FRET-based genetically encoded sensors allow high-resolution live cell imaging of Ca2+ dynamics


  • Melanie Krebs,

    1. Department of Developmental Biology, Centre for Organismal Studies (COS), University of Heidelberg, Im Neuenheimer Feld 230, 69120 Heidelberg, Germany
    Search for more papers by this author
  • Katrin Held,

    1. Molecular Genetics and Cell Biology of Plants, Institute of Plant Biology and Biotechnology, Westfälische Wilhelms-Universität Münster, Schlossplatz 4, 48149 Münster, Germany
    Search for more papers by this author
    • Both authors contributed equally to this work.

  • Andreas Binder,

    1. Institute of Genetics, Biocenter University of Munich (LMU), Großhaderner Straße 4, 82152 Martinsried, Germany
    Search for more papers by this author
    • Both authors contributed equally to this work.

  • Kenji Hashimoto,

    1. Molecular Genetics and Cell Biology of Plants, Institute of Plant Biology and Biotechnology, Westfälische Wilhelms-Universität Münster, Schlossplatz 4, 48149 Münster, Germany
    Search for more papers by this author
  • Griet Den Herder,

    1. Institute of Genetics, Biocenter University of Munich (LMU), Großhaderner Straße 4, 82152 Martinsried, Germany
    Search for more papers by this author
    • Present address: Ablynx nv, Technologiepark 21, 9052 Ghent, Belgium.

  • Martin Parniske,

    1. Institute of Genetics, Biocenter University of Munich (LMU), Großhaderner Straße 4, 82152 Martinsried, Germany
    Search for more papers by this author
  • Jörg Kudla,

    Corresponding author
    1. Molecular Genetics and Cell Biology of Plants, Institute of Plant Biology and Biotechnology, Westfälische Wilhelms-Universität Münster, Schlossplatz 4, 48149 Münster, Germany
    Search for more papers by this author
  • Karin Schumacher

    Corresponding author
    1. Department of Developmental Biology, Centre for Organismal Studies (COS), University of Heidelberg, Im Neuenheimer Feld 230, 69120 Heidelberg, Germany
    Search for more papers by this author

(e-mail fax +49-6221-546404; e-mail


Temporally and spatially defined calcium signatures are integral parts of numerous signalling pathways. Monitoring calcium dynamics with high spatial and temporal resolution is therefore critically important to understand how this ubiquitous second messenger can control diverse cellular responses. Yellow cameleons (YCs) are fluorescence resonance energy transfer (FRET)-based genetically encoded Ca2+ -sensors that provide a powerful tool to monitor the spatio-temporal dynamics of Ca2+ fluxes. Here we present an advanced set of vectors and transgenic lines for live cell Ca2+ imaging in plants. Transgene silencing mediated by the cauliflower mosaic virus (CaMV) 35S promoter has severely limited the application of nanosensors for ions and metabolites and we have thus used the UBQ10 promoter from Arabidopsis and show here that this results in constitutive and stable expression of YCs in transgenic plants. To improve the spatial resolution, our vector repertoire includes versions of YCs that can be targeted to defined locations. Using this toolkit, we identified temporally distinct responses to external ATP at the plasma membrane, in the cytosol and in the nucleus of neighbouring root cells. Moreover analysis of Ca2+ dynamics in Lotus japonicus revealed distinct Nod factor induced Ca2+ spiking patterns in the nucleus and the cytosol. Consequently, the constructs and transgenic lines introduced here enable a detailed analysis of Ca2+ dynamics in different cellular compartments and in different plant species and will foster novel approaches to decipher the temporal and spatial characteristics of calcium signatures.


Due to their sessile life style, higher plants have to adapt rapidly to variable environmental conditions. Signalling networks that perceive, transmit and integrate information translate environmental cues into appropriate cellular programs. Changes in cytosolic free Ca2+ ([Ca2+]cyt), represent a fundamental concept in signalling in all eukaryotic cells. In plants, the second messenger Ca2+ is associated with various abiotic and biotic stimuli such as salt and osmotic stress, oxidative stress, wounding, low temperature, pathogen attack and nodulation (Dodd et al., 2010). Levels of cytosolic Ca2+ are tightly regulated by coordinated transport processes between the cytosol and the sites of Ca2+ storage (Berridge et al., 2000; Kudla et al., 2010). Ca2+ signals are characterized by a transient rise of cytosolic free Ca2+ but vary in their temporal and spatial qualities in a stimulus-dependent way and have therefore been designated as Ca2+ signatures (McAinsh et al., 1992; McAinsh and Hetherington, 1998). Spatial specificity of the Ca2+ signal results from the combination of distinct calcium releases from different intra- and extracellular storage compartments combined with the extraordinary slow cytoplasmic diffusion rate of Ca2+ -ions (Kudla et al., 2010). Such local and transient changes of calcium near the release sites at cellular membranes, termed calcium microdomains, are fundamental for cellular calcium signalling and regulation processes (Webb et al., 1996). Analyzing Ca2+ signatures and elucidating the contribution of different compartments as sources and reservoirs for Ca2+ signalling requires tools that allow monitoring of Ca2+ dynamics with high spatial and temporal resolution.

Different probes for in vivo measurements of Ca2+ dynamics are available that can be categorized into the small-molecule fluorescent indicators, such as Fura-2, and in the genetically encoded Ca2+ sensors (GECIs). For live cell Ca2+ imaging GECIs have several advantages over small-molecule fluorescent dyes such as the ability to fuse the indicator to a protein or tag of interest to monitor Ca2+ in individual subcellular locations and specific microdomains and to generate transgenic lines expressing the sensor within tissues that are not accessible for loading with fluorescent dyes. The GECIs can be further divided in chemiluminescent Ca2+ sensors, based on the protein aequorin and fluorescent Ca2+ indicator proteins (FCIPs) (Palmer, 2009; Contreras et al., 2010). The aequorin-based system is inherently different from the FCIPs, as the emitted light is generated by a chemical reaction that requires reconstitution of aequorin with a co-factor (McCombs and Palmer, 2008). The aequorin method is well established in plants and has contributed substantially to our current knowledge of Ca2+ mediated signalling processes (Knight et al., 1991, 1992, 1996; Logan and Knight, 2003; Kaplan et al., 2006; Tanaka et al., 2010). Since light emission by aequorin does not depend upon optical excitation, problems with chromophore bleaching and autofluorescence, observed when using FCIPs, do not occur (Plieth, 2001). However, the main inconvenience of the aequorin system is the low light emission, which severely limits the spatio-temporal resolution during imaging as Ca2+ signals have to be detected from whole seedlings or tissues (Plieth, 2001; Alonso et al., 2009). Mathematical simulation has shown that oscillatory single-cell Ca2+ signals cannot be resolved with aequorin, at least when using conventional light detection devices (Dodd et al., 2006). Therefore fluorescence microscopy is required if high temporal and spatial resolution Ca2+ imaging is to be achieved.

Yellow cameleons, developed in the laboratory of Roger Tsien more than 10 years ago (Miyawaki et al., 1997) are composed of a donor chromophore (CFP), calmodulin (CaM), a glycylglycine linker, the CaM-binding peptide of myosin light-chain kinase (M13), and an acceptor chromophore (YFP). Ca2+ binding to CaM initiates an intramolecular interaction between CaM and M13 which changes the protein from an extended to a more compact conformation resulting in an increased FRET-efficiency between CFP and YFP (Miyawaki et al., 1997).

In plants, YC2.1 was successfully used to study guard cell-specific Ca2+ dynamics (Allen et al., 1999, 2000, 2001). Although the constitutive and strong CaMV 35S promoter was used to ensure high level expression, in most transgenic lines YC2.1 expression was patchy and often restricted to symplastically isolated stomatal guard cells that are protected from global silencing mechanisms (Wille and Lucas, 1984; Himber et al., 2003). The 35S promoter has been associated with gene silencing and co-suppression events (Palauqui et al., 1996; Elmayan et al., 1998; Mishiba et al., 2005; Daxinger et al., 2008; Ülker et al., 2008). Accordingly, 35S controlled expression of genetically encoded glucose nanosensors is limited to early stages of development (Chaudhuri et al., 2011) or can only be achieved in silencing-deficient mutants of Arabidopsis (Deuschle et al., 2006).

We thus decided to equip our vectors with the UBIQUITIN10 (UBQ10) promoter from Arabidopsis (Norris et al., 1993) that has been demonstrated to confer moderate and stable in planta expression of target genes (Geldner et al., 2009; Grefen et al., 2010). Furthermore, our vectors include the yellow cameleons YC3.6 and D3cpv respectively. Both contain a circularly permutated Venus (cpVenus) that was demonstrated to be brighter and more efficient in maturation and acid stability compared to conventional enhanced yellow fluorescent protein (EYFP) used in YC2.1 (Miyawaki et al., 1999; Nagai et al., 2002, 2004; Palmer et al., 2006). YC3.6 as well as D3cpv have a suitable dissociation constant (Kd′) for measuring Ca2+ concentrations within the cytosolic range (Palmer and Tsien, 2006) and have been shown to be functional in transgenic plants (Monshausen et al., 2008; Costa et al., 2010; Tanaka et al., 2010). Our basic vectors are available with different plant selection markers to have maximum flexibility when creating transgenic lines carrying multiple resistance genes. To expand the spatial resolution of Ca2+ imaging, our vector set contains tagged versions of YCs that mediate specific targeting of the Ca2+ reporter to defined subcellular domains such as the cytoplasm, the nucleoplasm, the plasma membrane and the tonoplast. In addition, we provide cameleon constructs that are appropriate for N- and C-terminal fusions with additional proteins or tags of interest. Using this improved toolkit for analyzing Ca2+ dynamics we identified temporally and spatially specific Ca2+ signatures in root cells in response to several stimuli.

Results and Discussion

In order to facilitate the analysis of Ca2+ dynamics in different plant tissues and cell types we generated an extended set of binary vectors for in planta expression of the FRET-based YC Ca2+ sensors YC3.6 and D3cpv. A subset of pGPTVII-based (Walter et al., 2004) vectors (Figure 1 and Table 1) contains a cassette for expression of YC3.6 (Nagai et al., 2004) under control of an optimized Arabidopsis UBQ10 promoter (Grefen et al., 2010) and the terminator of the nos gene from Agrobacterium tumefaciens. For selection of transgenic plants, the pGPTVII-based vectors confer resistance to either BASTA (bar), kanamycin (nptII) or hygromycin B (hptII). The pGPTVII-based vectors are suitable for measuring Ca2+ dynamics in the cytosol and nucleus and should be chosen when attempting to combine cameleon constructs or lines with plants already carrying a selectable marker.

Figure 1.

 pGPTVII-based cameleon expression vectors.
Schematic representation of the pGPTVII-based cameleon expression vectors. The YC3.6 expression cassette is flanked by the UBIQUITIN10 (UBQ10) promoter and the terminator of nopaline synthase (nos). The vectors are available with different plant selection genes, bar, nptII and hptII, conferring resistance to BASTA, kanamycin and hygromycin B, respectively. Expression of the selection markers is controlled by the nos promoter and the terminator of gene7 (Ag7). Regulatory and coding elements are indicated by the colour code.

Table 1.   Overview of cameleon constructs
  Cameleon expression cassettePlant selection markerBacterial selection marker
Construct nameVector backbonePromoterN-tagCameleonC-tagTerminator
  1. Overview of cameleon expression vectors and their corresponding vector backbones. UBQ10, UBIQUITIN10; NES, nuclear export signal; NLS, nuclear localization signal; PM, plasma membrane; TP, tonoplast; nos, nopaline synthase; rbcS, ribulose bisphosphate carboxylase oxygenase small subunit; bar, BASTA resistance gene; nptII, neomycin phosphotransferase II; hptII, hygromycin phosphotransferase II; aadA, aminoglycoside-3-adenyltransferase.

NES-YC3.6pTKanUBQ10NESYC3.6 rbcSnptIIaadA
NLS-YC3.6pTKanUBQ10 YC3.6 rbcSnptIIaadA
TP-D3cpvpTKanUBQ10TPD3vpv rbcSnptIIaadA

In addition to the pGPTVII-derived vectors the collection also includes pTkan-based vectors equipped with the UBQ10 promoter and the rbcS terminator to control expression of yellow cameleon YC3.6 or D3cpv (Palmer et al., 2006) respectively (Figures S1 and 4; Table 1). The pTKan-derived subset of constructs is comprised of basic vectors carrying a multiple cloning site (MCS) for C- and N-terminal fusions with a protein or tag of interest (Figure S1; Table 1) as well as vectors containing tagged versions of YCs that are appropriate for measuring Ca2+ dynamics in specific subcellular domains. Experimental procedures describing the vector design are provided online (Appendix S1) and a detailed overview of all cameleon constructs, their specifications and properties is given in Table 1.

Stable and ubiquitous in planta expression of yellow cameleons using the UBQ10 promoter of Arabidopsis

We generated transgenic lines and analyzed the expression pattern of the UBQ10 promoter using confocal laser scanning microscopy (CLSM). In agreement with recent reports (Geldner et al., 2009), we observed uniform UBQ10 promoter activity in all types of cells and tissues and during all stages of development so far analyzed (Figure 2). In plants expressing YC3.6-Bar, we observed cytoplasmic and nucleoplasmic localization of YC3.6 (Figure 2a). Expression was detected during embryo development (Figure 2b), in leaf epidermal cells including guard cells (Figure 2c), in cells of the hypocotyl (Figure 2d), in roots and in root hair cells (Figure 2e). Another obvious advantage of the UBQ10 promoter is that it also drives expression of the reporter proteins in pollen and pollen tubes (Figure 2f), which is not accomplished when using the 35S promoter for gene expression. It is important to mention that UBQ10 mediated gene expression was observed to be stable and consistent at least until the T3 generation in all described plant lines so far examined.

Figure 2.

 The UBQ10 promoter of Arabidopsis is active throughout development.
CLSM analysis was used to follow UBQ10-driven expression of YC3.6 in Arabidopsis plants stably transformed with YC3.6-Bar. Fluorescence was monitored at different developmental stages using settings for cpVenus excitation and emission. UBQ10 promoter activity was found in hypocotyl cells (a, d), during embryo development (b), in leaf epidermal cells (c), in the root, including root hair cells (e) and in pollen (f). Scale bars represent 10 μm in (a) and 50 μm in (b–f).

Ubiquitously localized YC3.6 reports calcium transients in Arabidopsis roots and cotyledons

The functionality of the ubiquitously localized sensor was tested by performing live cell Ca2+ imaging using an epi-fluorescence microscope. To apply different external stimuli, the 5–6-day-old Arabidopsis seedlings stably transformed with YC3.6-Bar, were fixed in a custom-built perfusion chamber to allow exchange of liquid using a peristaltic pump. According to the protocol of Allen et al. (2000, 2001), we imposed [Ca2+]cyt transients by alternate application of depolarisation (100 mm KCl, 10 mm MES–Tris pH 6.15) and hyperpolarisation (0.1 mm KCl, 10 mm CaCl2, 10 mm MES–Tris pH 6.15) buffer (Figure 3). Hyperpolarisation-activated Ca2+ transients have been documented in guard cells (Grabov and Blatt, 1998; Allen et al., 2000; Hamilton et al., 2000), but have been also reported for root hair cells of Arabidopsis (Véry and Davies, 2000). Here, we measured hyperpolarisation-activated [Ca2+]cyt transients in cells of the root hair zone of young Arabidopsis seedlings (Figure 3a, b; Movie S1). Buffer exchange was performed approximately every 60 sec and the whole root section depicted in Figure 3(b) was considered as region of interest (ROI) for signal detection. During hyperpolarising conditions the emission intensity of the FRET donor (ECFP) decreased concurrently with an increasing emission intensity of the FRET acceptor (cpVenus) indicating an increased FRET efficiency due to a higher level of cytosolic Ca2+ (Figure 3a). The FRET efficiency can be expressed as a ratio value that is calculated by dividing the emission intensity of cpVenus by the emission intensity obtained for ECFP (Figure 3a). Thereby the ratio of cpVenus/ECFP is proportional to the concentration of free [Ca2+]cyt in such that the ratio value increases with rising Ca2+ levels. Compared to the typical ‘spike and shoulder’ shaped Ca2+ transients observed in Arabidopsis guard cells (Allen et al., 1999, 2000, 2001), we found that in response to hyperpolarising conditions, the shoulder of the Ca2+ peaks in root cells was much less pronounced (Figure 3a). This observation could be due to a combination of reasons. First, Allen et al. did single-cell measurements, while the Ca2+ transients shown in Figures 3 and 5 are derived from a whole population of cells. Second, instead of YC2.1 used by Allen et al., our constructs contain YC3.6 that has a higher dynamic range and a lower affinity for Ca2+ (Miyawaki et al., 1999; Nagai et al., 2004). Third, the respective Ca2+ influx and efflux elements in guard cells might have a different kinetic per se compared to those of root cells. To visualize a single hyperpolarisation-activated Ca2+ transient in a whole root segment, ratiometric image analysis was performed (Figure 3b). Ratiometric imaging is based on the pixel-by-pixel division of the cpVenus image by the ECFP image and generates a third, ratiometric image that allows monitoring of tissue or cell-specific changes in free [Ca2+]cyt. A time laps containing part of the ratiometric images corresponding to Figure 3(a,b) is provided online (Movie S1).

Figure 3.

 Ubiquitously localized YC3.6 reports cytosolic Ca2+ dynamics in different tissues.
Live cell Ca2+ imaging was performed on roots (a, b) and cotyledons (c, d) of 5–6-day-old Arabidopsis seedlings expressing YC3.6-Bar. Cytosolic Ca2+ transients were either induced by alternate application of depolarisation and hyperpolarisation buffer (a, b) or by exchanging room-tempered with ice-cold medium (c, d). The emission intensities of cpVenus and ECFP are expressed as relative fluorescence units (RFUs) and were monitored over time (a and c). The calculated emission ratios between cpVenus and ECFP fluorescence intensities indicate changes in free cytosolic [Ca2+] (a, c). ROIs considered for measuring RFU in (a) and (c) correspond to whole root sections depicted in (b) and (d), respectively. Time series of ratiometric images illustrates a single cytosolic Ca2+ transient in root (b) and in cotyledons (d). Changes in cytosolic free Ca2+ are visualized in false colours as indicated by the colour bars. Scale bars represent 100 μm (b, d). Part of the ratiometric image sequences are provided online (Movies S1 and S2).

To further examine the reliability and consistency of our Ca2+ reporter system, we analyzed in vivo Ca2+ dynamics in cotyledons of stable transformed 6-day-old Arabidopsis seedlings expressing YC3.6-Bar (Figure 3c,d; Movie S2). To induce cytosolic Ca2+ influx we applied cold-shock treatment that has been shown before to trigger transient changes in [Ca2+]cyt (Knight et al., 1996; Allen et al., 2000; Carpaneto et al., 2007). Detached cotyledons were fixed in a custom-built chamber and rapid exchange of room-tempered versus ice-cold media (1/10 MS, 0.5% sucrose, 10 mm MES–KOH pH 5.8) was performed manually using pipettes. Cold treatment was applied approximately every 5 min and the emission intensities of cpVenus and ECFP and their corresponding ratio were followed over time (Figure 3c). The whole leaf segment depicted in Figure 3(d) was chosen as a ROI for signal detection. Like for hyperpolarisation-activated [Ca2+]cyt transients in roots we also observed a sharp spike-like shape of the peak, indicating that the majority of cells respond almost simultaneously in a defined temporal frame to the treatment. We found that the amplitude of cold-induced [Ca2+]cyt transients is attenuated after repetitive cold-shock treatments (Figure 3c), an observation which has been made previously by Knight et al. (1996). A single cold-induced [Ca2+]cyt transient in epidermal cells of cotyledons is visualized in the ratiometric image sequence shown in Figure 3(d). A movie comprising part of the ratiometric image sequence shown in Figure 3(c) is provided online (Movie S2).

Subcellular targeting of yellow cameleons

The specificity of Ca2+ signatures depends on the spatio-temporal composition of the signal, which is thought to be confined to the vicinity of Ca2+ release sites and to be shaped by the involvement of different Ca2+-stores. In order to facilitate the identification and characterization of components and mechanisms leading to spatially defined changes in [Ca2+]cyt we developed several targeted versions of yellow cameleons to increase the resolution of live cell Ca2+ imaging in plants. The localized YC versions were generated by fusing either peptide tags or small membrane-integral proteins to the N- or C-terminus of the sensor that confer localization either in the cytoplasm (NES), the nucleoplasm (NLS) or at the cytosolic faces of the plasma membrane (PM) and the tonoplast (TP) (Figure 4; Table 1).

Figure 4.

 Differential targeting of yellow cameleons.
N- or C-terminal fusion of targeting peptides or proteins directs yellow cameleon to specific cellular sites (a–h). CLSM analysis was performed in roots of 5-day-old Arabidopsis seedlings using settings for cpVenus excitation and emission. (a, b) NES-YC3.6 is only detected in the cytosol but not in the nucleoplasm.
(c, d) Vice versa, NLS-YC3.6 is exclusively localized to the nucleus. (c) Maximum projection of 20 optical sections.
(e, f) In lines expressing PM-YC3.6-LTI6b yellow cameleon is localized at the plasma membrane.
(g, h) Tonoplast localization was inefficient in plants expressing TP-D3cpv. Besides the vacuolar membrane, cpVenus fluorescence is also detected to considerable amounts within the cytosol. (a–h) Scale bars represent 20 μm. NES, nuclear export signal; NLS, nuclear localization signal; PM, plasma membrane; TP, tonoplast.

To be able to distinguish between nuclear and cytosolic Ca2+ signals, we expressed YC3.6 fused to a nuclear localization signal (NLS) and a nuclear export signal (NES), respectively. For nuclear export we have chosen the NES of rabbit heat stable protein kinase inhibitor α (PKIα) (Wen et al., 1995). Nuclear localization was achieved using the NLS of simian virus 40 (SV40) large-T protein (Kalderon et al., 1984). Both peptides have been successfully applied in Arabidopsis for the targeting of proteins of interest (Matsushita et al., 2003). We performed CLSM analysis with roots of 5-day-old transgenic Arabidopsis seedlings using settings for cpVenus to verify the correct localization of the targeted YC constructs. NES-YC3.6 was found to be homogeneously distributed within the cytosol and as expected no cpVenus signal was detected in the nuclei (Figure 4a,b). Vice versa, in transgenic seedlings expressing NLS-YC3.6 fluorescence was exclusively detectable in the nuclei (Figure 4c,d).

Targeting of YC3.6 to the PM was achieved by fusing the sensor to the N-terminus of the small membrane resident protein LTI6b that has been shown previously to localize specifically to the PM (Cutler et al., 2000; Robert et al., 2005; Schleifenbaum et al., 2010). In plants expressing PM-YC3.6-LTI6b, distinct labelling of the plasma membrane was observed (Figure 4e,f).

As the vacuole constitutes one of the major internal Ca2+ storage compartments (White and Broadley, 2003), we also used the N-terminal peptide of the calcineurin-B-like protein 2 (CBL2) that was shown to confer tonoplast localization of fusion proteins (Batistic et al., 2008, 2010) to generate a Ca2+ sensor associated with the tonoplast. Stably transformed Arabidopsis seedlings expressing TP-D3cpv were examined for fluorescence of cpVenus using CLSM analysis. Besides labelling of the tonoplast, we observed also moderate labelling of the cytosol in at least 12 independent T-DNA insertion lines (Figure 4g,h). This situation restricts the usefulness of the TP-D3cpv expressing lines, since the detection of vacuolar-derived Ca2+ fluxes might be compromised by signals that originate from other cellular Ca2+ stores. However, comparative analyses of Ca2+ dynamics that combine TP-D3cpv with PM-YC3.6-LTI6b should allow us to distinguish Ca2+ signatures emanating from extracellular or tonoplast stores, respectively.

Yellow cameleons visualize subcellular calcium dynamics in Arabidopsis with high temporal and spatial resolution

To verify the functionality of the targeted sensors NES-YC3.6 and NLS-YC3.6, we performed Ca2+ measurements in 5- to 6-day-old transgenic Arabidopsis seedlings using epi-fluorescence microscopy (Figure S2). Whole seedlings were mounted in a custom-built perfusion chamber to achieve exchange of the surrounding liquid. In plants expressing NLS-YC3.6 we could induce repetitive nuclear Ca2+ transients by alternate application of ice-cold versus room-tempered buffer (Figure S2a,b). In addition to the cold treatment, consecutive sequences of de- and hyperpolarisation buffer also led to repetitive Ca2+ transients (Figure S2c–f). Ratio changes given in Figure S2(a,c,e) were recorded from root sections shown in Figure S2(b,d,f). Interestingly, we observed that hyperpolarisation-activated Ca2+ transients in the nucleoplasm and the cytoplasm exhibit a similar shape and duration (compare Figure S2d and f), which suggests that for these stimuli nuclear Ca2+ changes mirror the Ca2+ dynamics in the cytoplasm.

Application of cold-shock and hyperpolarising conditions induce very fast and global Ca2+ responses throughout cells and tissues. In order to demonstrate the capability of the targeted sensors to visualize Ca2+ transients with high spatial and temporal resolution, we performed CLSM-based Ca2+ imaging using ATP as an external stimulus. Although there is not much known about the signalling mechanism that follows extracellular ATP recognition in plants, the phenomenon of ATP-induced Ca2+ transients has been documented very well in Arabidopsis (Demidchik et al., 2003; Jeter et al., 2004; Rincón-Zachary et al., 2010; Tanaka et al., 2010). We monitored Ca2+ dynamics in epidermal root cells of 5–7-day-old Arabidopsis seedlings expressing NES-YC3.6, NLS-YC3.6 or PM-YC3.6-LTI6b, respectively (Figure 5). To allow application of ATP during the measurements, the seedlings were imaged in a custom-built perfusion chamber filled with a defined volume of liquid 0.5× MS media (0.5× MS salt, 1% sucrose, 10 mm MES–KOH pH 5.8). In seedlings expressing NES-YC3.6, Ca2+ oscillations were recorded around 420 sec after adding ATP to a final concentration of 100 μm (100 μm MgATP, 100 μm Tris-MES, pH 7.5) to the bathing solution (Figure 5a,b). Oscillations took place non-synchronously, even within one and the same cell (compare ROI1 and ROI2, Figure 5a,b). In addition we found, that changes in cytosolic Ca2+ are restricted to certain regions within the imaged tissue (compare RO1 and RO2, with ROI3, Figure 5a,b). Ca2+ oscillations were also recorded in transgenic lines expressing the nuclear Ca2+ indicator NLS-YC3.6. Approximately 15 min after applying ATP to a final concentration of 100 μm, several nuclei of epidermal root cells showed non-synchronous Ca2+ oscillations (Figure 5c,d; Movie S3). When we applied 100 μm ATP to seedlings expressing the PM targeted Ca2+ sensor PM-YC3.6-LTI6 (Figure 5e), we were not able to observe Ca2+ transients (data not shown). A previous study, performed in human cell lines, reported reduced sensitivity of a plasma membrane targeted YC sensor (Heim and Griesbeck, 2004). It was speculated that endogenous CaM might bind to YC which would reduce its dynamic range (Palmer et al., 2006). Accordingly, when we increased the extracellular ATP concentration to 1 mm (1 mm MgATP, 1 mm Tris–MES, pH 7.5), we were able to monitor spatially defined Ca2+ transients in PM-YC3.6-LTI6b expressing seedlings (Figure 5e,f). ROI1 and ROI2 of Figure 5(e,f) illustrate consecutive Ca2+ transients of two adjacent cells. It is remarkable that a region within the same cell (ROI3, Figure 5e,f), shows a much less pronounced Ca2+ response, which indicates that there are indeed local maxima of [Ca2+]cyt close to the site of Ca2+ release. Using ATP as a trigger for cytosolic Ca2+ release and CLSM for imaging, allowed us to demonstrate that YC sensors that are specifically targeted to defined cellular compartments are capable of resolving locally defined and temporally specific and distinct Ca2+ signalling events. These findings establish them as powerful tools for the in-depth analysis of microdomain Ca2+ dynamics.

Figure 5.

 YC3.6 constructs report Ca2+ transients in high spatial and temporal resolution.
Live cell imaging was performed on roots of 5- to 7-day-old Arabidopsis seedlings expressing different versions of the targeted sensors. To induce Ca2+ transients, ATP was added to the bathing solution at different time points and in different concentrations.
(a, b) In root cells expressing NES-YC3.6 spatially defined Ca2+ oscillations were detected approximately 420 sec after application of 100 μm ATP.
(c, d) Around 15 min after application of 100 μm ATP, root cell cells expressing NLS-YC3.6, showed oscillation in nuclear Ca2+ concentration.
(e, f) PM-YC3.6-LTI6b expressing root cells reported localized changes of submembrane Ca2+ levels 280 sec after application of 1 mm ATP. (a, c, e) Scale bars represent 100 μm. ROIs considered for ratiometric measurements are indicated in coloured boxes. (b, d, f) Data shown are representative for at least five independent experiments.

Yellow cameleons indicate nuclear and cytoplasmic Nod factor induced calcium spiking in Lotus

To investigate function of the UBQ10 promoter and the Ca2+ sensors in plant species other than Arabidopsis, we measured Ca2+ oscillations that occur as part of the symbiotic interaction of legumes with rhizobia. This so-called Ca2+ spiking can be detected in root cell nuclei, as well as in the adjacent cytoplasm, following perception of the bacterial nodulation (Nod) factors (Ehrhardt et al., 1996; Sieberer et al., 2009). Stable transgenic Lotus japonicus lines expressing the Ca2+ reporter constructs NLS-YC3.6 or NES-YC3.6 were successfully generated by A. tumefaciens mediated transformation. CLSM analysis of the respective lines revealed a clear nuclear localization of YC3.6, without cytoplasmic background signal (Figure 6a). We found that within the nucleus the fluorescence is uniformly distributed, except for the nucleolus that showed a slightly weaker signal. In case of NES-YC3.6 transformed Lotus plants, fluorescence was exclusively observed in the cytoplasm (Figure 6b), while the nuclei were devoid of any signal (data not shown). In general, the signal intensities detected in NES-YC3.6 expressing lines were found to be lower than in NLS-YC3.6 expressing plants, except for root hair cells with a cytoplasmic accumulation in the tip region (Figure 6b). The localization of both cameleon constructs appeared ubiquitous throughout the plant and the expression of the transgenes had no obvious effect on root hair or overall plant growth and morphology.

Figure 6.

 Nod factor induced Ca2+ spiking in Lotus.
Single cell Ca2+ imaging was performed in L. japonicus roots expressing YC3.6 reporter constructs. cpVenus and brightfield images of root hairs expressing NLS-YC3.6 (a); and NES-YC3.6 (b) are shown. Regions of interest used to measure cpVenus and ECFP intensities are indicated by white circles. Nuclear Ca2+ spiking occurred after 10−8 m Nod factor addition and continued for 60 min in root hair 1 (a1, c), whereas spiking stopped after 30 min in root hair 2 (a2, c). In addition to nuclear Ca2+ oscillations, continuous cytosolic Ca2+ spiking following Nod factor addition could be visualized in the cytoplasmic tip region by use of the NES-YC3.6 construct (b, d). A side-by-side comparison of 5-min measurements between nuclear (e) and cytoplasmic (f) spiking in a root hair cell and nuclear spiking in an upper cortex cell (g) shows slight differences in the overall spike shape. Scale bars represent 50 μm (a, b).

Ratiometric measurements of Ca2+ spiking in Lotus roots were conducted using CLSM analysis. We focussed on growing root hair cells, since these are likely the primary physiological targets of rhizobial Nod factor perception (Oldroyd and Downie, 2008). To assess the basic signal intensity, the measurements were started 5 min prior to Nod factor addition. After pipetting 10−8 m of aqueous Nod factor solution directly onto the root, both nuclear (Figure 6a,c) and cytoplasmic spiking (Figure 6b,d) were observed in root hairs of NLS-YC3.6 and NES-YC3.6 expressing plants respectively. For the majority of root hair cells (>85%) Ca2+ spiking continued for the duration of the measurement (Figure 6c–1,d). In rare cases (one out of 20 cells) either no spiking could be measured or spiking stopped (two out of 20 cells) after several minutes (Figure 6c–2). Measurements of random cells at later time points revealed that spiking continued several hours after Nod factor addition.

Previously it was shown that Ca2+ spiking is a cell-autonomous process (Ehrhardt et al., 1996; Sieberer et al., 2009). We could confirm this observation in such as adjacent root hair cells initialized Ca2+ oscillations at different time points after Nod factor addition and ongoing spiking was found to be independent and non-synchronous between the cells (Figure 6a,c). Like shown for perinuclear spiking in L. japonicus (Harris et al., 2003) and nuclear spiking in Medicago truncatula (Sieberer et al., 2009), we also observed a typical biphasic asymmetric shape of the Ca2+ spikes after Nod factor treatment in NLS-YC3.6 transformed plants (Figure 6e,g). An initial fast phase is most likely caused by the opening of Ca2+ channels, allowing Ca2+ to flow from its stores down its concentration gradient, while a longer downward phase corresponds to a slower active Ca2+ reuptake into the internal stores (Oldroyd and Downie, 2006). However, for cytoplasmic spiking in root hair cells this characteristic shape was much less pronounced (Figure 6f), possibly because the cytosolic tip region that was measured is some distance away from the Ca2+ stores in the nuclear envelope and the endoplasmic reticulum, which are assumed to drive Nod factor induced Ca2+ spiking via the action of associated Ca2+ channels and pumps. Both the NLS-YC3.6 and the NES-YC3.6 constructs represent valuable tools for studying Nod factor induced calcium oscillations. Compared with the laborious injection of calcium sensitive dyes into root hairs to measure calcium spiking, using transgenic plants stably expressing these Ca2+ reporters enables the screening of much larger numbers of events and offers the additional possibility to easily study Ca2+ oscillations in non-root hair cells.

Concluding Remarks

The set of vectors and transgenic lines presented here combines stable and constitutive expression of the cameleons YC3.6 and D3cpv with highly improved spatial resolution through subcellular targeting. All lines are phenotypically indistinguishable from wild-type plants, indicating that expression of the sensor does not interfere with calcium signalling. Most importantly, these vectors enable analyses of Ca2+ dynamics in a non-preceded spatial resolution and specificity. Already our first comparative investigations of Ca2+ dynamics in root cells of Medicago and Arabidopsis, respectively, provided clear evidence for remarkable and stimulus-specific differences in nuclear and cytoplasmic calcium responses. Moreover, the availability of targeted Ca2+ sensors for plant research facilitates the design and performance of novel experimental approaches that address the interconnection of Ca2+ in distinct cellular compartments. For stimuli that evoke simultaneous Ca2+ rises in the nucleus and the cytoplasm, a simultaneous combination of PM-YC3.6-LTI6b with NLS-YC3.6 would allow to compare Ca2+ dynamics at plasma membrane microdomains with the Ca2+ dynamics in the cytoplasm. In summary, the reported cameleon vector set represents a comprehensive toolbox for live cell Ca2+ imaging that holds potential for novel approaches in exploring Ca2+ -mediated signalling processes in different plant species.

Experimental Procedures

Stable transformation of Arabidopsis plants

The binary constructs were introduced into the Agrobacterium tumefaciens strain GV3101:pMP90 and selected on 5 μg ml−1 rifampicin, 10 μg ml−1 gentamycin and 100 μg ml−1 spectinomycin. Arabidopsis thaliana ecotype Col-0 wild-type plants were used for transformation using standard procedures.

Stable transformation of Lotus japonicus plants

Lotus japonicus ecotype Gifu and MG20 plants stably expressing NES-YC3.6 and NLS-YC3.6 were generated by Agrobacterium meditated transformation as described (Kato et al., 2005) with the following minor modifications: Plates were made with 0.6% Gelrite® (Carl Roth, instead of phytagel and the antibiotic claforan was replaced by cefotaxim (Hexal AG,

Plant materials and growth conditions for Arabidopsis

Arabidopsis thaliana ecotype Col-0 seeds were grown on media containing 1× MS, 1% sucrose with pH adjusted to 5.8 with KOH. Plates were solidified using 0.6% phyto agar. Agar and MS basal salt mixture were purchased from Duchefa ( Seeds were surface-sterilized with ethanol followed by stratification for 48 h at 4°C. Seedlings were grown at 22°C with cycles of 16 h light and 8 h darkness. Transgenic plants were selected on medium supplied with either 4 μm l-Methionine Sulfoximine (MSX) or 50 μg μl−1kanamycin.

Plant materials and growth conditions for Lotus

Lotus japonicus T2 seeds stably transformed with the NLS-YC3.6 and NES-YC3.6 constructs were scarified with sandpaper, sterilized with 2% NaOCl containing 0.1% sodium dodecylsulfate, washed and incubated over night in sterile water. Germinated seedlings were placed on filter paper on 0.8% Bacto-Agar plates and grown for 6 days (24°C, 16 h light/8 h dark). The plants were transferred to 0.8% Gelrite® (Carl Roth) plates with ½ strength B&D medium (Broughton and Dilworth, 1971) supplemented with 2 mm MgSO4 and 0.1 μm AVG [l-α-(2-aminoethoxyvinyl)-glycine; Sigma-Aldrich,]. The roots were covered with a gas permeable plastic membrane (lumox® Flim 25; Sarstedt, and the plants were grown for at least four more days up to a maximum of 4 weeks prior to imaging.

Nod factor extraction

Mesorhizobium loti harboring the pMP2112 plasmid (Niwa et al., 2001), which expresses L. japonicus specific Nod factor upon induction with Naringenin (Carl Roth), were grown for 2–4 days at 28°C in TY medium (Beringer, 1974). The culture was centrifuged, washed and resuspended in modified B-medium (Spaink et al., 1991) to a final OD600 of 2. A 1:100 dilution of the resuspension was added to modified B-medium containing 1 μm Naringenin and grown for 24 h at 28°C. Cells were removed by centrifugation and the supernatant was run through a Sep-Pak C18 cartridge (Waters, at a flow rate of 300 ml h−1. The run-through was discarded. The retained exudate was eluted from the C18 cartridge by a series of methanol dilutions (20, 40, 60, 80 and 100%). The methanol was evaporated and the fractions were analyzed by high-performance liquid chromatography (HPLC) revealing high Nod factor concentrations in the 80 and 100% fraction. Stock solutions of Nod factor were diluted to 10−6 m in H2O. A working concentration of 10−8 m was diluted in ½ strength B&D for Ca2+ spiking induction.

Localization studies

The subcellular localization of the YC3.6 constructs was assessed by CLSM using either a Leica TCS SP2 or a Leica TCS SP5II ( confocal laser scanning microscope. The microscopes were equipped with a Leica HCX PL APO ×63 water immersion objective. For image acquisition, the respective Leica confocal software has been used. cpVenus was excited using the 488 nm laser line and the emission was detected between 520 and 570 nm.

Calcium imaging using fluorescence microscopy

Ca2+ measurements using epi-fluorescence microscopy (Figures 3 and S2a,b) were performed on an inverted fluorescence microscope (DM IRBE; Leica) equipped with a cooled charge-coupled (CCD) device camera (SenSys Photometrics, and a Leica HC PL FLUOTAR ×20/0.5 air objective. ECFP was excited using a UV (ultraviolet) mercury arc lamp and a 436/20 excitation filter. Excitation light intensity was reduced by a 5% transmission neutral density filter. Dual ratiometric measurements between cpVenus and ECFP emission were made by interchanging 535/30 nm (cpVenus) and 480/40 nm (ECFP) filters in front of the CCD camera. Filter changes and shutter control for periodic excitation were controlled by a MAC 5000 controller system (Ludl Electronic Products, and ratiometric images were collected in 5 sec intervals using MetaFluor 6.2r5 software (Universal Imaging, For imaging, the 5- to 6-day-old plate grown Arabidopsis seedlings were sandwiched between a cover slip and a thin layer of cotton in a custom-built perfusion chamber. To hold the seedlings in position, the sample-cotton sandwich was slightly pressed against the cover slip by adding an open cover plate from the top. Prior to measurement seedlings were allowed to recover in reaction media or buffer, for approximately 15–30 min. Repetitive Ca2+ transients in roots were either induced by exchanging depolarisation (100 mm KCl, 10 mm MES–Tris pH 6.15) with hyperpolarisation buffer (0.1 mm KCl, 10 mm CaCl2, 10 mm MES–Tris pH 6.15) or by treatment with ice-cold buffer (1/10 MS salt, 0.5% sucrose, 10 mm MES–KOH pH 5.8). For hyperpolarisation treatments, buffer changes were achieved using a peristaltic pump, whereas changes between room-tempered and ice-cold buffer were carried out manually using a pipette.

Ca2+ measurements shown in Figure S2(c–f) were analyzed on an inverted fluorescence microscope (DMI 6000B; Leica) equipped with an emission filter wheel (Ludl Electronic Products) and an Orca ER CCD camera (Hamamatsu, Excitation was provided by a xenon lamp through a 436/20 nm filter and emission filters were 485/30 nm (ECFP) and 535/40 nm (cpVenus). Image acquisition and ratio calculations were performed using Openlab 5.2 software (Perkin Elmer, Images were collected every 6 sec and the ratio between cpVenus and ECFP emission intensities was calculated in ROIs. For fixation, seedlings grown on vertical plates were placed in a custom-built perfusion chamber onto the surface of a liquid low melt agarose bed (0.7%, 25–30°C, Ultrapure low-melting point (LMP) agarose; Gibco BRL, Seedlings were allowed to recover for 1 h in the light at 22°C, before starting the measurements. Repetitive Ca2+ transients in roots were induced by alternate pulses of depolarising (50 mm KCl, 10 mm MES–Tris pH 5.6) and hyperpolarising buffer (1 mm KCl, 1 mm CaCl2, 10 mm MES–Tris pH 5.6).

Calcium imaging using CLSM

CLSM-based Ca2+ imaging illustrated in Figure 5 was performed on a Leica SP5II system equipped with an inverted DMI6000 microscope stand. Images were recorded using a HCX PL APO lambda blue 63.0 × 1.20 WATER UV objective. The imaging parameters were as follows: image dimension (512 × 512), pinhole (3.69 airy unit), scanning speed (400 Hz), line average (2). Images acquisition was proceeded every 3 sec. ECFP was excited using the 458 nm laser line of the Argon laser. Fluorescence intensity values for ECFP (465–500 nm) and cpVenus (520–570 nm) were detected simultaneously in defined ROIs. After background subtraction the ECFP/cpVenus ratio was calculated followed by normalising the ratio to the ratio value at time point zero (R0). For imaging, the 5- to 7-day-old Arabidopsis seedlings were sandwiched between a cover slip and a thin layer of cotton in a custom-built perfusion chamber as described above. Prior to measurement seedlings were allowed to recover in liquid media for approximately 15–30 min. To obtain the desired final concentration of ATP, an appropriate volume of stock solution was added to the bathing medium without interrupting image acquisition.

FRET-based Ca2+ imaging shown in Figure 6 was performed with an upright Leica SP5 confocal laser scanning microscope using a long-distance Leica HCX APO L40×/0.80 W U-V-1 objective. During culturing and microscopy the seedlings were constantly kept under a gas permeable plastic film (lumox® Flim). As the membrane has the same refraction index as water, direct microscopy without removal of this film is possible using a dipping lens. This setup has been shown to be advantageous for prolonged in vivo microscopy, as the film limits potential contaminations, reduces autofluorescence and enables a stable imaging over longer time periods by preventing evaporation and fixing the plants in place (Fournier et al., 2008). For measuring Ca2+ oscillations, resolution was set to 256 × 256 pixels, pinhole to 4 airy units (corresponding to a Z volume of 7.25 μm), scanning speed to 700 Hz, line average to 2 and an image was collected every 1.5 sec. For still images, the resolution was set to 512 × 512 pixels and the frame average to 4. FRET-based cpVenus/ECFP ratio shifts were measured by exciting the cameleon sensor with the 458 line of the Argon laser at 10% laser power and measuring the emission simultaneously at around 470–500 nm for ECFP and 530–580 nm for cpVenus. cpVenus and ECFP intensity values were measured for a defined ROI and the background fluorescence in a region outside of the root was subtracted. To induce Ca2+ spiking, 10−8 m of aqueous Nod factor dilution was applied directly onto the roots below the plastic film (lumox® Flim) with a syringe and needle. Imaging was initialized for 5 min prior to Nod factor addition to assess the baseline signal levels and then 2–10 min after Nod factor application for up to 1 h. After this time the signal strength did not decay for more than 10%, therefore much longer measurements potentially using a slower scanning rate should be easily possible.


The authors would like to thank Zhao-Xin Wang and Beate Schöfer for excellent technical assistance. We are also grateful to Julian Schroeder for kindly providing the pNB96-YC3.6 construct. This work was supported by the Deutsche Forschungsgemeinschaft through Grants within FOR964 to M.P., J.K. and K.S.