Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Fluorescent indicators for Ca2+based on green fluorescent proteins and calmodulin


Important Ca2+ signals in the cytosol and organelles are often extremely localized and hard to measure. To overcome this problem we have constructed new fluorescent indicators for Ca2+ that are genetically encoded without cofactors and are targetable to specific intracellular locations. We have dubbed these fluorescent indicators ‘cameleons’. They consist of tandem fusions of a blue- or cyan-emitting mutant of the green fluorescent protein (GFP)1,2, calmodulin3,4,5, the calmodulin-binding peptide M13 (ref. 6), and an enhanced green- or yellow-emitting GFP7,8,9. Binding of Ca2+ makes calmodulin wrap around the M13 domain, increasing the fluorescence resonance energy transfer (FRET) between the flanking GFPs2. Calmodulin mutations can tune the Ca2+ affinities to measure free Ca2+ concentrations in the range 10−8 to 10−2 M. We have visualized free Ca2+ dynamics in the cytosol, nucleus and endoplasmic reticulum of single HeLa cells transfected with complementary DNAs encoding chimaeras bearing appropriate localization signals. Ca2+ concentration in the endoplasmic reticulum of individual cells ranged from 60 to 400 µM at rest, and 1 to 50 µM after Ca2+ mobilization. FRET is also an indicator of the reversible intermolecular association of cyan-GFP-labelled calmodulin with yellow-GFP-labelled M13. Thus FRET between GFP mutants can monitor localized Ca2+ signals and protein heterodimerization in individual live cells.


Cytosolic and organellar free Ca2+ concentrations are among the most important and dynamic intracellular signals, and are usually measured using synthetic fluorescent chelators10,11,12,13 or recombinant aequorin14,15. The chelators are easily imaged but are difficult to target precisely to specific intracellular locations, whereas aequorin is easily targeted but requires the incorporation of coelenterazine, is irreversibly consumed by Ca2+, and is very difficult to image because its luminescence produces less than one photon per molecule. To combine the brightness of fluorescent indicators with the targetability of a biosynthetic indicator, we have used GFP, a spontaneously fluorescent protein from the jellyfish Aequorea victoria. Its cDNA can be concatenated with those encoding many other proteins, and the resulting fusion proteins are usually fluorescent and often preserve the biochemical functions and cellular localizations of the partner proteins. Mutagenesis has produced GFP mutants with shifted wavelengths of excitation or emission1,2,7,9 that can serve as donors and acceptors for FRET. FRET is a non-destructive spectroscopic method that can monitor the proximity and relative angular orientation of fluorophores in living cells. The donor and acceptor fluorophores can be entirely separate or attached to the same macromolecule. In the latter case, ligand-induced conformational changes might be monitored by FRET if the amino and carboxy termini of the binding protein are fused to a donor and acceptor GFP. This approach has several advantages over the usual covalent labelling with fluorescent probes. First, the indicator is generated in situ by gene transfer into the cells or organisms, obviating large-scale expression, purification, labelling and microinjection of recombinant proteins that must be soluble. Second, the sites of fusions are defined exactly, giving a molecularly homogenous product without the use of elaborate protein chemistry. Third, the chromophore of GFP is fixed in the protein9. If the GFP donor and acceptor are rigidly fused to a host protein, minor changes in the relative orientation of the ends of the latter would alter FRET. In contrast, most conventional fluorescent labels are attached by flexible linkers that at least partly decouple the fluorophore orientation from that of the protein to which it is attached.

Many effects of Ca2+ in cells are mediated by the binding of Ca2+ to calmodulin, which causes calmodulin to bind and activate target proteins3. Based on the nuclear magnetic resonance solution structure of calmodulin bound to M13, the 26-residue calmodulin-binding peptide of myosin light-chain kinase6, the C terminus of calmodulin has been fused to the M13 by means of a Gly-Gly spacer16. Ca2+ binding switched the resulting hybrid protein (calmodulin–M13) from a dumb-bell-like extended form to a compact globular form similar to the calmodulin–M13 intermolecular complex. We therefore sandwiched the calmodulin–M13 fusion between a blue (Y66H/Y145F)2 and a green (S65T)7 GFP mutant to construct a Ca2+ indicator (Fig. 1a). The amino-acid sequences of the boundary regions between the calmodulin–M13 hybrid and GFPs proved critical to the optimization of protein folding and the Ca2+ sensitivity of FRET. Numerous deletions, insertions and amino-acid substitutions were tested; the best splice sequences found are shown in Fig. 1b. The prototype fusion protein was efficiently expressed and folded in bacteria and increased its ratio of ultraviolet-excited 510:445 nm emissions by 70% upon binding Ca2+(Fig. 2a). The decrease in blue and increase in green emission indicated that Ca2+ increased the efficiency of FRET from Y66H/Y145F to S65T, consistent with the expected decrease in distance between the two ends of the protein. The Ca2+ response was fully reversible upon chelation of Ca2+. We have called this calmodulin-based indicator ‘cameleon-1’, because it readily changes colour and retracts and extends a long tongue (M13) into and out of the mouth of the calmodulin, often abbreviated CaM. Mg2+, pH and ionic strength did not alter the emission spectra of either the Ca2+-saturated or Ca2+-unsaturated forms (data not shown), in accord with the typical high specificity of calmodulin for Ca2+(ref. 5). Cameleon-1 was not affected by hydrophobic proteins, such as bovine serum albumin. Isolated calmodulin saturated with Ca2+ becomes sticky with hydrophobic amino acids exposed to the surface. The calmodulin in cameleon-1, however, seems to interact preferentially with its intramolecularly adjacent M13 peptide. Such a self-contained system should minimize the possibility that the protein might interact with endogenous calmodulin-binding sequences in eukaryotic cells.

Figure 1: Schematic structures and sequences of cameleons.
figure 1

a, Scheme showing how FRET between GFPs can measure Ca2+. The GFPs are drawn as simple rigid cylinders, reflecting their crystal structures9. Schematic structures of calmodulin (CaM) without Ca2+, disordered unbound M13, and the Ca2+–calmodulin–M13 complex are derived from crystallography4, NMR6 and modelling16, but the relative orientations of the GFPs are unknown. b, Domain structure of cameleons expressed in bacteria for in vitro characterization, showing sequences of the boundaries between the donor GFP and Xenopus calmodulin (XCaM) and between M13 and the acceptor GFP. c, Consitition of cameleons for expression and imaging in mammalian cells. CRsig, calreticulin signal sequence, MLLSVPLLLGLLGLAAAD24; nls, nuclear localizaiton signal, PKKKRKVEDA; KDEL, ER retention signal24; kz, Kozak consensus sequence for optimal translational iniitation in mammalian cells28. The significant mutations in the four GFPs are: EBFP, F64L/Y66H/Y1452; EGFP, F64L/S65T8; ECFP, F64L/S65T/Y66W/N146I/M153T/V163A/N212K2; and EYFP, S65G/S72A/T203Y9.

Figure 2: Properties of cameleons in vitro.
figure 2

a, Emission spectra of cameleon-1 (excited at 380 nm) at the indicated free Ca2+ values at pH 7.47. b, Ca2+ titration curves of cameleon-1 (open circles), cameleon-1/E104Q (filled circles), and cameleon-1/E31Q (open triangles), combining 5 independent experiments using Ca2+/EGTA and Ca2+/HEEDTA systems10 below 10−5 M free Ca2+ and unbuffered Ca2+ above. The changes in emission ratio (510 to 445 nm) were normalized to the effects of full Ca2+ saturation, which were 60–80% increases in ratio over the values at zero Ca2+. The fitted curves correspond to the apparent dissociation constants and Hill coefficients given in the text. c, Relaxation rate constants, kobs (=kon[Ca2+] + koff), for reaction of cameleon-1/E104Q with Ca2+; kon and koff are the association and dissociation rate constants. d, Emission spectrum of yellow cameleon-2 in vitro, excited at 432 nm, at zero and saturating Ca2+.

The original calmodulin–M13 hybrid protein without GFPs displayed a biphasic Ca2+ binding with two dissociation constants, 2 µM and 80 nM, resulting from independent Ca2+ binding to the N- and C-terminal domains of calmodulin, respectively16. The emission ratio of cameleon-1 has a similar biphasic Ca2+ dependency (Fig. 2b, open circles) with apparent dissociation constants (Kd) of 11 µM and 70 nM, and Hill coefficients (n) of 1.0 and 1.8, respectively. Therefore cameleon-1 can report a very wide range of Ca2+ concentration, from <10−7 to >10−4 M. Many site-directed mutations have been studied for their effects on the Ca2+ binding and Ca2+-induced conformational changes of calmodulin17,18. Mutation of the conserved bidentate glutamic acid (E104) at position 12 of the third Ca2+-binding loop to glutamine (Q) eliminated the high-affinity component of the Ca2+ response (Fig. 2b, filled circles). The resulting monophasic response (Kd, 4.4 µM; n, 0.76) corresponded fairly closely to the low-affinity component of the cameleon-1 response. An analogous mutation in the first Ca2+-binding loop (E31Q) resulted in a large rightward shift of the low-affinity component of cameleon-1, without affecting the high-affinity component (Fig. 2b, triangles). Cameleon-1/E31Q showed a biphasic response (Kd, 83 nM and 700 µM; n, 1.5 and 0.87. Much more tuning of Ca2+ affinities for particular applications should be possible.

Because cameleon-1/E104Q had the simplest Ca2+-response curve, it was chosen for initial explorations of Ca2+-binding kinetics by stopped-flow measurements of the acceptor S65T emission after rapid mixing with Ca2+. The approach to equilibrium was reasonably exponential, and the inverse time constants were a linear function of free [Ca2+], as expected for simple 1 : 1 binding (Fig. 2c). The association and dissociation rate constants correspond tothe slope and intercept of the linear plot and were (2.4 ± 0.2) × 106 M−1 s−1 and 13 ± 9 s−1, in 100 mM KCl at 25 °C. Their ratio is 5.4 µM, which is consistent with the apparent dissociation constant (4.4 µM) independently derived from equilibrium measurements (Fig. 2b). The dissociation rate is similar to that (12 ± 2 s−1) for calmodulin and M13 without the GFPs19, so the GFPs do not perturb the dissociation even though they are much larger than the interacting calmodulin and M13 domains.

Can cameleons work as Ca2+ indicators in mammalian cells? Initial transfections with the gene encoding cameleon-1 in the pcDNA3 vector (with a cytomegalovirus promoter) gave inadequate fluorescence because the blue was not bright enough. The simplest remedy was to introduce mammalian codon bias and mutate Phe 64 to Leu, which improves expression and folding of GFPs at 37 °C without greatly affecting the spectra of the final well-folded molecules8,20. The resulting enhanced blue and green fluorescent proteins (‘EBFP’ and EGFP, respectively) fused to calmodulin–M13 constituted cameleon-2 (Fig. 1c), which indeed greatly improved mammalian cell expression. However, an even better approach for single-cell imaging proved to be replacement of the EBFP and EGFP by similarly enhanced cyan and yellow fluorescent mutants, ‘ECFP’ and ‘EYFP’, respectively (Fig. 1c). ECFP contains Trp at residue 66 in place of the His of EBFP and Tyr of EGFP, whereas EYFP contains the normal Tyr 66 in the fluorophore but is shifted to a yellowish emission by mutation of Thr 203 to Tyr (ref. 9). Both genes have mammalian codon usage and encode appropriate mutations to improve folding20. The substitution of ECFP for EBFP shifted the excitation peak from 381 to 433 nm and improved the brightness, signal-to-noise ratio and duration of recording, yet decreased potential concerns regarding autofluorescence, photodamage and incompatibility with caged compounds. The ‘yellow cameleons’ incorporating ECFP and EYFP have three minor disadvantages. The maximal ratio changes (Fig. 2d) are slightly smaller (1.5-fold) than for the original EBFP/EGFP cameleons (1.8-fold), perhaps because the emission spectrum of ECFP encroaches more on the EYFP emission. Yellow cameleons are perturbed by acidification below pH 7.0, which mimics a fall in Ca2+. Also, EYFP can undergo reversible photochromism if illuminated too strongly (A.Miyawaki and J. Llopis, unpublished observations). Other properties, such as Ca2+ affinities, as well as the qualitative behaviour inside cells, seem to be the same for the two families of cameleons.

When yellow cameleon-2 was transfected into HeLa cells, the fluorescence was uniformly distributed in the cytosolic compartment but excluded from the nucleus (Fig. 3a), as expected for a 74 kDa protein without targeting signals. The time course of the spatially averaged yellow : cyan emission ratios from a single HeLa cell expressing yellow cameleon-2 is shown in Fig. 3b. A submaximal dose of histamine (0.5 µM) produced a significant increase in the emission ratio indicating an initial peak of about 3 µM cytosolic free Ca2+, [Ca2+]c. This response was gradually converted into oscillations and eventually desensitized altogether. However, ATP added as a purinergic agonist could still stimulate another increase in [Ca2+]c of similar amplitude to the initial histamine response. These responses agree well with the known behaviour of HeLa cells to histamine and ATP21,22. Application of the Ca2+ionophore ionomycin followed by a high concentration (20 mM) of extracellular Ca2+ gave a large increase in the ratio, which should correspond to the maximal ratio Rmax. To establish the minimal ratio Rmin, as required for a simple in situ calibration, the cell was loaded with the permeant chelator ester BAPTA/AM, and external Ca2+ was clamped to zero with EGTA in the presence of ionomycin. The ratio Rmax/Rmin inside cells was about 1.6, equal or slightly greater than that observed in vitro. This correspondence argues that the indicator in cells behaved in reasonable accord with its in vitro calibration, and that the calibration protocol effectively achieved saturating and zero [Ca2+]c.

Figure 3: Imaging of cameleons in HeLa cells.
figure 3

a, Fluorescence image (440 ± 10 nm excitation, 535 ± 12.5 nm emission) of yellow cameleon-2 shows cytosolic localizaiton in HeLa cells. Scale bar, 10 µm in all fluorescence images. b, Ratios of 535 ± 12.5 nm to 480 ± 15 nm emissions from yellow cameleon-2 in the cytoplasm of the leftmost cell of a, monitored every 7 s by digital imaging microscopy. The ratios were calibrated in terms of absolute pCa on the right-hand ordinate axis, using Rmin (arrowhead) and Rmax (arrow) determined in situ. ‘Ca2+ free’ indicates medium containing 0.5 mM MgCl2. c, d, Matched phase-contrast and fluorescence (480 ± 15 nm excitation, 535 ± 22.5 nm emission) images of two HeLa cells showing cameleon-2nu localized to the nuclei. e, Fluorescence image of yellow cameleon-3er, taken as in a. The same reticular pattern of fluorescence was observed for yellow cameleon-4er. f, Immunogold localizaiton of cameleon-3er in ultrathin cryosections using a polyclonal antibody against GFP. The cameleon-3er was abundant in the ER (er), nuclear envelope (ne and arrows) and intermediate compartment (arrowheads), but absent from the cis-most Golgi cisternae (Gc), mitochondria (m) and nucleus (n). ‘Cis’ and ‘trans’ show the orientation of the Golgi stack but do not mark specific structures. Scale bars, 0.1 µm. g, The 535 to 480 nm emission ratios from yellow cameleon-3er in two cells. The right-hand ordinate gives Rmin (arrowhead), Rmax (arrows) and pCa calibrations for each cell. h, Analogous emission ratios from yellow cameleon-4er in two cells. The right-hand ordinate calibrates [Ca2+]er.

Cameleons should be readily directable to interesting sites by fusion to appropriate organellar targeting signals or localized host proteins. For example, addition of a nuclear localization signal to cameleon-2 yielded a Ca2+ indicator, ‘cameleon-2nu’ (Fig. 1c), fluorescence of which was tightly localized to nuclei (Fig. 3c, d) but excluded from nucleoli. The time course of nuclear [Ca2+] proved generally similar to those of [Ca2+]c, confirming previous comparisons of nuclear and cytosolic aequorin23. We also succeeded in monitoring agonist-induced changes in the free Ca2+ inside the endoplasmic reticulum ([Ca2+]er) in individual, intact, non-perfused cells. Two low-affinity indicators, yellow cameleon-3er and 4er (for endoplasmic reticulum) (Fig. 1c), whose Ca2+ responses should correspond to the filled circles and open triangles respectively of Fig. 2b, were engineered to reside in the lumen of endoplasmic reticulum (ER) by addition of a signal sequence at the N terminus and a KDEL signal for ER retention at the C terminus24. Reticular patterns of fluorescence were seen in HeLa cells expressing the protein (Fig. 3e). Targeting to the ER was verified by electron-microscopic immunogold localization (Fig. 3f), which showed that almost all of the GFP immunoreactivity was within the ER and nuclear envelope. A few percent of gold particles were in the intermediate compartment between ER and Golgi, but the Golgi apparatus itself was essentially devoid of immunoreactivity. [Ca2+]er reported by yellow cameleon-3er in two neighbouring cells is shown in Fig. 3g. In one cell (filled circles), the indicator was initially saturated ([Ca2+]er > 100 µM), as shown by an emission ratio matching Rmax and a time delay in response to the Ca2+-mobilizing effect of histamine. The agonist eventually depleted [Ca2+]er to about 30 µM, within the dynamic range of the indicator. In the other cell (open circles), [Ca2+]er started at about 60 µM and responded to histamine with much less time delay, soon levelling off at 1.5 µM. The depletion of [Ca2+]er was promptly reversed by a histamine antagonist, cyproheptadine, and recapitulated by ATP. Responses to thapsigargin blockage of the ERCa2+-ATPase were also heterogeneous, whereas ionomycin immediately clamped [Ca2+]er almost to zero. To measure resting [Ca2+]er values above 100 µM, we turned to the lowest-affinity indicator, yellow cameleon-4er (Figs 2band 3h). This variant was clearly not saturated, because the ratio started well below Rmax, fell immediately upon histamine stimulation, but was delayed in recovering after antagonist application. Combined results from the two indicators separately applied to 75 cells showed [Ca2+]er ranging from 60 to 400 µM before and 1 to 50 µM after application of a maximal dose of agonists. Our estimates for basal [Ca2+]er fall roughly in the middle of previous estimates, which range from 1 to 2,000 µM (refs 11, 14, 15, 25). Previous measurements using small-molecule indicators are subject to severe uncertainties in dye localization and Mg2+ interference. Aequorin experiments can only be done on cell populations and are complicated by the opposing abilities of Ca2+ both to stimulate light output and to irreversibly destroy the indicator. ER-targeted cameleons avoid both sets of problems. Ca2+ stores in the ER have recently been suggested to contain spatially and functionally distinct compartments13,26. Cameleons fused to different ER sequences should be able to examine such heterogeneity with the unique combination of precise molecular targeting, single-cell physiological imaging, and ultrastructural localization.

Is FRET between GFPs limited to detecting intramolecular proximity, or can it also detect intermolecular associations? Yellow cameleon-2 was expressed as two separate pieces: ECFP–calmodulin and M13–EYFP (Fig. 1c). When these two proteins were mixed in vitro, the spectral difference between zero and saturating Ca2+ was similar to but somewhat larger than that for the corresponding intramolecular association in the fully fused construct (data not shown), presumably because the unfused partners could get much further apart from each other in the absence of Ca2+. When expressed in HeLa cells, ECFP–calmodulin and M13–EYFP were distributed throughout both the cytoplasm and the nucleus, as would be expected from their molecular weights (Mr 44K and 30K, respectively). This pattern is quite different from the exclusion of intact yellow cameleon-2 from the nucleus. Histamine mobilization of Ca2+ elevated the emission ratio of EYFP to ECFP in both cytoplasm and nucleus, an effect reversed by a histamine antagonist (Fig. 4, circles). ATP caused a similarly reversible increase in the ratio. Such physiological [Ca2+]c transients caused reciprocal changes in the EYFP and ECFP intensities (Fig. 4, lines without symbols), as expected for association/dissociation effects on FRET, whereas artefacts such as a change of focus caused parallel changes that were cancelled by calculating the ratios. Thus GFP-based FRET can monitor the dynamics of reversible heterodimerizaiton in live cells. In a recent alternative approach27, blue- and green-emitting GFPs were linked through the calmodulin-binding domain of smooth-muscle myosin light-chain kinase. FRET between the GFPs was disrupted in vitro upon binding of unlabelled (Ca2+)4-calmodulin to the linker. The fusion protein was introduced into cells only by microinjection, and showed small fluorescence response to [Ca2+]c that were amplifiable by co-injection of exogenous calmodulin.

Figure 4: GFP-based detection of protein association/dissociation.
figure 4

Top, 535 to 480 nm emission ratios (filled and open circles) in two adjacent cells co-transfected with M13–EYFP and ECFP–calmodulin. Bottom, 535 (broken line, left-hand scale) and 480 nm (solid line, right-hand scale) emissions, background-corrected, from the cell corresponding to the filled circles.

Many further improvements in cameleons should be possible by mutagenesis, whereas small-molecule indicators can only be optimised by more laborious chemical resynthesis. The targetability of cameleons may eventually allow Ca2+ measurements at previously inaccessible sites, such as the immediate vicinity of synaptic vesicles or Ca2+ channels. Cameleons should be well suited to genetically tractable organisms such as bacteria, yeast, nematodes, flies, plants and transgenic mice. Fluorescent indicators for analytes other than Ca2+ might be engineered by sandwiching other conformationally responsive receptors between GFP donors and acceptors. Fluorescence readout of intermolecular associations by FRET (as in Fig. 4) is non-destructive, quantifiable with high spatiotemporal resolution, and probably applicable to any compartment of the cell and to many proteins other than calmodulin and M13. However, the ratio of intensities in the acceptor and donor channels starts from a non-zero baseline even in the absence of FRET. Therefore, rare associations will be hard to detect, and positive and negative controls are crucial, such as the ability to switch the putative association on and off, ideally in situ. The dynamic range of such measurements both of Ca2+ and of protein association will benefit from further improvements in the GFP mutants to reduce leakage of the donor emission into the acceptor emission band.


Gene construction. The cDNA of the GFP blue mutant Y66H/Y145F (ref. 1) was amplified by the polymerase chain reaction (PCR) with a sense primer containing a BamHI site and a reverse primer containing an SphI site and eliminating the GFP stop codon. Similarly the cDNA of S65T was amplified with a SacI site and an EcoRI site introduced to the 5′ and 3′ ends of the gene, respectively. Two restriction sites (SphI and SacI) were introduced by PCR into the 5′ and 3′ ends of the calmodulin–M13 gene, respectively, using pHY1 (ref. 16) as a template. The restricted products were ligated and cloned in-frame into the BamHI/EcoRI sites of pRSETB (Invitrogen). The modifications of the boundary regions between Y66H/Y145F and calmodulin and between M13 and S65T were performed by PCR or by a combined use of restriction enzymes, Klenow fragment of DNA polymerase I, T4 DNA polymerase, mung-bean exonuclease, and T4 DNA ligase. Enhanced GFP mutants EBFP, ECFP and EYFP were created by introducing appropriate substitutions2,9 into the commercially available phEGFP plasmid from Clontech, which has mammalian codon usage20 and mutations F64L and S65T. Oligonucleotide-directed mutageneses were performed using the Muta-Gene Phagemid in vitro kit (Bio-Rad). The 5′ end of the EBFP or ECFP gene was modified by PCR to have a HindIII restriction site followed by a Kozak consensus sequence28 (ACCGCC-ATG). The HindIII/EcoRI fragment encoding the entire chimaeric protein was subcloned in the mammalian expression vector pcDNA3 (Invitrogen). For cameleon-2nu, the cameleon-2 cDNA was extended by PCR at the 3′ end with the sequence encoding the nuclear localization signal (PKKKRKVEDA). Yellow cameleon-3er and -4er cDNA were obtained by extending yellow cameleon-3 cDNA at the 5′ end with the sequence encoding the signal peptide from calreticulin (MLLSVPLLLGLLGLAAAD), and at the 3′ end with the sequence encoding the ER retention signal (KDEL)24. The cDNAs coding for ECFP–calmodulin and M13–EYFP were obtained from yellow cameleon-2 by PCR and cloned into pRSETB behind a polyhistidine tag for bacterial expression and into pcDNA3 behind a Kozak sequence for mammalian expression.

Protein expression, in vitro spectroscopy, Ca2+ titrations, and reaction kinetics. Chimaeric proteins were expressed in Escherichia coli, purified and spectroscopically characterized as previously described2. Ca2+ titrations were performed by reciprocal dilution of Ca2+-free and Ca2+-saturated buffers10. In situ calibration for [Ca2+] used the equation29 [Ca2+] = Kd [(RRmin)/ (RmaxR)](1/n), where Kd is the apparent dissociation constant corresponding to the Ca2+ concentration at which R is midway between Rmax and Rmin, and n is the Hill coefficient. Fast kinetic experiments were performed in a stopped-flow spectrofluorometer (Applied Photophysics). After rapid mixing of purified cameleon-1/E104Q (0.8 µM final concentration) and Ca2+ of various concentrations in 20 mM MOPS, 100 mM KCl, pH 7.2, at 25 °C, the fluorescence was monitored using an excitation wavelength of 380 nm with an emission cut-on filter of 500 nm. The observed first-order constant (kobs) was calculated from each averaged set of data by nonlinear regression analysis.

Electron microscopy and immunolabelling. Immunogold labelling of ultrathin sections was performed as described previously30. The polyclonal antibody against GFP was a gift from C. Zuker (University of California, SanDiego).

Imaging. Between 2 and 5 days after cDNA transfection with lipofectin (Gibco BRL), HeLa cells at 22 °C were imaged on a Zeiss Axiovert microscope with a cooled CCD camera (Photometrics, Tucson, AZ), controlled by MetaFluor 2.75 software (Universal Imaging, West Chester, PA). Dual-emission ratio imaging of yellow cameleons used a 440DF20 excitation filter, a 455DRLP dichroic mirror, and two emission filters (480DF30 for ECFP, 535DF25 for EYFP) alternated by a filter changer (Lambda 10-2, Sutter Instruments, San Rafael, CA). Interference filters were obtained from Omega Optical and Chroma Technologies (Brattleboro, VT).


  1. Heim, R., Prasher, D. C. & Tsien, R. Y. Wavelength mutations and post-translational autooxidation of green fluorescent protein. Proc. Natl Acad. Sci. USA 91, 12501–12504 (1994).

    Google Scholar 

  2. Heim, R. & Tsien, R. Y. Engineering green fluorescent protein for improved brightness, longer wavelengths and fluorescence energy transfer. Curr. Biol. 6, 178–182 (1996).

    Google Scholar 

  3. Crivici, A. & Ikura, M. Molecular and structural basis of target recognition by calmodulin. Annu. Rev. Biophys. Biomol. Struct. 24, 85–116 (1995).

    Google Scholar 

  4. Babu, Y. S., Bugg, C. E. & Cook, W. J. Structure of calmodulin refined at 2.2 Å resolution. J. Mol. Biol. 204, 191–204 (1988).

    Google Scholar 

  5. Falke, J. J., Drake, S. K., Hazard, A. L. & Peersen, O. B. Molecular tuning of ion binding to calcium signaling proteins. Q. Rev. Biophys. 27, 219–290 (1994).

    Google Scholar 

  6. Ikura, al. Solution structure of a calmodulin-target peptide complex by multidimensional NMR. Science 256, 632–638 (1992).

    Google Scholar 

  7. Heim, R., Cubitt, A. B. & Tsien, R. Y. Improved green fluorescence. Nature 373, 663–664 (1995).

    ADS  CAS  Article  Google Scholar 

  8. Cormack, B. P., Valdivia, R. H. & Falkow, S. FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173, 33–38 (1996).

    Google Scholar 

  9. Ormö, al. Crystal structure of the Aequorea victoria green fluorescent protein. Science 273, 1392–1395 (1996).

    Google Scholar 

  10. Grynkiewicz, G., Poenie, . & Tsien, R. Y. Anew generation of Ca2+ indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450 (1985).

    Google Scholar 

  11. Tse, F. W., Tse, A. & Hille, B. Cyclic Ca2+ changes in intracellular stores of gonadotropes during gonadotropin-releasing hormone-stimulated Ca2+ oscillations. Proc. Natl Acad. Sci. USA 91, 9750–9754 (1994).

    Google Scholar 

  12. Hofer, A. M. & Schulz, I. Quantification of intraluminal free [Ca] in the agonist-sensitive internal calcium store using compartmentalized fluorescent indicators: some considerations. Cell Calcium 20, 235–242 (1996).

    Google Scholar 

  13. Golovina, V. A. & Blaustein, M. P. Spatially and functionally distinct Ca2+ stores in sarcoplasmic and endoplasmic reticulum. Science 275, 1643–1648 (1997).

    Google Scholar 

  14. Montero, al. Monitoring dynamic changes in free Ca2+ concentration in the endoplasmic reticulum of intact cells. EMBO J. 14, 5467–5475 (1995).

    Google Scholar 

  15. Kendall, J. M., Badminton, M. N., Sala-Newby, G. B., Campbell, A. K. & Rembold, C. M. Recombinant apoaequorin acting as a pseudo-luciferase reports micromolar changes in the endoplasmic reticulum free Ca2+ of intact cells. Biochem. J. 318, 383–387 (1996).

    Google Scholar 

  16. Porumb, T., Yau, P., Harvey, T. S. & Ikura, M. Acalmodulin-target peptide hybrid molecule with unique calcium-binding properties. Protein Eng. 7, 109–115 (1994).

    Google Scholar 

  17. Maune, J. F., Klee, C. B. & Beckingham, K. Ca2+ binding and conformational change in two series of point mutations to the individual Ca2+-binding sites of calmodulin. J. Biol. Chem. 267, 5286–5296 (1992).

    Google Scholar 

  18. Gao, Z. al. Activation of four enzymes by two series of calmodulin mutants with point mutations in individual Ca2+ binding sites. J. Biol. Chem. 268, 20096–20104 (1993).

    Google Scholar 

  19. Martin, S. al. Spectroscopic characterization of a high-affinity calmodulin-target peptide hybrid molecule. Biochemistry 35, 3508–3517 (1996).

    Google Scholar 

  20. Zolotukhin, S., Potter, M., Hauswirth, W., Guy, J. & Muzyczka, N. A“humanized” green fluorescent protein cDNA adapted for high levels of expression in mammalian cells. J. Virol. 70, 4646–4654 (1996).

    Google Scholar 

  21. Smit, M. al. Extracellular ATP elevates cytoplasmatic free Ca2+ in HeLa cells by the interaction with a 5′-nucleotide receptor. Eur. J. Pharmacol. 247, 223–226 (1993).

    Google Scholar 

  22. Bootman, M. D., Cheek, T. R., Moreton, R. B., Bennett, D. L. & Berridge, M. J. Smoothly graded Ca2+ release from inositol 1,4,5-trisphosphate-sensitive Ca2+ stores. J. Biol. Chem. 269, 24783–24791 (1994).

    Google Scholar 

  23. Brini, M., Marsault, R., Bastianutto, C., Pozzan, T. & Rizzuto, R. Nuclear targeting of aequorin. A new approach for measuring nuclear Ca2+ concentration in intact cells. Cell Calcium 16, 259–268 (1994).

    Google Scholar 

  24. Kendall, J. M., Dormer, R. L. & Campbell, A. K. Targeting aequorin to the endoplasmic reticulum of living cells. Biochem. Biophys. Res. Commun. 189, 1008–1016 (1992).

    Google Scholar 

  25. Bygrave, F. L. & Benedetti, A. What is the concentration of calcium ions in the endoplasmic reticulum? Cell Calcium 19, 547–551 (1996).

    Google Scholar 

  26. Button, D. & Eidsath, A. Aequorin targeted to the endoplasmic reticulum reveals heterogeneity in luminal Ca++ concentration and reports agonist- or InsP3-induced release of Ca++. Mol. Biol. Cell 7, 419–434 (1996).

    Google Scholar 

  27. Romoser, V. A., Hinkle, P. M. & Persechini, A. Detection in living cells of Ca2+-dependent changes in the fluorescence emission of an indicator composed of two green fluorescent protein variants linked by a calmodulin-binding sequence. J. Biol. Chem. 272, 13270–13274 (1997).

    Google Scholar 

  28. Kozak, M. The scanning model for translation: an update. J. Cell Biol. 108, 229–241 (1989).

    Google Scholar 

  29. Adams, S. R., Bacskai, B. J., Taylor, S. S. & Tsien, R. Y. Fluorescent Probes for Biological Activity of Living Cells–A Practical Guide(ed. Mason, W. T.) 133–149 (Academic, New York, (1993)).

    Google Scholar 

  30. Balch, W. E., McCaffery, J. M., Plutner, H. & Farquhar, M. G. Vesicular stomatitis virus glycoprotein is sorted and concentrated during export from the endoplasmic reticulum. Cell 75, 841–852 (1994).

    Google Scholar 

Download references


We thank S. R. Adams and A. B. Cubitt for advice, C. Zuker for the GFP antibody, and C. Klee for the original Xenopus calmodulin gene. This work was supported by HHMI (R.Y.T.), NIH (R.Y.T.), HFSP (R.Y.T. and A.M.), MRCC (M.I.) and the Spanish Ministry of Science (J.L.). M.I. is an HHMI international research scholar and MRCC scholar.

Author information

Authors and Affiliations


Rights and permissions

Reprints and Permissions

About this article

Cite this article

Miyawaki, A., Llopis, J., Heim, R. et al. Fluorescent indicators for Ca2+based on green fluorescent proteins and calmodulin. Nature 388, 882–887 (1997).

Download citation

  • Received:

  • Accepted:

  • Issue Date:

  • DOI:

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing