenic plants expressing mitochondria-targeted YC3.6
The construct 4mt-YC3.6 was introduced into Arabidopsis plants (Columbia ecotype) by the floral-dip method (Clough and Bent, 1998), and eight independent transgenic lines were obtained. In order to verify ubiquitous expression of the probe, we performed confocal laser scanning microscopy (CLSM) analyses in 7-day-old Arabidopsis seedlings (Figure 1). Figure 1(a–f) shows expression of the 4mt-YC3.6 probe in leaves, where fluorescence was clearly detected in epidermal, stomata (Figure 1a–c) and mesophyll cells (Figure 1d–f). The signal was also present in cells of the hypocotyl (Figure 1g–i), mature root cells (Figure 1j,k), including root hairs and in all the cell types present in the root tip region: meristematic cells, transition and elongation zones (Figure 1 l-m) ). We did not observed silencing effects when using this construct.
In order to confirm the correct mitochondrial localization of 4mt-YC3.6, we performed co-localization experiments by staining 4mt-YC3.6-expressing root tissues (Figure 2a) with the tetramethylrhodamine methyl ester dye (TMRM), a potentiometric cationic probe (Figure 2b) (Zottini et al., 2006). Fluorescence of 4mt-YC3.6 perfectly merged with the TMRM signal of stained cells (Figure 2c). In order to determine whether the presence of YC3.6 in the mitochondria could have undesirable effects on them, we analyzed independently, by means of CLSM, mitochondria of transgenic plants (Figure 2d,e) and mitochondria of wild-type plants stained with TMRM (Figure 2f,g). Mitochondria morphology and distribution appeared comparable in the analyzed tissues of the two lines. Moreover, time-lapse analyses, performed in leaf epidermal cells of 4mt-YC3.6 transgenic (Movie S1) and wild-type (Movie S2) lines, showed typical plant mitochondria motility (Logan, 2010).
Osmotic stress induces Ca2+ transients in the cytoplasm, nucleus and mitochondria of Arabidopsis leaf cells
In order to evaluate the functionality of 4mt-YC3.6 localized to mitochondria, we performed a series of experiments in guard cells, which are often used as model system for single-cell analyses of Ca2+ dynamics (Allen et al., 1999; Hetherington and Brownlee, 2004; Costa et al., 2010; Kim et al., 2010). We specifically assayed guard cells undergoing osmotic stress, a condition known to induce substantial and reproducible cytoplasmic Ca2+ transients in plant cells (Knight et al., 1998; Logan and Knight, 2003). In parallel, we monitored Ca2+ dynamics in the cytoplasm and nucleus. For this second group of experiments, we used Arabidopsis transgenic plants expressing YC3.6 targeted to the cytoplasm (NES-YC3.6) or the nucleus (NLS-YC3.6) (Krebs et al., 2012).
First we tested the effect of 250 mm sorbitol (as osmoticum) on pre-opened stomata (see Data S1) in leaf epidermal strips (Costa et al., 2010) of 4–5-week-old 4mt-YC3.6 Arabidopsis plants (Figure 3) using a simple protocol. Guard cells bathed in standard solution (see Data S1) were first challenged with 250 mm sorbitol and then returned to the standard normo-osmotic medium. Both addition of sorbitol (hyper-osmotic stress) and removal of sorbitol (hypo-osmotic stress) caused an increase in the 540/480 nm fluorescence emission ratio ΔR/R0Δ(R/R0. ratio at time “tn”-ratio at time “t0”/ratio at time “t0”), which is proportional to the [Ca2+]. Pseudocolor ratio images of a typical cell (taken at the times indicated in Figure 3b) are shown in Figure 3(a). The kinetics of the ΔR/R0 changes for the same cell are shown in Figure 3(b). A rapid [Ca2+]m increase was induced by hyper-osmotic stress, followed by a slow [Ca2+]m decrease toward the initial basal level. Similar behavior was observed after sorbitol removal (hypo-osmotic stress) (Figure 3a,b), but the mitochondria Ca2+ peak was larger (see also Table 1). The ΔR/R0 for the second peak was 0.48 ± 0.16, a value that falls in the probe linear range response in mitochondria (ΔRmax/R0 0.73 ± 0.14), as determined in experiments performed in permeabilized leaf epidermal cells (see Experimental Procedures). Figure 3(c) shows that, after application of hypo- and hyper-osmotic stresses, the fluorescence emissions at 540 nm (cpVenus) and 480 nm (CFP) mostly showed anti-parallel behavior (Figure 3c). In some cases (e.g. Figure 3c), a movement artifact occurred, resulting in a parallel change in the fluorescence intensity at both wavelengths. However, this artifact is corrected by the ratiometric values, which show no change. Other possible artifacts, for example due to focus changes associated with stomata morphological modification after release of the osmotic stress (Figure 3a), are similarly corrected by the ratiometric calculation (Rudolf et al., 2004, and data not shown).
Table 1. Values of ΔR/R0 (means ± SD) corresponding to the first and second Ca2+ peaks measured in the nucleus, cytoplasm and mitochondria after addition and removal of 250 mm of sorbitol
T/2 corresponds to the time at which [Ca2+] was halved in the various compartments.
2.73 ± 0.82
8.64 ± 2.34
2.74 ± 0.55
13.5 ± 4.74
1.86 ± 0.28
5.91 ± 3.02
2.05 ± 0.29
14.5 ± 4.97
0.2 ± 0.1
31.08 ± 10.84
0.48 ± 0.16
199.22 ± 34.74
The same osmotic stress assayed in 4mt-YC3.6 plants was then repeated using Arabidopsis plants expressing the YC3.6 probe in the cytoplasm and nucleus (Figure 4a,d). Both [Ca2+]c and the nuclear Ca2+ concentration ([Ca2+]n) showed qualitatively a very similar behavior (Figure 4b,e): a steep [Ca2+] increase was evoked by addition of sorbitol, followed by a fast decrease with complete recovery of the basal [Ca2+] in <20 sec (Table 1). In the cytoplasm, the first peak observed after hyper-osmotic stress was less intense than that induced by the hypo-osmotic one (as in mitochondria), but the amplitudes of the two peaks were not statistically different in the nucleus (Table 1). The absolute amplitude of the peak in the cytosol was smaller than that in the nucleus. However, the kinetics of the [Ca2+] decreases were not significantly different in the two compartments. Thus, in guard cells subjected to osmotic stress, the responses in terms of the ΔR/R0 for cytoplasm and nuclear dynamics are similar, but not identical. The reasons for the larger peaks in the nucleus compared to the cytoplasm are presently unknown, and may depend either on a real difference in the [Ca2+] of the two compartments or different behavior of the probe in the two environments. In kinetic terms, however, it can be concluded that, at least for the osmotic stress stimuli investigated here, the nuclear Ca2+ dynamics mirror the cytosolic dynamics. This result is consistent with what is now generally agreed for various mammalian cell types (Rizzuto and Pozzan, 2006; Giacomello et al., 2010), but is partially in contrast with previous data in plants (Mazars et al., 2011, and references therein). The osmotic stress-induced Ca2+ increases measured here depend on Ca2+ influx from the extracellular environment, as the presence of 500 μm EGTA in the bath solution prevented any change in [Ca2+] in the various compartments (data not shown).
Two major differences were noted between the increases in ΔR/R0 in mitochondria and those in the cytoplasm and nucleoplasm: (i) the amplitude of the ΔR/R0 in the mitochondria is much smaller, and (ii) while accumulation of Ca2+ in the mitochondria is similar to that in the other two compartments, the return to basal level is much slower in mitochondria (Figures 3 and 4, and Table 1). With regard to the difference in amplitudes, it should be stressed that the dynamic range of the YC3.6 within mitochondria (ΔRmax/R0 0.73 ± 0.14) is much smaller than that in the nucleus (ΔRmax/R0 3.31 ± 0.46), as assessed in permeabilized leaf epidermal cells (see Experimental Procedures). With regard to the prolonged Ca2+ retention in mitochondria, the simplest explanation is that the Ca2+ efflux mechanism of the organelles in plants is intrinsically slow, and the Ca2+ released from mitochondria is either rapidly sequestered by other compartments such as the endoplasmic reticulum and/or extruded into the external medium without leading to a detectable cytoplasmic Ca2+ increase.
Finally, we observed that only half of the cells responded to the addition of sorbitol by a Ca2+ increase in all compartments, but a Ca2+ increase was observed in all cells analyzed after removal of sorbitol. One explanation could be that the turgor reached by guard cells during stomata pre-opening differed between cells, and the most turgid ones did not respond to hyper-osmotic stress. However, removal of sorbitol was always perceived, probably through activation of channels, such as stretch-activated channels, that mediate Ca2+ entrance after being specifically stimulated by hypotonic conditions, as reported for Vicia faba (Zhang et al., 2007).
The experiments here reported were performed in guard cells, but a similar behavior was observed in epidermal cells in response to osmotic treatment (particularly sorbitol removal), with an increase in Ca2+ concentration in the cytoplasm, nucleoplasm and mitochondria. In all cells of the imaged leaf (Movie S3), epidermal cells expressing NES-YC3.6 responded with a series of cytoplasmic Ca2+ oscillations. In epidermal cells, the dynamics of [Ca2+]m were very similar to what was observed in guard cells, with two rapid Ca2+ increases after addition and removal of sorbitol, followed by a slow decrease in [Ca2+]m (Movie S4). However, no significant Ca2+ oscillations were observed in mitochondria.
Plant cells co-expressing nuclear and mitochondrial targeted Cameleons allow simultaneous analyses of Ca2+ dynamics in two compartments of guard cells
The measurements reported in Figure 3 and 4 were performed in parallel in cells from different transgenic plants, thus preventing direct comparison between the kinetics of mitochondrial and cytoplasmic Ca2+ signals in the same cell. Crossing plants expressing a cytosolic Cameleon and a mitochondrial Cameleon cannot resolve this issue, as the two signals cannot be spatially resolved. However, experiments performed in guard cells exposed to osmotic stresses demonstrate that the Ca2+ transients in nucleoplasm and cytoplasm are kinetically similar, suggesting that the nuclear Ca2+ signature mirrors what occurs in the cytoplasm. The mitochondrial and nuclear signals are spatially segregated and can be independently monitored in the same cell (Giacomello et al., 2010). We thus crossed Arabidopsis plants expressing the Cameleon in the nucleus (NLS-YC3.6) with plants expressing the Cameleon in mitochondria (4mt-YC3.6). In order to determine whether the two probes were simultaneously expressed in different tissues/organs, we performed CLSM analyses in 7-day-old Arabidopsis seedlings (Figure S3). As observed with parental plants, the expression of YC3.6 present in both, nuclei and mitochondria, was observed in epidermal cells, stomata (Figure S3a–c) and mesophyll cells (Figure S3d–f) of young leaves. The two probes were also co-expressed in other tissues/organs, such as hypocotyl (Figure S3g–i), root and the root tip (Figure S3j–m).
To monitor the Ca2+ dynamics in the two compartments, we subjected leaf epidermal strips of plants, expressing Cameleons in both compartments, to the osmotic stress protocol described in Figures 3 and 4. Pre-opened stomata were perfused with 250 mm of sorbitol, and then the osmoticum was removed (Figure 5a,b). The osmotic stress induced both nuclear and mitochondria Ca2+ transients, whose dynamics mimicked those observed in the respective single transgenic lines (Figures 3b and 4e). Only half of the analyzed cells showed Ca2+ transients in response to hyper-osmotic stress, whereas all cells responded to hypo-osmotic stress. Moreover, a mitochondrial Ca2+ transient was observed only if there was a corresponding Ca2+ transient at the level of the nucleoplasm/cytoplasm: in no cells did mitochondria accumulate Ca2+ in the absence of a Ca2+ increase in the nucleoplasm/cytoplasm (Figure 5b, after hyper-osmotic stress). Application of osmotic stress in the presence of the Ca2+ chelator EGTA (500 μm) in the bath solution completely abolished both nucleoplasm/cytoplasm and mitochondrial Ca2+ transients (data not shown).
Mitochondria Ca2+ transient analyses in Arabidopsis root tips evoked by eATP
In order to study mitochondrial Ca2+ handling in a more physiological context, we analyzed the response in the root tip, an organ that integrates various environmental stimuli, often mediated by Ca2+, to direct root growth (Fasano et al., 2002). Extracellular ATP (eATP) has recently been recognized as a key signaling molecule in plant roots (Roux and Steinebrunner, 2007; Tanaka et al., 2010a) that is able to inhibit growth and cause root curling in Arabidopsis seedlings (Tang et al., 2003). In particular, Tanaka et al. (2010b) showed that Arabidopsis root cells responded to eATP with a large and rapid cytoplasmic Ca2+ increase. Recent data strongly suggest that the eATP response is mediated by ATP receptors located in the plasma membrane of root cells (Demidchik et al., 2009), although their molecular identity is still unknown.
We studied the reciprocal interaction in terms of Ca2+ handling between the cytoplasm, nucleus and mitochondria in response to eATP, performing the experiments in parallel using independent plants expressing YC3.6 targeted to the various compartments. We then monitored the dynamics of [Ca2+] in Arabidopsis roots in various regions of the root tip (from the meristematic cells to the cells of the elongation zone) (Figure 6, blue, red and green traces). Administration of eATP led to a rapid [Ca2+] increase in both cytoplasm and nucleoplasm (Figure 6a,c), with the external cells being the first to respond to the stimulus (Figure 6b,d). Seedlings expressing YC3.6 in the cytoplasm and nucleus showed Ca2+ transients that spread throughout the entire tip region (Figure 6b,d and Movies S5 and S6), reaching a maximum ΔR/R0 after 31.25 ± 4.78 and 38 ± 8.3 sec, respectively. Analysis of the various tip regions (meristematic and transition/elongation zones in Figure 6) showed that the Ca2+ transients differed in terms of magnitude, but not shape, within the cytoplasm and nucleoplasm (Figure 6a,c). The typical Ca2+ signature comprised a first Ca2+ peak followed by a series of smaller peaks, with complete recovery of the basal [Ca2+] in <5 min (Figure 6a,c). Notably, the Ca2+ signals were indistinguishable between nucleus and cytoplasm in the various root regions in terms of both amplitude and kinetics. The lack of clear recognition of second and third Ca2+ peaks (and higher orders) in the very tip region (blue in Figure 6) is due to the resolution power of the epifluorescence microscope. We then performed the same analyses in the root tip of Arabidopsis 4mt-YC3.6 plants (Figure 6e,f). The first cells that responded to eATP were those in the meristematic region (blue), and then a Ca2+ wave propagated to the entire tip region (Figure 6f and Movie S7). However, the dynamics of mitochondrial Ca2+ accumulation were quite different from those of the other two compartments: mitochondria showed only one main peak with no significant oscillations, occurring in the meristematic/transition zone 128.3 ± 8.2 sec after eATP treatment (Figure 6e, blue and red traces), and in the upper part of the tip 234.75 ± 64 sec after eATP treatment (Figure 6e, green trace). After reaching a maximum, [Ca2+]m started to decrease, approaching the basal level after 20 min; in the meristematic region, [Ca2+]m remained high for longer durations. These experiments indicate that mitochondrial Ca2+ accumulation in roots in response to eATP shows different dynamics in different regions that, appear to depend on the magnitudes of cytoplasmic/nuclear Ca2+ increase (Figure 6a,c). The fastest mitochondrial Ca2+ accumulation was observed in the meristematic zone (Figure 6e, blue trace), i.e. the same region that showed the highest cytoplasmic/nuclear Ca2+ increase. To confirm this observation, we performed an eATP dose–response analysis. We subjected Arabidopsis root tip seedlings to three eATP concentrations (2, 0.5 and 0.01 mm), and analyzed the response, in terms of FRET changes, of cytoplasm and mitochondria in the tip region (Figure 7). An increase in eATP concentration led to a clear dose-dependent response in terms of the maximum [Ca2+]c (Figure 7a,c) and [Ca2+]m (Figure 7b,c) reached by the cells. The experiments demonstrate that the larger the cytoplasm [Ca2+]c response, the larger the mitochondrial response, an observation that is consistent with the presence of a low-affinity, highly regulated, Ca2+ transport system in the membrane of plant mitochondria. The mechanism responsible for mitochondrial Ca2+ uptake in plants could be the so-called ‘Ca2+ uniporter’, a low-affinity Ca2+ channel recently identified at the molecular level in mammalian cells (Baughman et al., 2011; De Stefani et al., 2011), for which six predicted isoforms are present in Arabidopsis (Stael et al., 2011).
Using aequorin, Logan and Knight (2003) showed that the maximum Ca2+ concentration in the cytoplasm and mitochondria was reached almost simultaneously for all tested stimuli. Indeed, differences in the kinetics of [Ca2]m changes among cell types in response to eATP were clearly revealed by our analysis. These differences may depend on the fact that the Cameleon allowed monitoring of specific regions in the root, whereas the response monitored by aequorin reflected the cell population of the seedling.
Mitochondria Ca2+ analyses in single cells of plant roots
As mentioned above, use of Cameleons allows imaging of changes in Ca2+ of single cells, or even single organelles. In the case of guard cells, for example, this can easily be achieved using a wide-field fluorescence microscope, but a confocal microscope is necessary to study single cells in a more complex organ such as the root tip (Tanaka et al., 2010b). We therefore used CLSM to monitor the Ca2+ responses in the root tip of 4mt-YC3.6-expressing plants, measuring single-cell responses to eATP administration (Figure 8a and Movie S8). As shown in Figure 8(a), the resolution of the CLSM was sufficient to allow monitoring of mitochondrial Ca2+ accumulation in single cells of the root transition zone (indicated by red and green arrows in Figure 8a). Figure 8(b) shows that the two cells analyzed, from two different root cell layers, responded with very similar Ca2+ kinetics. It is interesting to note that the single-cell mitochondrial Ca2+ accumulation kinetics are similar to those observed in the entire root tip (blue trace in Figure 8a,b) or using the wide-field fluorescence microscope (Figures 6e and 7b). These results demonstrate that use of the 4mt-YC3.6 Cameleon itself provides sufficient resolution to observe a single cell in a single focal plane even in a complex organ, such as the root tip. Interestingly, also with the CLSM, we observed what we call a ‘mitochondrial Ca2+ wave’ (Movie S8), but, compared with wide-field fluorescence microscopy analyses, the ‘wave’ was more uniform (Movie S8). The drawbacks of using CLSM for such analyses were the high laser power required for imaging cells in the deeper root tissues and the time required for each acquisition, which meant that, at this time, we are unable to obtain better resolution or perform very long experiments. The development of new high-resolution and less-invasive imaging systems will overcome these limitations (Fischer et al., 2011), and potentially offer a means to monitor mitochondrial Ca2+ handling during long time-lapse experiments, such as during development. Moreover, in the future, it will be interesting to compare the responses of plant expressing mitochondria targeted Cameleon or aequorin in the same organs/tissues. This will be feasible by using the new generation of ultrasensitive Electron Multiplying CCD (EMCCD) camera for aequorin–GFP detection.
We have generated Arabidopsis transgenic plants harboring a genetically encoded Ca2+ probe, Cameleon YC3.6, that is specifically targeted to the mitochondria. We monitored mitochondrial Ca2+ dynamics in various tissues and cell types in response to two stimuli (osmotic stress and eATP treatment). Generation of plant cells co-expressing nuclear and mitochondrial targeted Cameleons allowed comparison of mitochondrial [Ca2+] kinetics with cytosolic/nuclear signals. We demonstrated that mitochondrial Ca2+ accumulation in response to osmotic stress is strictly dependent on [Ca2+]c increase, and that, even for the eATP response, mitochondrial Ca2+ accumulation strictly depends on the cytoplasmic Ca2+ increase. Finally, use of CLSM and the mitochondrial targeted Cameleon allowed monitoring of dynamic changes in Ca2+ within the organelles of single root cells. Plants expressing Cameleons in both the nucleus and the mitochondria represent a useful tool for better understanding of the in vivo impact of mitochondrial Ca2+ handling on the cytoplasmic Ca2+ signature and specific downstream cellular responses. Additionally, simultaneous analyses of Ca2+ in two compartments may shed light on the mechanism and role of mitochondria retrograde signaling.
Plant material and growth conditions
All Arabidopsis thaliana plants used in this study were of the Columbia ecotype. Plants for guard cell imaging were grown in Jiffy pots (http://www.jiffypot.com/) under 16 h light (70 μmol m−2 sec−1)/8 h dark at 22°C and 75% relative humidity. Seeds of Arabidopsis were surface-sterilized by vapor-phase sterilization (Clough and Bent, 1998), and plated on half-strength MS medium (Murashige and Skoog, 1962) ((Duchefa, http://www.duchefa.com/) supplemented with 0.1% sucrose, 0.05% MES, pH 6.0, and solidified using 0.8% plant agar (Duchefa). After stratification at 4°C in the dark for 3 days, the seeds were transferred to the growth chamber under 16 h light (70 μmol m−2 sec−1)/8 h dark at 24°C. The plates were kept vertically. The seedlings used for the analyses were 11–12 days old, with a mean root length of 4.5 cm.
Transgenic NES-YC3.6 Arabidopsis plants, in which the Cameleon is targeted to the cytoplasm by a nuclear export signal (NES) and NLS-YC3.6 Arabidopsis plants, in which the Cameleon is targeted to the nucleus by a nuclear localization signal (NLS) were kindly provided by Dr Karin Schumacher (Department of Developmental Biology, University of Heidelberg, Germany) (Krebs et al., 2012).
The YC3.6 coding sequence was digested from the pcDNA3-YC3.6 vector (Nagai et al., 2004) using HindIII and EcoRI restriction enzymes, and ligated into the p35S-2 vector (http://www.pgreen.ac.uk/JIT/JIT_fr.htm). In order to generate the 4mt-YC3.6 construct, the 4mt targeting peptide was isolated by digestion of pcDNA3-4mt-D1cpv (Zampese et al., 2011) with HindIII, and ligated into the 35S-YC3.6. The clone obtained was sequenced to verify the correct orientation of the targeting peptide, and then the entire cassette (35S-4mt-YC3.6-Ter) was PCR-amplified using Phusion® DNA polymerase (Finnzymes, http://www.finnzymes.fi/). For 35S-4mt-YC3.6-Ter cassette amplification, we used the primers 5′-CATGGGTACCGATATCGTACCCCTACTCCAAAAAT-3′ (forward) and 5′-CATGGGTACCGATATCGATCTGGATTTTAGTA-3′ (reverse), in which KpnI restriction sites (underlined) are present at the 5′ and 3′ ends, respectively. The amplicon for the entire expression cassette was digested using KpnI, and ligated into the pGreen 0179 binary vector (Hellens et al., 2000). The binary vector was then introduced in the Agrobacterium tumefaciens GV3101 strain.
The Agrobacterium strains obtained were used to generate transgenic Arabidopsis plants by the floral-dip method (Clough and Bent, 1998). For each construct, independent transgenic lines were selected, and two lines were used for imaging experiments.
Confocal microscopy analyses
Confocal microscopy analyses were performed using a Leica SP5 laser scanning confocal imaging system (http://www.leica-microsystems.com). For cpVenus fluorescence, excitation was at 514 nm and emission was between 525/540 nm. For chlorophyll detection, excitation was at 514 nm and detection at >600 nm. For TMRM analysis, the seedlings were stained for 10 min in 5 mm KCl, 10 mm MES, 10 mm Ca2+, pH 5.8, supplemented with 500 nm TMRM. Seedlings were washed for 5 minutes with the same solution and analyzed by means of a confocal microscope. Excitation was at 543 nm and emission was between 590–620 nm. For the co-localization analyses of cpVenus and TMRM, sequential excitation in the confocal microscope scanning configuration was used. Image analyses were performed using ImageJ software (http://rsb.info.nih.gov/ij/).
For Ca2+ imaging analyses, the roots were imaged using a 40× lens (HCX PL APO CS 40×/1.25–0.75 oil), and YC3.6 was excited using the 458 nm line of the argon laser with 15% total power. The CFP and cpVenus emissions were collected at 473–505 and 526–536 nm, respectively, and the pinhole diameter was 2 airy units. Images were collected every 7.5 sec. False-color ratio images were obtained using the ImageJ ‘Ratio Plus Plugin’ (Palmer and Tsien, 2006).
Guard cell and root tip imaging
The imaging techniques used are described in Data S1.
Measurement of the YC3.6 dynamic range in mitochondria and nuclei
In order to determine the in vivo dynamic range of the probe (ΔRmax/R0) within mitochondria and the nucleus, leaf epidermal cells were permeabilized by treating them for 4 min with 0.2 mm digitonin in an intracellular-like medium containing 100 mm potassium gluconate, 1 mm MgCl2, 10 mm HEPES, pH 7.5, and 5 mm EGTA. The digitonin was then removed, and the cells were held for 5 min in the same medium containing 5 mm EGTA, and finally washed and maintained in 1 mm EGTA. In order to measure the ΔRmax/R0, 10 mm Ca2+ was added to the medium. The ratio measurements were performed by observing single non-moving mitochondria (n = 9) or nuclei (n = 13). The ratios reported are means ± SD.
All the data are representative of at least nine cells or roots. Reported traces correspond to the typical observed responses.
We thank Karin Schumacher (Department of Developmental Biology, University of Heidelberg, Germany) for providing the NES-YC3.6 and NLS-YC3.6 Arabidopsis plants. We tank Paulo Magalhães for help with ImageJ analyses. This work was supported by a grant from the Ministero dell’Istruzione, dell’Università e della Ricerca, fondi PRIN to F.L.S., and grants from the Ministero dell'Istruzione, dell'Universitá e della Ricerca (Futuro in Ricerca Bando 2010) to A.C., and from the Veneto Region (Biotech 2) to T.P.