Cytoplasmic free calcium ([Ca2+]cyt) acts as a stimulus-induced second messenger in plant cells and multiple signal transduction pathways regulate [Ca2+]cyt in stomatal guard cells. Measuring [Ca2+]cyt in guard cells has previously required loading of calcium-sensitive dyes using invasive and technically difficult micro-injection techniques. To circumvent these problems, we have constitutively expressed the pH-independent, green florescent protein-based calcium indicator yellow cameleon 2.1 in Arabidopsis thaliana (Miyawaki et al. 1999; Proc. Natl. Acad. Sci. USA 96, 2135–2140). This yellow cameleon calcium indicator was expressed in guard cells and accumulated predominantly in the cytoplasm. Fluorescence ratio imaging of yellow cameleon 2.1 allowed time-dependent measurements of [Ca2+]cyt in Arabidopsis guard cells. Application of extracellular calcium or the hormone abscisic acid (ABA) induced repetitive [Ca2+]cyt transients in guard cells. [Ca2+]cyt changes could be semi-quantitatively determined following correction of the calibration procedure for chloroplast autofluorescence. Extracellular calcium induced repetitive [Ca2+]cyt transients with peak values of up to approximately 1.5 μM, whereas ABA-induced [Ca2+]cyt transients had peak values up to approximately 0.6 μM. These values are similar to stimulus-induced [Ca2+]cyt changes previously reported in plant cells using ratiometric dyes or aequorin. In some guard cells perfused with low extracellular KCl concentrations, spontaneous calcium transients were observed. As yellow cameleon 2.1 was expressed in all guard cells, [Ca2+]cyt was measured independently in the two guard cells of single stomates for the first time. ABA-induced, calcium-induced or spontaneous [Ca2+]cyt increases were not necessarily synchronized in the two guard cells. Overall, these data demonstrate that that GFP-based cameleon calcium indicators are suitable to measure [Ca2+]cyt changes in guard cells and enable the pattern of [Ca2+]cyt dynamics to be measured with a high level of reproducibility in Arabidopsis cells. This technical advance in combination with cell biological and molecular genetic approaches will become an invaluable tool in the dissection of plant cell signal transduction pathways.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Many physiological stimuli in plant cells induce elevations in the cytoplasmic free calcium concentration ([Ca2+]cyt) which acts as a second messenger in signal transduction cascades (Sanders et al. 1999; Trewavas & Malhó 1997). Changes in [Ca2+]cyt occur in stomatal guard cells in response to a wide variety of exogenous and endogenous signals. Changes in [Ca2+]cyt regulate the activity of ion transporters in the plasma and vacuolar membranes which control the turgor of guard cells and hence regulate stomatal aperture (for reviews see Assmann & Shimazaki 1999; McAinsh et al. 1997). Guard cells have become a widely used model cell system for studying Ca2+- and membrane-based signal transduction in plants.
In the present study, we transformed Arabidopsis plants to express genetically encoded calcium indicators based on green fluorescent protein (GFP) (Miyawaki et al. 1997; Miyawaki et al. 1999). These calcium indicators consist of an enhanced cyan fluorescent protein (ECFP) linked via calmodulin and a calmodulin binding peptide (M13), to an enhanced yellow fluorescent protein (EYFP) (as developed by Miyawaki et al. 1997; Miyawaki et al. 1999). The binding of calcium ions to the calmodulin domain causes it to associate with the M13 peptide. This causes a conformational change in the molecule, bringing the two GFPs into closer molecular proximity, which in turn enhances the efficiency of fluorescent resonance energy transfer (FRET) between the two GFPs. FRET occurs when excitation energy from the higher-energy GFP (the donor, in this case the ECFP) is passed to the lower-energy GFP (the acceptor, EYFP), causing an increase in emission fluorescence of the acceptor GFP. This process is fully reversible when calcium dissociates from calmodulin, giving rise to a ratiometric indicator of cytoplasmic free calcium (see Figs 1a in Miyawaki et al. 1997). These calcium indicators have been termed ‘cameleons’ (Miyawaki et al. 1997).
To test cameleon expression in Arabidopsis guard cells, we used yellow cameleon 2.1 (YC2.1). This form of cameleon carries two mutations in the acceptor EYFP (V68L and Q69K), which render the fluorescence relatively insensitive to pH changes in the physiological range (Miyawaki et al. 1999). The pH insensitivity is a critical factor for guard cells as cytoplasmic pH changes are an important signalling mechanism in these cells (Blatt & Armstrong 1993; Irving et al. 1992; for review see Assmann & Shimazaki 1999; MacRobbie 1997). Three major patterns of [Ca2+]cyt change associated with stomatal closure have been previously reported in Commelina communis or Vicia faba guard cells which are large enough to be micro-injected with Ca2+-sensitive dyes. These are:
In this paper, we report that Arabidopsis guard cells expressing YC2.1 exhibited all three [Ca2+]cyt responses listed above. In addition, we find that [Ca2+]cyt increases do not necessarily coincide in the two guard cells of a single stomate. Using cameleon instead of conventional micro-injection methods to study calcium signalling in guard cells has a number of practical advantages (see Discussion). In addition, cameleon expression in Arabidopsis combined with cell biological, molecular biological and genetic approaches will become an invaluable tool to study the molecular mechanisms of calcium-based signal transduction in plants.
Cameleon expressed in Arabidopsis guard cells is located in the cytoplasm
We cloned the YC2.1 construct (Miyawaki et al. 1999) into a plant expression vector as described in Experimental procedures. Expression of YC2.1 was driven by the CaMV 35S promoter. Confocal microscopy of rosette leaf epidermal fragments at excitation and emission wavelengths corresponding to those of the EYFP indicated that YC2.1 was expressed at similar levels in guard cells of transformed plants (11 separate lines analysed) (Fig. 1a,b). Higher magnification using the confocal microscope revealed a cytoplasmic localization of YC2.1 (Fig. 1c,d). Fluorescence was brightest from the region surrounding the nucleus where the majority of the guard cell cytoplasm is located, but could also be observed from the nucleoplasm, from cytoplasm surrounding the vacuoles, in a thin layer of cytoplasm between vacuoles and the plasma membrane, and in cytoplasmic strands traversing the vacuole and surrounding the chloroplasts.
Confocal microscopy only allowed visualization of the EYFP emission due to the 488 nm excitation wavelength of the laser. High-resolution deconvolution fluorescence microscopy allowed three-dimensional reconstruction of separate images for both emission wavelengths of YC2.1, as shown in Fig. 1(e–h). These images also revealed YC2.1 accumulation in the cytoplasm and nucleoplasm, as has also been observed for GFP expression in Arabidopsis (Haseloff et al. 1997; Müller-Röber et al. 1998). The pattern of fluorescence was the same at both emission wavelengths (480 and 535 nm) (Fig. 1e,f), except at 480 nm a significant fluorescence was observed from the chloroplasts (Fig. 1f). Overlaying the two images demonstrated that 535 and 480 nm emissions coincide throughout the cell except in the chloroplasts where the 480 nm emission only was still observed (Fig. 1g). The clear demarcation of chloroplasts, vacuoles and the nuclear region and the observed coincidence of the 480 and 535 nm signals (Fig. 1g) indicate that mobile cytosolic factors did not affect the apparent YC2.1 distribution.
Ratiometric images produced by pixel-by-pixel division of the 535/480 wavelength images show the true YC2.1 distribution, as emission at only one wavelength would produce an infinitely small or infinitely large ratio value (Fig. 1h, pseudo-coloured red for clarity). No ratiometric signal was observed from the chloroplasts, suggesting that YC2.1 was not expressed in this organelle. This was confirmed by measuring 535 and 480 nm emissions from non-transformed stomata as shown in Fig. 1(i,j). Autofluorescence from chloroplasts was minimal at 535 nm emission (Fig. 1i, enhanced for this image of 535 emission by increasing excitation time threefold) but was clearly visible from the chloroplasts at 480 nm emission (Fig. 1j). When the 535 and 480 nm emission images were overlaid, autofluorescence from the chloroplasts coincided, and a small additional autofluorescence was observed at the lignified lip of the stomatal pores at 535 nm (Fig. 1k). Interestingly, autofluorescence from chloroplasts in mesophyll cells was significantly lower at 480 nm than in guard cells, but higher at 535 nm (data not shown), suggesting that these two cell types have different pigment contents. In addition, cameleon expression was visualized in root tips and root hairs and was also observed to be cytoplasmic in these tissues (data not shown).
These data indicate that YC2.1 expressed in guard cells is predominantly cytoplasmic, and 480 nm signals from the chloroplasts are due to autofluorescence and not YC2.1 expression in these organelles.
YC2.1 reports repetitive [Ca2+]cyt increases in Arabidopsis guard cells induced by increases in extracellular calcium
Increases in the extracellular calcium concentration have been shown to induce [Ca2+]cyt oscillations in Commelina and Arabidopsis guard cells micro-injected or acid-loaded with Fura-2 (Allen et al. 1999; McAinsh et al. 1995). When the extracellular calcium concentration bathing Arabidopsis guard cells expressing YC2.1 was increased, repetitive transients in the YC2.1 ratio were observed (Fig. 2a). This response was observed in all 30 guard cells tested. Ratio increases resulted from an increase in the fluorescence of the acceptor EYFP (Fig. 2a, lower panels). However, no perceptible or only very small decreases in the fluorescence of the donor ECFP were observed (see also Fig. 3a). This is due to the high level of chloroplast autofluorescence which increases the background against which ECFP changes are measured (see Fig. 1j). The high background at this wavelength reduces the magnitude of calcium-induced fluorescent change and therefore reduces the sensitivity of the ratio change. Nevertheless, the signal-to-noise level was sufficient to allow ratio changes to be clearly observed, and subsequently accounting for autofluorescence in the YC2.1 calibration procedure allowed approximate quantification of the [Ca2+]cyt increase (see below).
Images of the extracellular calcium-induced [Ca2+]cyt increase were captured at points corresponding to the peaks and troughs of the calcium increase occurring immediately following calcium addition. These images are shown in pseudo-colour in Fig. 2(b). Fluorescence ratio changes could be observed throughout the cells but the limited resolution in these non-confocal images did not allow any spatial analysis of the calcium increase.
The [Ca2+]cyt response immediately following an increase in extracellular calcium was always observed in both guard cells of any single stomate (Fig. 2a, left and right panels, n = 15 stomates). Transients then repeated one to four times following this initial calcium increase, an example of two increases for both cells being shown in Fig. 2(a). However, the number of repetitive transients observed in each guard cell of any single stomate was not always the same, or of the same magnitude. In Fig. 2(c), emission ratios are shown from the two cells of another stomate expressing YC2.1 and challenged with 10 mm extracellular Ca2+ where indicated. Both cells showed a characteristic [Ca2+]cyt transient immediately following Ca2+ addition (#1). However, cell 1 then showed a second subsequent increase (#2) which is not clearly defined in cell 2. This difference in the pattern of [Ca2+]cyt dynamics between cells in a single stomate was observed in 9 of the 15 stomates challenged with 10 mm external Ca2+, indicating that [Ca2+]cyt elevations are not necessarily identical in all guard cell pairs.
Addition of 1 mm Ca2+ to the extracellular medium could also induce repetitive increases in [Ca2+]cyt in Arabidopsis guard cells expressing YC2.1 (n = 6 cells from three stomata) (data not shown). As observed for 10 mm extracellular calcium, the number of repetitive transients induced by 1 mm extracellular calcium was not necessarily the same in the two guard cells of any single stomate.
YC2.1 reports repetitive [Ca2+]cyt transients induced by ABA in Arabidopsis guard cells
Many studies have indicated that ABA elicits stomatal closure by causing transient increases in [Ca2+]cyt (Allan et al. 1994; Grabov & Blatt 1998a; Irving et al. 1992; McAinsh et al. 1992; Schroeder & Hagiwara 1990). Recently it has been shown that ABA elicits repetitive [Ca2+]cyt increases in Commelina or Vicia guard cells if the cells are maintained in a hyperpolarized state by reducing the KCl concentration in the bathing buffer (Grabov & Blatt 1998a; Staxén et al. 1999). We therefore measured the effect of ABA on [Ca2+]cyt in Arabidopsis guard cells expressing YC2.1 in a buffer containing 5 mm KCl. To test for a stable resting [Ca2+]cyt, the YC2.1 ratio was monitored for 15–20 min before applying ABA (see below). In cells exhibiting a stable resting [Ca2+]cyt, addition of 10 μm ABA to the bath caused repetitive, transient increases in [Ca2+]cyt (Fig. 3a, n = 24 cells from 12 stomata). Repetitive [Ca2+]cyt transients induced by ABA have been observed in Vicia or Commelina guard cells loaded with Fura-2 (Grabov & Blatt 1998a; Schroeder & Hagiwara 1990; Staxén et al. 1999). Increases in [Ca2+]cyt occurred 12.6 ± 2.1 min after ABA addition (range 4–28 min, n = 24), and the number of transients induced by ABA varied between one and four (three repeats shown in Fig. 3a, cell 1). Increases were observed in 85% of cells treated with 10 μm ABA; only in four cells (of two stomates) from a total of 28 cells tested did ABA fail to elicit a measurable [Ca2+]cyt response (data not shown).
Measuring [Ca2+]cyt dynamics in both cells of a single stomate revealed that in the majority of stomates (9 of 12) ABA-induced [Ca2+]cyt increases did not coincide in the two cells. For the stomate in Fig. 3(a,b), both cells exhibit an initial ABA-induced [Ca2+]cyt increase at the same time following ABA addition (marked as *2). However, the increase was smaller and more prolonged in cell 2 than in cell 1. Subsequently, cell 1 showed a repetitive [Ca2+]cyt response, increasing transiently twice more, whereas similar large increases were not observed in cell 2, particularly at mark *3. Images from these cells are shown in Fig. 3(b). The spatial resolution in these images is limited because to achieve an adequate signal-to-noise ratio the pixel size was expanded to its maximum (16 pixels per bin compared to 4 pixels per bin in Fig. 2b). Nevertheless, the differential response of the two cells was clearly observed. In panel *2 (corresponding to point *2 in Fig. 3a), ratiometric increases were observed in both cells. However, in panel *3 (corresponding to point *3 in Fig. 3a), a clear transient increase was only observed in cell 1 (upper cell).
Although the cells in most stomates did not respond in a synchronized manner, occasionally (3 of 12 stomates) a more coordinated response was observed. In Fig. 3(c), emission ratios are shown from two cells of a single stomate expressing YC2.1, challenged with 10 μm ABA where indicated. Both cells show characteristic [Ca2+]cyt transients following ABA addition which appear similar in magnitude and timing in each of the two cells.
The differential response of the two cells of a single stomate to ABA might be explained if YC2.1 expression were different in the each of the cells, as higher expression levels might buffer [Ca2+]cyt changes. Expression levels could be directly assessed from the intensity of the EYFP signal from each cell. Of all stomates measured in this study (n = 35), the maximum difference in fluorescent intensity between two cells of any single stomate was 23% (see Fig. 2a) and for the cells which were treated with ABA the variation between cells of single pairs was only 6.1 ± 2.3%. These data indicate that expression of YC2.1 was very similar in the two cells of a stomate. The differential responses of the two cells to either calcium (Fig. 2c) or ABA (Fig. 3a) cannot therefore be explained by different levels of YC2.1 buffering calcium changes differently in the two cells. This conclusion is highlighted by the data in Fig. 3(a) where the EYFP are virtually identical in the two cells (lower panels) indicating equal YC 2.1 expression, yet the cells respond to ABA with a different pattern of [Ca2+]cyt change. Differential responses could also not be attributed to differences in YC2.1 expression between transgenic lines. Expression levels in the 11 independent transgenic lines selected in this study were very similar (variation of only 25% between the highest and the lowest) and typical patterns of ABA and calcium responses were seen in all lines. Therefore, it can be concluded that the expression levels of YC2.1 observed here did not disrupt the patterns of [Ca2 + ]cyt changes induced by ABA or external calcium.
Yellow cameleon expression does not disrupt signal transduction in Arabidopsis guard cells
Increasing the extracellular Ca2+ concentration leads to an increase in [Ca2+]cyt (Fig. 2) and induces stomatal closure, as demonstrated in Arabidopsis (Allen et al. 1999; Webb & Hetherington 1997) and other species (DeSilva et al. 1985; McAinsh et al. 1995; Schwartz 1985). ABA also causes stomatal closure via a transduction pathway involving an increase in [Ca2+]cyt (Allen et al. 1999; McAinsh et al. 1990; Schroeder & Hagiwara 1989; Webb & Hetherington 1997). We tested whether expression of yellow cameleon 2.1 in Arabidopsis guard cells interfered with signalling pathways activated by cytoplasmic calcium, by measuring calcium- and ABA-induced closure of pre-opened stomata in these transgenic plants. Stomatal aperture measurements in Fig. 4 clearly demonstrate that calcium- and ABA-induced closure occurs to the same extent (P > 0.3) in plants expressing YC2.1 as in non-expressing plants. Normal stomatal closure in plants expressing YC2.1 indicates that calcium-dependent signal transduction pathways are unaffected by YC2.1 expression.
Spontaneous [Ca2+]cyt transients measured by YC2.1 in Arabidopsis guard cells
In Arabidopsis guard cells expressing YC2.1 (bathed in 5 mm KCl and 50 μm CaCl2), spontaneous, repetitive increases in [Ca2+]cyt were observed in some cells (in both cells in three of 35 stomata measured) in the absence of an exogenously added ABA or Ca2+ stimulus. Figure 5 shows spontaneous transients in YC2.1 ratio measured in two guard cells of a single stomate. As observed for ABA-induced [Ca2+]cyt transients, spontaneous increases in [Ca2+]cyt were not necessarily synchronized between the two guard cells. In Fig. 5, the point *1 marks an increase in cell 1 which was not observed in cell 2, whereas #2 marks a rise in cell 2 which was absent in cell 1. In addition, cell 1 had an elevated [Ca2+]cyt when the recording commenced, which was not observed in cell 2.
Calibrating the magnitude of the [Ca2+]cyt change measured with YC2.1
The YC2.1 ratio in cells before stimulation with calcium or ABA was 2.02 ± 0.27 (n = 70, range 1.85–2.31). Addition of 10 mm extracellular calcium increased the ratio to 3.24 ± 0.22 (n = 30) measured at the peak of the transients, whereas 10 μm ABA produced transients with a mean peak ratio of 2.75 ± 0.39 (n = 24). To calibrate these increases, Rmin and Rmax were determined in vitro as described in Experimental procedures. Values for Rmin of 1.81 and Rmax of 3.66 were obtained following autofluorescence correction (Experimental procedures) and were used to convert YC2.1 ratios into approximate [Ca2+]cyt concentrations by fitting to the YC2.1 in vitro calibration curve (see Miyawaki et al. 1999, Fig. 2B). The resting [Ca2+]cyt range approximated by this calibration was approximately 10–60 nm. Extracellular calcium elevated [Ca2+]cyt to peak levels ranging from approximately 450 to approximately 1600 nm, whilst ABA elevated [Ca2+]cyt with peaks levels up to approximately 580 nm. Despite the need to correct for autofluorescence, these values are similar to the magnitudes of stimulus-induced [Ca2+]cyt changes measured in plant cells using ratiometric dyes (Gilroy et al. 1991; McAinsh et al. 1990; McAinsh et al. 1992; McAinsh et al. 1995) or aequorin (Johnson et al. 1995; Knight et al. 1991; Knight et al. 1996).
Given the intrinsic errors associated with calibrating all known calcium indicators, the ability to ratiometrically measure the patterns of [Ca2+]cyt changes and differences in [Ca2+]cyt dynamics has been demonstrated to be more reliable in animal cell studies than quantification of the absolute [Ca2+]cyt concentration, and therefore often only ratio values and relative changes are appropriately reported (Dolmetsch et al. 1998; Li et al. 1998). The data presented in Figs 2, 3 and 5 not only indicate that YC2.1 expression reports [Ca2+]cyt changes in response to stimuli previously reported to modulate [Ca2+]cyt in guard cells, but also demonstrates that reproducible measurement of the pattern of [Ca2+]cyt dynamics can now be readily achieved in guard cells of Arabidopsis.
Yellow cameleon calcium indicators are expressed in Arabidopsis guard cell cytoplasm
We have created a vector (p35S-YC2.1-Kan) to constitutively express YC2.1 in plant cells under control of the CaMV 35S promoter (Fig. 6). In addition, this vector has been engineered to contain features that will be useful for further applications (Fig. 6 and Experimental procedures). Unique cloning sites have been added upstream and downstream of the YC2.1 insert. This will enable N or C terminal fusion of signal transduction proteins with YC2.1, addition of subcellular localization sequences, or transfer of YC2.1 into different expression vectors. Also, a six-histidine tag was added to the YC2.1 insert to enable purification of the YC2.1 protein (Fig. 6). We also created a similar vector (p35S-YC3.1-Kan) to express YC3.1 in plants (Miyawaki et al. 1997), although the lower affinity for calcium of this cameleon makes it less suitable for studying cytoplasmic calcium dynamics in plant cells and therefore it was not used for this initial study.
Arabidopsis plants transformed with the p35S-YC2.1-Kan vector expressed YC2.1 in the guard cells and YC2.1 accumulation was observed in the cytoplasm with distribution similar to that observed in non-confocal images of calcium dyes micro-injected into guard cells (Gilroy et al. 1991; McAinsh et al. 1990; McAinsh et al. 1992).
In this paper, we have demonstrated that GFP-based cameleon calcium indicators can be used as a novel tool for studying guard cell [Ca2+]cyt. Using Arabidopsis guard cells expressing YC2.1, three typical calcium responses of guard cells could be measured.
Firstly, increasing extracellular calcium concentration gives rise to [Ca2+]cyt increases (Fig. 2) that are similar to those previously measured in micro-injected Commelina guard cells (McAinsh et al. 1995; Webb et al. 1996). In Commelina, increases in extracellular calcium induce [Ca2+]cyt increases which have a characteristic shape, consisting of an initial ‘spike’ followed by a more prolonged ‘shoulder’ (McAinsh et al. 1995; Webb et al. 1996). The response to extracellular calcium is observed so consistently that it has been used to confirm intact calcium homeostatic mechanisms in guard cells which fail to respond to other stimuli (Webb et al. 1996). In this study, the extracellular calcium-induced [Ca2+]cyt increases measured in Arabidopsis guard cells expressing YC2.1 were similar, a ‘spike and shoulder’ being observed in most of the cells (n = 25 from 30) following an increase in extracellular calcium to 10 mm (Fig. 2a).
This characteristic transient has also been measured in other plant cells or tissues responding to physiological stimuli such as osmotic shock (Cessna et al. 1998) or ozone (Clayton et al. 1999). This response may arise from calcium influx inducing internal calcium-induced calcium release mechanisms (Cessna et al. 1998; McAinsh et al. 1995). The physiological relevance of this signal in guard cells is not clear although it is interesting to note that apoplastic calcium concentration in the vicinity of guard cells is tightly regulated (DeSilva et al. 1996; Ruiz & Mansfield 1994).
Secondly, addition of extracellular ABA gives rise to repetitive [Ca2+]cyt transients in Arabidopsis guard cells expressing YC2.1. Previous studies using calcium-sensitive dyes have indicated that ABA elicits [Ca2+]cyt increases in guard cells (Allan et al. 1994; Grabov & Blatt 1998a; Irving et al. 1992; McAinsh et al. 1990; McAinsh et al. 1992; Schroeder & Hagiwara 1990; Staxén et al. 1999), although the characteristics of these increases varied considerably between these studies. Recently, Staxén et al. (1999) have demonstrated that maintaining intact Commelina guard cells in lower KCl concentrations than in most previous studies (5 mm versus 50 mm) elicits repetitive ABA-induced [Ca2+]cyt transients which had a rapid onset and a slower decline in [Ca2+]cyt. In this study, similar asymmetric repetitive ABA-induced [Ca2+]cyt transients were measured in Arabidopsis guard cells expressing YC2.1 maintained in 5 mm KCl buffer (Fig. 3, point *3, cell 1 in particular). This again suggests that YC2.1 acts as an accurate calcium indicator in Arabidopsis guard cells.
Thirdly, spontaneous transients can be measured in Arabidopsis guard cells expressing YC2.1 and maintained in 5 mm KCl buffer (Fig. 5). Similar transients have been observed in dye-loaded Commelina and Vicia guard cells and may be activated by rapid transient hyperpolarizations in the plasma membrane potential (Grabov & Blatt 1998a; Staxén et al. 1999). In addition to guard cells, repetitive [Ca2+]cyt transients have also been measured in root cells in response to nodulation signals (Ehrhardt et al. 1996), indicating the fundamental importance of dynamic calcium signals in encoding specificity in plant cell signal transduction cascades.
Although chloroplast autofluorescence at 480 nm masked large changes in the ECFP signal, correcting for this during the YC2.1 calibration enabled ratios to be converted to approximate [Ca2+]cyt values. The level of autofluorescence was relatively constant when measured in non-expressing cells (variation less than 10%, 32 ± 3 at 535 nm and 51 ± 3 at 480 nm, Experimental procedures) and these values were used to correct for autofluorescence during calibration. Although this offset in the cyan 480 nm wavelength reduced the absolute magnitude of YC2.1 ratio changes, the data in Figs 2, 3 and 5 indicate that the pattern of [Ca2+]cyt changes is not affected by this static background. In order to allow subcellular imaging in the future and to more accurately calibrate the ratio change, we propose techniques to measure autofluorescence on a cell-by-cell basis by measuring the ratio of autofluorescence at 480 nm to that at 600 nm and correcting on a pixel-by-pixel basis accordingly. At 600 nm emission, chloroplast fluorescence can be measured without interference from cameleons. This will require further technical development and could not be applied to the cells measured in this study where ratios were calculated from the average intensity at each wavelength for each cell. The 480 nm autofluorescence was only observed from guard cell chloroplasts and not from mesophyll chloroplasts. Changes in the light intensity or growth conditions may affect pigment accumulation and autofluorescence in the guard cell. Note that the above limitations do not impair the ability to resolve the temporal dynamics of [Ca2+]cyt changes in Arabidopsis guard cells.
The expression of YC2.1 in both cells of a single stomate allowed us for the first time to observe whether the two cells in one stomate showed identical patterns in [Ca2+]cyt changes. The responses to extracellular calcium or ABA and the spontaneous transients were not necessarily the same in each cell (Figs 2, 3 and 5). This is an interesting observation as the two adjacent cells have the same developmental history and experience identical growth conditions, two factors which have been proposed to explain heterogeneity in stomatal [Ca2+]cyt signalling (Allan et al. 1994; McAinsh et al. 1997; Trewavas & Malhó 1997).
Advantages of using cameleon to measure guard cell [Ca2+]cytUsing a molecularly encoded ratiometric calcium indicator in guard cells has a number of specific advantages over using ratiometric dyes or aequorin.
1 Using YC2.1 allows cytoplasmic calcium to be measured in guard cells without using technically difficult micro-injection protocols (Gilroy et al. 1991; Grabov & Blatt 1998a; Grabov & Blatt 1998b; McAinsh et al. 1990; McAinsh et al. 1995). Micro-injection is particularly difficult in the small guard cells of Arabidopsis. Using YC2.1 circumvents such technical difficulties and will significantly expedite the study of plant cell Ca2+ signalling. Note, however, that we have also adapted a technique for non-invasively loading fura-2 and indo-1 into Arabidopsis guard cells (Allen et al. 1999). This alternative approach has shown a reduction in ABA-induced [Ca2+]cyt elevations in the abi1-1 and abi2-1 mutants (Allen et al. 1999).
2 Following micro-injection or acid-loading, calcium-sensitive dyes tend to leak from plant cells. The expression of the calcium-sensitive protein YC2.1 prevents leakage problems. Although YC2.1 fluorescence can bleach with intense violet UV illumination, [Ca2+]cyt measurements from Arabidopsis guard cells could be performed for up to 1 h with YC2.1.
3 In relation to point 1, many agonists or antagonists of [Ca2+]cyt signalling are membrane-impermeable. Introduction of such compounds (in their free or caged forms) into the guard cell cytoplasm has previously required double micro-injection protocols (one for the dye and one for the agonist). The second micro-injection may result in cell death in up to 75% of cells surviving the first micro-injection (Leckie et al. 1998). The constitutive expression of a calcium indicator in these cells will allow calcium signalling modifiers to be introduced into the cytoplasm by a single micro-injection.
4 Expression of YC2.1 in both guard cells of a single stomate allows simultaneous measurement in both cells. This study has revealed that both cells do not necessarily always respond with the same pattern of [Ca2+]cyt changes.
5 Due to a higher photon production per molecule, the fluorescent intensity of cameleons is far higher than that of the calcium-sensitive luminescent protein aequorin. In addition, cameleons are not irreversibly consumed by Ca2+ and do not require an additional co-factor such as coelenterazine. Although aequorin has been used successfully to monitor whole seedling or organ [Ca2+]cyt changes (Johnson et al. 1995; Knight et al. 1991), and single time-point images of stomatal luminescence have been acquired from Arabidopsis (Trewavas & Malhó 1998), the high fluorescent intensity of cameleons makes them far more suitable for measuring single cell [Ca2+]cyt dynamics in guard cells.
6 Localization signals have been used to direct cameleon expression to various subcellular compartments in animal cells (Miyawaki et al. 1997). Although subcellular calcium measurements have been achieved in plant cells using aequorin (Johnson et al. 1995; Knight et al. 1996), these can only be measured at the whole plant or organ level and are not ratiometric. The high fluorescent intensity of cameleon combined with improved imaging techniques should allow calcium dynamics to be measured on a subcellular level in individual cells. In addition, fusing cameleon to signalling proteins will allow calcium dynamics to be measured in microdomains surrounding particular protein types important in signal transduction. For example, this approach should allow calcium to be measured in the immediate molecular vicinity of ion channels, pumps or enzymes. To enable this long-term approach, we have engineered multiple cloning sites in the p35S-YC2.1 expression vector upstream and downstream of the YC2.1 coding sequence to enable C and N terminal cameleon tagging of plant proteins (see Fig. 6).
7 Although points 1–6 demonstrate the practical advantages of using cameleon calcium indicators in plants, the major advance presented in this study is the ability to readily study [Ca2+]cyt dynamics in single cells of Arabidopsis. The future combination of Ca2+ imaging together with cell biology, molecular genetics, genomics and physiology made possible by this fundamental advance will provide a powerful tool for the molecular dissection of signal transduction pathways in plants.
Standard PCR reactions and subcloning procedures were used to engineer the cameleon constructs described below. PCR was performed using PanVera X-Taq (PanVera Corporation, Madison, Wisconsin, USA). All PCR-derived sequences were sequenced to ensure the absence of PCR mistakes.
p35S-YC2.1-Kan (p35S-YC2.1-JFH6mp10) and p35S–YC3.1-Kan (p35S–YC3.1-JFH5mp102) express derivatives of cameleons YC2.1 and YC3.1, respectively. These constructs are based on a vector derived from pBIN19 (accession number U09365; Frisch et al. 1995) modified to include a CaMV 35S promoter from a pRT-derived vector (Topfer et al. 1987) and a translational enhancer derived from a tobacco etch virus (sequence from 20–155 shown in Harper et al. 1998). Cameleon sequences were PCR-amplified from their original vectors (Miyawaki et al. 1997). The 5′ primer was CAG GAA TTC CTC GAG GGC GCG CCC CTA GGT GGT GCG GCC GCC ACC ATG GTG AGC, and the 3′ primer was CAT TCC CGG GTC GAG ACC CTT GTA CAG CTC GTC CAT GCC GAG. A six-histidine affinity tag was added to the C-terminal end of both cameleons. Restriction enzyme sites were added to both ends to allow translational fusions with other proteins. The constructs above contain a kanamycin resistance marker for selection in Escherichia coli, Agrobacterium and plants. An identical set of cameleons was engineered into another pBIN19-derived vector that was modified to include a BAR gene to provide Basta resistance instead of kanamycin resistance in plants (Becker et al. 1992). These constructs are called p35S-YC2.1-BAR (p35S-YC2.1-JFH4-mp3) and p35S-YC3.1-BAR (p35S-YC3.1-JFH3-mp1).
Transgenic plants were generated by a vacuum infiltration protocol (Bechtold et al. 1993) using Agrobacterium tumefaciens strain GV3101. Infiltration was performed a few days after clipping primary bolts. Dry seeds were harvested, sterilized in 20% bleach for 20 min, plated on Gamborg's B5 media containing kanamycin (50 mg l–1) and carbenicillin (300 mg l–1). Kanamycin-resistant plants (T1) were identified and grown for seed. Experiments were conducted with T1 or T2 plants.
Arabidopsis thaliana (ecotype Landsberg erecta) seedlings expressing YC2.1 were grown in soil (Redi-Earth Peat-Lite Mix; Scotts, Marysville, Ohio, USA) in a controlled environment growth chamber (Conviron model E15; Controlled Environments, Asheville, North Carolina, USA) under a 16 h light/8 h dark cycle, a photon fluency rate of 100 μmol m–2 sec–1 and at a temperature of 20°C. Pots were watered every 2–3 days with deionized water and plants were misted with deionized water daily to keep the humidity close to 70%.
Confocal and deconvolution microscopy
The cytoplasmic localization of YC2.1 expressed in guard cells was assessed by scanning confocal laser microscopy in epidermal fragments prepared by blending rosette leaves as described for stomatal aperture measurements (see below). Epidermal fragments were mounted between two cover slips on a Zeiss Axiovert 35M microscope (with a 63×, 1.4 n.a., oil immersion lens) coupled to a Bio-Rad MCR-1000 scanning laser confocal system. Argon laser light (488 nm wavelength, 30% power) was used to excite the EYFP of cameleon (the brighter of the two mutant GFPs) and emission light was measured at 522 nm (DF32). Transmission images were collected in parallel.
To assess cytoplasmic localization and autofluorescence at both emission wavelengths of cameleon, YC2.1 was visualized in epidermal fragments using high-resolution deconvolution microscopy. Fragments were mounted on an Olympus IX-70 inverted microscope with a 100× (1.3 n.a.) oil immersion objective. Excitation light was provided by a mercury arc lamp and a 436 nm (DF10) filter. Emission of the ECFP was measured at 473 nm (DF30) and of the EYFP at 535 nm (DF30) using a cooled CCD camera (Photometrics, Tuscon, Arizona, USA). Images were collected through an entire stomatal complex in 60 0.2 μm optical sections (z-axis). Excitation light was delivered for 10 sec for each wavelength in each section giving a 20 min total imaging time. The long time course of exposure necessary for deconvolution microscopy prevented time-dependent ratio measurements on this high-resolution system. Following acquisition, images were deconvolved using Delta Vision 2.1 software (Applied Precision Instruments, Issaquah, Washington State, USA) and three-dimensional images were reconstructed. Representative images from particular three-dimensional projections were converted into Adobe Photoshop files and are presented in Fig. 1.
Epidermal strip preparation and fluorescence ratio measurements
To measure dynamic [Ca2+]cyt changes in guard cells, epidermal peels of Arabidopsis rosette leaves were used. To prepare peels, leaves were detached and the adaxial surface was gently pressed onto a piece of double-sided sided sticky tape that was mounted on the bench top. A small epidermal strip (typically 3 × 5 mm in size) was removed from the abaxial surface using needle-nosed tweezers and immediately mounted on an ultra-thin (0.083 mm) glass cover slip that had been coated with an very thin film of high vacuum grease (Dow Corning, Midland, Michigan, USA) to prevent strip movement. This cover slip formed the bottom of a microwell chamber (Mat Tek Corp., Ashland, Massachusetts, USA) which was immediately filled with 4 mL of buffer containing 5 mm KCl, 50 μm CaCl2, 10 mm MES-Tris pH 6.2. To promote stomatal opening, the chamber was illuminated for 2 h (photon fluency rate of 100 μmol m–2 sec–1) at 22°C before measurements commenced. Ratiometric [Ca2+]cyt measurements were made only on stomata that maintained equal turgor in both guard cells, had large apertures and intact organelle structure.
The microwell chamber was mounted on a Zeiss Axiovert microscope equipped for imaging with a cooled CCD camera (Photometrics, Tuscon, Arizona, USA). Excitation light was provided by a Xenon lamp and a 440 nm (DF20) filter. Excitation light intensity was reduced by a 10% transmission neutral density filter and directed through a 63× (1.4 n.a.) Nikon oil immersion objective by a 455DRLP dichroic mirror. Ratiometric measurements were made by interchanging 480 nm (DF30) and 535 nm (DF25) emission filters in front of the CCD camera. Emission filter changes and shutter control for periodic excitation were controlled by a Lambda 10-2 filter changer (Sutter Instruments, San Rafael, California, USA) and ratiometric images were collected using Metafluor 2.75 software (Universal Imaging, West Chester, Pennsylvania, USA). Excitation light was provided for 2 sec for each emission wavelength measured (4 sec total illumination), and this was repeated on an 8 or 10 sec cycle. Ratios were calculated online from the average pixel intensity for each emission wavelength in regions corresponding to each of the two cells of the stomatal complex. Images were only permanently recorded at specific events using image capture options in the software. Background fluorescence from the mounting grease and cover slip was subtracted from each wavelength online. It was not possible to account for autofluorescence in cells expressing YC2.1 as the cell could not be observed without a combined emission signal from YC2.1 and autofluorescence. Autofluorescence was accounted for in the YC2.1 calibration (see below). In order to gain a satisfactory signal-to-noise ratio in most experiments, pixels were maximized (16 pixels per bin) to increase fluorescent intensity, although this limited spatial image resolution (compare Fig. 2b at 4 pixels per bin and Fig. 3b at 16 pixels per bin). Cells were monitored at 1 min intervals for 10–15 min before recordings to check for spontaneous [Ca2+]cyt changes. ABA was only added to cells which exhibited a stable resting ratio.
Calibration of YC2.1 ratio changes
Minimum and maximum ratio values were determined in vitro as ionophores failed to consistently raise the [Ca2+]cyt as commonly observed for plant cells. Serial dilutions of purified YC2.1 protein were made in a solution of 80 mm K-gluconate, 20 mm KCl, 10 mm MES-Tris pH 7.2 (to mimic the cytoplasm) containing either 1 mm CaCl2 or 100 μm EGTA. Drops of 2 μl were placed on a cover slip and the correct focal depth found by focusing on the edge of the drop. The YC2.1 dilution which gave similar 535 and 480 nm intensities to a YC2.1-expressing guard cell was used to determine Rmin and Rmax. To account for autofluorescence (particularly at 480 nm, Fig. 1j), 535 nm and 480 nm signals were measured from 20 non-expressing guard cells from three separate plants. The average fluorescence was determined for each emission wavelength for each cell; the average values were 32 ± 3 for 535 nm and 51 ± 3 for 480 nm. The low errors indicate that autofluorescence was relatively consistent between cells. These values were then added back to the signals from the purified YC2.1 protein during calibration (after background subtraction from the cover slip) to mimic the autofluorescence from a cell. The values of Rmin and Rmax were used to convert YC2.1 ratios into [Ca2+]cyt by fitting to the YC2.1 in vitro calibration curve (Figs 2B in Miyawaki et al. 1999).
Stomatal aperture bioassays
To measure calcium-induced stomatal closing, rosette leaves from 4–6-week-old plants were detached from the plants and floated in opening solution consisting of 5 mm KCl, 50 μm CaCl2 10 mm MES-Tris pH 6.15 for 2 h in the light (photon fluency rate of 100 μmol m–2 sec–1). After 2 h, either ABA (10 or 100 μm) or CaCl2 (1 or 10 mm) were added to the buffer. After a further 2 h, the leaves were blended in 400 ml of deionized water in a Waring Blender for 20 sec. The resulting epidermal fragments were filtered out with a 30 μm nylon mesh, placed on a microscope slide and covered with a cover slip. Aperture ratios (width/height) were measured as previously described (Pei et al. 1997).
Thanks to Malcolm Wood (Scripps Research Institute) for assistance with deconvolution microscopy. The research was supported by National Science Foundation MCB-9506191 (J.I.S.), an NSF-REU supplement and Department of Energy grants DE-FG03-94-ER20148 (J.I.S.) and DE-FG03-94ER20152 (J.F.H.), a joint grant from NASA and the National Science Foundation (IBN-9416038; J.F.H.) and by Human Frontiers Science Program fellowships (G.J.A and J.M.K).