Red fluorescent CEPIA indicators for visualization of Ca²⁺ dynamics in mitochondria

Abstract

Mitochondrial Ca²⁺ dynamics are involved in the regulation of multifarious cellular processes, including intracellular Ca²⁺ signalling, cell metabolism and cell death. Use of mitochondria-targeted genetically encoded Ca²⁺ indicators has revealed intercellular and subcellular heterogeneity of mitochondrial Ca²⁺ dynamics, which are assumed to be determined by distinct thresholds of Ca²⁺ increases at each subcellular mitochondrial domain. The balance between Ca²⁺ influx through the mitochondrial calcium uniporter and extrusion by cation exchangers across the inner mitochondrial membrane may define the threshold; however, the precise mechanisms remain to be further explored. We here report the new red fluorescent genetically encoded Ca²⁺ indicators, R-CEPIA3mt and R-CEPIA4mt, which are targeted to mitochondria and their Ca²⁺ affinities are engineered to match the intramitochondrial Ca²⁺ concentrations. They enable visualization of mitochondrial Ca²⁺ dynamics with high spatiotemporal resolution in parallel with the use of green fluorescent probes and optogenetic tools. Thus, R-CEPIA3mt and R-CEPIA4mt are expected to be a useful tool for elucidating the mechanisms of the complex mitochondrial Ca²⁺ dynamics and their functions.

Introduction

Mitochondrial Ca²⁺ dynamics contribute to the control of various cellular functions such as formation of spatiotemporal patterns of cytosolic Ca²⁺ dynamics, cellular metabolism and cell survival¹. Ca²⁺ concentrations in the mitochondrial matrix are regulated by the balance between the influx of Ca²⁺ through the mitochondrial Ca²⁺ uniporter (MCU) and the efflux of Ca²⁺ by Na⁺/Ca²⁺ or H⁺/Ca²⁺ exchangers^1,2,3. Recent studies have elucidated the molecular identity of the channel and regulatory components of MCU^{4,5,6,7,8,9,10,11} as well as Na⁺-Ca²⁺-Li⁺ exchanger in the mitochondrial inner membrane^12,13. Furthermore, the advent of genetically encoded Ca²⁺ indicators (GECIs) that are targeted to the mitochondrial matrix has enabled monitoring mitochondrial Ca²⁺ dynamics with high spatiotemporal resolution, revealing both the subcellular and intercellular heterogeneity of mitochondrial Ca²⁺ responses upon agonist-induced increases in the cytosolic Ca²⁺ concentration^14,15,16. These imaging results suggest that the threshold of the net Ca²⁺ flux into the mitochondrial matrix is differentially determined in individual cells or even in each subcellular mitochondrial domain. However, the mechanism of these heterogeneous mitochondrial Ca²⁺ dynamics and their functional significance remains to be clarified. Thus, further analyses combined with high-resolution mitochondrial Ca²⁺ imaging are required.

We have previously developed a Ca²⁺ indicator protein family of Calcium-measuring organelle-Entrapped Protein IndicAtors (CEPIA) to visualize Ca²⁺ signals in both the endoplasmic reticulum (ER) and mitochondria¹⁵. ER-targeted CEPIAs have K_d values for Ca²⁺ ranging between 558 and 672 µM, which are higher than those of other ER-targeted GECIs such as D1ER¹⁷, GCaMPer¹⁸, ER-GCaMPs¹⁹, and ER-LAR-GECOs²⁰. Mitochondrial Ca²⁺ imaging analyses using green fluorescent protein (GFP)-based CEPIA variants with lower Ca²⁺ affinities (K_d = 14.5 or 90.2 µM) than cytosolic Ca²⁺ indicators suggested that mitochondrial Ca²⁺ concentrations can increase beyond 50 µM in a small fraction of HeLa cells. While a red fluorescent protein (RFP)-based low-affinity GECI, mito-LAR-GECO1.2 (K_d = 12 µM)²⁰, has been developed, lower-affinity mitochondrial RFP-based CEPIAs have not been developed, yet. Moreover, GFP-based GECIs cannot be used simultaneously with other green fluorescent imaging tools, including synthetic Ca²⁺ indicators, nor are they simultaneously used with optogenetic tools that are activated by blue light, such as channelrhodopsin-2 and OptoXRs^21,22. These limitations can be circumvented by GECIs with longer excitation and emission wave lengths. Therefore, CEPIA variants with red fluorescence may allow simultaneous use of other optical tools to increase the utility of mitochondria-targeted GECIs.

To this end, we generated the red-fluorescent CEPIAmt variants, R-CEPIA3mt and RCEPIA4mt, of which the Ca²⁺-affinity was optimized to measure mitochondrial Ca²⁺ concentrations. These variants allow visualization of mitochondrial Ca²⁺ signals with high spatiotemporal resolution that enables the detection of mitochondrial Ca²⁺ dynamics at subcellular local domains. Furthermore, simultaneous use of green fluorescent Ca²⁺ indicators and optogenetic tools is possible. Thus, R-CEPIA3mt and R-CEPIA4mt are expected to be a valuable tool for obtaining deeper insight into the cellular functions of mitochondrial Ca²⁺ dynamics.

Results

In vitro characterization of R-CEPIA3mt and R-CEPIA4mt

On the basis of an amino acid substitution strategy to produce low Ca²⁺ affinity variants of CEPIA¹⁵, we generated R-CEPIA3mt and R-CEPIA4mt by modifying one (E31D) and three (E31D, F92W and D133E) amino acids, respectively, in the calmodulin domain of R-GECO1mt (Supplementary Fig. 1A). As expected, these mutant indicators had reduced Ca²⁺ affinities (K_d = 3.7 µM for R-CEPIA3mt; K_d = 26.9 µM for R-CEPIA4mt) and high dynamic ranges (Fig. 1A and Table 1) without apparent alterations in the extinction coefficient, both excitation and emission spectra, and pH dependence of the original R-GECO1mt (Fig. 1B,C and Table 1). Notably, R-CEPIA3mt had a broader dynamic range with reduced cooperativity compared with those of R-GECO1mt and R-CEPIA4mt, indicating that R-CEPIA3mt may be useful for detecting dynamic changes in mitochondrial Ca²⁺ levels ranging from 0.5 to 100 µM.

Figure 1

In vitro characterization of R-CEPIA3mt and R-CEPIA4mt. All the extracted parameters are summarized in Table 1. (A) In vitro Ca²⁺ titration curves of R-GECO1mt (gray), R-CEPIA3mt (red) and R-CEPIA4mt (brown) at pH 7.2 (left) or 8.0 (right) solution. (B) Absorption (upper), excitation and emission (lower) spectra of R-GECO1mt (left), R-CEPIA3mt (middle) and R-CEPIA4mt (right) in Ca²⁺-containing (1 mM, magenta) or Ca²⁺-free (1 mM EGTA, black) solution. (C) pH titration curves of R-GECO1mt (left), R-CEPIA3mt (middle) and R-CEPIA4mt (right) in Ca²⁺-containing (1 mM, magenta) or Ca²⁺-free (1 mM EGTA, black) solution. The plots were fitted by a single Hill equation. Mean ± SEM (n = 3).

Table 1 Properties of R-CEPIA3mt and R-CEPIA4mt.

Mitochondrial Ca²⁺ signals visualized by R-CEPIA3mt and R-CEPIA4mt

Using these CEPIA variants, we visualized mitochondrial Ca²⁺ signals in HeLa cells. The mitochondrial distribution of both variants was confirmed by colocalization with MitoTracker Green (Fig. 2A). Simultaneous Ca²⁺ imaging in mitochondria and the cytosol (the latter was visualized by the green fluorescent Ca²⁺ indicator, Cal-520) demonstrated that only a fraction of the cells (35.7%, n = 34 for R-CEPIA3mt; 28.6%, n = 38 for R-CEPIA4mt) showed a transient mitochondrial Ca²⁺ increase in response to cytosolic Ca²⁺ elevations induced by the inositol trisphosphate-producing agonist, histamine (Fig. 2B,C). Interestingly, mitochondrial Ca²⁺ transients were elicited only by the initial peak of the cytosolic Ca²⁺ oscillations, which reached the threshold of mitochondrial Ca²⁺ increases. Similar observations have previously been reported using GFP-based mitochondrial CEPIAs¹⁵.

Figure 2

Global mitochondrial Ca²⁺ signals visualized by R-CEPIA3mt and R-CEPIA4mt. (A) Representative images of HeLa cells expressing R-CEPIA3mt (upper) or R-CEPIA4mt (lower). Fluorescence of R-CEPIAs (left), MitoTracker Green staining (middle) and the merged images (right) are shown. Scale bars, 5 µm. (B) Time course of agonist-induced Ca²⁺ response in the mitochondria (upper) and cytosol (lower) in three representative HeLa cells expressing R-CEPIA3mt (left) or R-CEPIA4mt (right). Cytosolic Ca²⁺ signals were visualized by a green fluorescent synthetic Ca²⁺ indicator, Cal-520. (C) Heat maps of cell population data (n = 14) of global Ca²⁺ signals in mitochondria (upper) and cytosol (lower). Each horizontal strip corresponds to the time course of the Ca²⁺ signal in each cell.

To examine the spatial resolution of R-CEPIA3mt and R-CEPIA4mt, we next focused on subcellular mitochondrial domains in single HeLa cells. Both R-CEPIAmt variants successfully detected heterogenous Ca²⁺ signals in mitochondrial subdomains within close proximity (Fig. 3A,B). We next performed simultaneous imaging of mitochondrial Ca²⁺ using two CEPIAmt variants with different colours and Ca²⁺ affinities to effectively expand the range of Ca²⁺ concentrations detectable with the indicators. HeLa cells were co-transfected with R-CEPIA3mt or R-CEPIA4mt and the GFP-based high Ca²⁺ affinity CEPIA2mt (K_d = 160 nM)¹⁵. Additionally, the cells were loaded with fura-2 for simultaneous detection of cytosolic Ca²⁺ dynamics. As shown in Fig. 3C,D, we found that individual HeLa cells had heterogenous mitochondrial domains, of which Ca²⁺ signals were detected either by both CEPIA2mt and one of the R-CEPIAmt variants (domain 1 in Fig. 3C,D) or only by CEPIA2mt (domain 2 in Fig. 3C,D). It is remarkable that agonist-induced mitochondrial Ca²⁺ dynamics were detected by indicators with varying affinities to Ca²⁺. On the basis of the calibrations of the indicators (Fig. 1A), these results suggest that each mitochondrial subdomain reaches distinct Ca²⁺ levels that range between 0.1 and 100 µM.

Figure 3

Local mitochondrial Ca²⁺ signals visualized by R-CEPIA3mt and R-CEPIA4mt. (A,B) Representative red fluorescence images and time courses of agonist-induced Ca²⁺ signals in subcellular mitochondrial domains in a HeLa cell expressing R-CEPIA3mt (A) or R-CEPIA4mt (B). Regions of interest (ROI) are indicated in high magnification images shown in the left lower panels. Scale bars, 10 µm (upper) and 1 µm (lower). (C,D) Simultaneous Ca²⁺ imaging of mitochondria with low Ca²⁺ affinity R-CEPIAs (red fluorescence), high Ca²⁺ affinity CEPIA2mt (green fluorescence) and the cytosolic Ca²⁺ indicator, fura-2 (405-nm excitation). Representative red fluorescence images and time courses of agonist-induced Ca²⁺ signals in subcellular mitochondrial domains in a HeLa cell are shown. ROI are indicated in high magnification images shown in the left lower panels. The cells expressing both R-CEPIA3mt and CEPIA2mt (C), and both R-CEPIA4mt and CEPIA2mt (D) were loaded with fura-2. Scale bars, 10 µm (upper) and 1 µm (lower).

pH dependence of R-CEPIA3mt and R-CEPIA4mt

Using HeLa cells co-expressing CEPIA2mt and one of the new R-CEPIAmt variants, we examined the pH dependence of R-CEPIA3mt and R-CEPIA4mt. Although alkalization by NH₄Cl induced a sustained elevation in the fluorescence intensity of all mitochondrial CEPIAs, the amplitude of the change in R-CEPIA3mt and R-CEPIA4mt was significantly smaller than that of CEPIA2mt (Supplementary Fig. 2). This is attributable to the acidity shifted acid dissociation constant of R-CEPIA3mt and R-CEPIA4mt (Table 1; pK_a = 6.5 in the presence of Ca²⁺ for both the R-CEPIAmt variants) compared with the GFP-based CEPIAs (pK_a = 7.0 in the presence of Ca²⁺ for CEPIA2mt, Supplementary Fig. 1B)¹⁵. Thus, R-CEPIA3mt and R-CEPIA4mt enable a more stable detection of Ca²⁺ signals in mitochondria, which may undergo dynamic pH changes.

Simultaneous use of an optogenetic tool and mitochondrial Ca²⁺ indicators

As the excitation and emission spectra of the R-CEPIAmt variants do not overlap with the excitation spectrum (peaking at ~450 nm) of the phospholipase C (PLC) activator, Opto-α₁AR-YFP²², we performed mitochondrial Ca²⁺ imaging during optogenetic activation. In response to 448-nm laser irradiation, several mitochondrial domains (positions 1, 2 and 3 in Fig. 4A) in a HeLa cell expressing Opto-α₁AR-YFP and R-CEPIA3mt showed Ca²⁺ increases (Fig. 4B), whereas another mitochondrial domain in the same cell (position 4 in Fig. 4A) did not show Ca²⁺ signals. The ensemble averaging of these local responses represents the global mitochondrial Ca²⁺ signal (bottom solid line in Fig. 4B). Another cell without Opto-α₁AR-YFP expression in the same dish failed to show mitochondrial Ca²⁺ signals in response to light activation (“Cont” in Fig. 4B). Bath application of the mitochondrial oxidative phosphorylation uncoupler, carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), rapidly decreased the fluorescence intensity of R-CEPIA3mt in all mitochondrial domains, indicating that R-CEPIA3mt reported Ca²⁺ increases in mitochondria induced by optogenetic activation of Ca²⁺ release from the ER. Averaged traces show that the global mitochondrial Ca²⁺ responses were detected by R-CEPIA3mt but not by R-CEPIA4mt (Fig. 4C), suggesting that optogenetic PLC activation may induce mitochondrial Ca²⁺ increases up to 1–2 µM at most in our experimental conditions.

Figure 4

Simultaneous application of mitochondrial Ca²⁺ imaging and an optogenetic activator of the Ca²⁺ release machinery. (A) Representative confocal section images of Opto-α₁AR-YFP and R-CEPIA3mt in a transfected HeLa cell. A merged image is shown in the lower panel. Scale bar, 10 µm. (B) Time courses of R-CEPIA3mt-reported Ca²⁺ signals in the subcellular mitochondrial domains indicated in (A) (Local; ROI 1–4). Laser irradiation (448 nm) during the period indicated by the blue box was used to activate Opto-α₁AR-YFP, which is a chimera protein of rhodopsin, α₁-adrenergic receptor and YFP. Traces in the right box show Ca²⁺ responses during bath application of the mitochondrial uncoupler, FCCP. Time courses of global mitochondrial Ca²⁺ signals in the same cell and in a negative control cell expressing R-CEPIA3mt (without expressing Opto-α₁AR-YFP) in the same culture dish are shown at the bottom (Global and Cont). (C) Averaged time courses of optogenetics-induced global mitochondrial Ca²⁺ signals that were reported by R-CEPIA3mt (red) or R-CEPIA4mt (brown). As a control, the averaged response of HeLa cells transfected with only R-CEPIA3mt is shown (grey dashed line). n = 6, 3 and 3 cells for R-CEPIA3mt, R-CEPIA4mt and control, respectively, from 2 culture dishes each.

Discussion

We succeeded to produce two variants of RFP-based mitochondrial CEPIAs, R-CEPIA3mt and R-CEPIA4mt. These new indicators have lower Ca²⁺ sensitivity than that of the previously developed high Ca²⁺ affinity RFP-based mitochondrial GECI, R-GECO1mt (Table 1)^15,23. The K_d values for Ca²⁺ of R-CEPIA3mt and R-CEPIA4mt are similar to those of CEPIA3mt and CEPIA4mt, respectively, which were developed in our previous study¹⁵. Owing to the large dynamic range and high brightness, R-CEPIA3mt and R-CEPIA4mt allow spatiotemporal resolution imaging that is high enough to visualize heterogenous Ca²⁺ dynamics in subcellular mitochondrial domains (Fig. 3). They can be used simultaneously with other tools, such as GFP-based CEPIAs, Cal-520 and fura-2 (Fig. 3) as well as optogenetic PLC activator (Fig. 4). Furthermore, they have a lower pH dependence than the GFP-based CEPIAmt variants (Supplementary Fig. 2).

We have previously shown that there is no significant pH change within the mitochondrial matrix in histamine-stimulated HeLa cells¹⁵. Thus, the mitochondrial Ca²⁺ changes reported by R-CEPIAmts in the present study are unlikely to be influenced by mitochondrial pH changes. It has been reported that under certain conditions mitochondria may undergo transient alkalizations called “mitochondrial flash” (estimated pH range: 7.5 to 8.0)^24,25. The effect of such pH changes would be less in R-CEPIA3mt and R-CEPIA4mt than in GFP-based CEPIAmts.

The mechanism underlying the intercellular and subcellular heterogeneity in mitochondrial Ca²⁺ dynamics is of great interest. Recent studies have identified molecules involved in Ca²⁺ influx through the MCU complex, including MCU, MCUb, EMRE and MICU1–3^{4,5,6,7,8,9,10,11}. Furthermore, the Na⁺-Ca²⁺-Li⁺ transporter has been identified as a Ca²⁺ extrusion mechanism on the mitochondrial inner membrane^12,13. Therefore, it is important to study whether there is intercellular and subcellular heterogeneity in the expression levels or activities of these Ca²⁺-handling proteins, and whether protein heterogeneity corresponds to the heterogeneity in mitochondrial Ca²⁺ dynamics. Moreover, it will be informative to study whether the heterogeneity in mitochondrial Ca²⁺ dynamics produces intercellular heterogeneity in cell death as well as subcellular heterogeneity in mitochondrial motility, ATP production, opening of permeability transition pores, and mitochondrial membrane potentials. Various optical tools for tagging, probing and controlling these cellular processes have been developed. Simultaneous use of these optical tools and the R-CEPIAmt variants is expected to make a great contribution to the field.

In summary, we added new members to the library of green and red emission mitochondrial Ca²⁺ indicators that cover a broad range of mitochondrial Ca²⁺ concentrations. Further imaging analyses using these potent and useful tools may facilitate uncovering the mode of action of mitochondrial Ca²⁺ dynamics and their functions in health and disease.

Methods

Gene manipulation and plasmid construction

For engineering R-CEPIA3 and R-CEPIA4, we introduced point mutations into pET19b R-GECO1¹⁵ using primers 1–6 (Supplementary Table 1). For mammalian expression, the cDNAs of R-CEPIA3 and R-CEPIA4 were subcloned into pCMV vector containing the mitochondria-targeting sequence (pCMV R-GECO1mt¹⁵) by restriction enzyme digestion.

Bacterial expression and in vitro spectroscopy

BL21-CodonPlus(DE3)-RIL bacteria (Agilent, Santa Clara, CA, USA) were transformed with the pET19b plasmids having R-CEPIA3, R-CEPIA4, R-GECO1 or CEPIA2. The bacteria were incubated for 16–24 h at 37 °C in 2 × YT medium containing ampicillin and chloramphenicol (20 μg•ml^–1). After harvesting the bacteria by centrifugation, the cells were resuspended in KCl/MOPS buffer (130 KCl, 50 MOPS in mM, pH 7.2) and lysed by French press (Thermo Fisher, Waltham, MA, USA) at 20,000 psi. The recombinant proteins were purified using TALON metal affinity resin (Takara Clontech, Shiga, Japan) and eluted with KCl/MOPS buffer containing 250 mM imidazole.

The absorbance spectra were measured with a NanoDrop 2000 spectrophotometer (Thermo Fisher) in KCl/MOPS buffer containing 1 mM EGTA or 1 mM CaCl₂. The protein concentration was calculated by measuring the absorbance following alkaline denaturation, assuming ε = 38,000 M^–1•cm^–1 at 455 nm for R-CEPIA3, R-CEPIA4, and R-GECO1, and ε = 44,000 M^–1•cm^–1 at 446 nm for CEPIA2²⁶. The molar extinction coefficient was calculated by dividing the peak absorbance value by the protein concentration.

Ca²⁺ titration curves were obtained by adding small aliquots of CaCl₂ to the recombinant indicators in KCl/MOPS or KCl/HEPES (130 KCl, 50 HEPES in mM, pH 8.0) buffer. The indicators’ concentration was 25–75 nM. The Ca²⁺ concentrations were clamped with any of the following Ca²⁺ buffers: 1 mM EGTA, 1 mM BAPTA, 1 mM Br₂BAPTA (5,5′-Dibromo BAPTA) and 1 mM Nitrilotriacetic acid (NTA). Free Ca²⁺ concentration was calculated by Maxchelator (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/). Fluorescence intensity was measured with F-4500 FL spectrofluorometer (Hitachi, Tokyo, Japan) at 562 ± 5/584 ± 5 nm (excitation/emission) wavelength for R-CEPIA3, R-CEPIA4, and R-GECO1, and 492 ± 10/514 ± 10 nm for CEPIA2. The obtained relationship between the Ca²⁺ concentration and the fluorescence intensity was fitted by the following single Hill plot equation with the KaleidaGraph software (Synergy Software, Reading, PA, USA).

$${F}={{F}}_{{\rm{m}}{\rm{i}}{\rm{n}}}+({{F}}_{{\rm{m}}{\rm{a}}{\rm{x}}}\,\mbox{--}\,{{F}}_{{\rm{m}}{\rm{i}}{\rm{n}}})\times {({[C{a}^{2+}]}_{{\rm{free}}})}^{{n}}/[{({[C{a}^{2+}]}_{{\rm{f}}{\rm{r}}{\rm{e}}{\rm{e}}})}^{{n}}+{({{K}{{\prime} }}_{{\rm{d}}})}^{{n}}].$$

K′_d represents apparent dissociation constant or the Ca²⁺ concentration at which half of the indicator molecules bind to Ca²⁺. n represents Hill coefficient. The fluorescence intensity at various Ca²⁺ concentrations was normalized by (F − F_min)/(F_max − F_min). The dynamic range of the indicator was calculated as the ratio of F_max to F_min.

pH titration curves for each indicator were obtained by measuring fluorescence intensity in the solutions containing 130 mM KCl and 50 mM pH buffer (MES for pH 5.0; MES/HEPES for pH 6.0; MOPS/HEPES for pH 7.0–7.5; HEPES for pH 8.0; HEPES/CHES for pH 9.0; CHES for pH 10.0) with 1 mM EGTA or 1 mM Ca²⁺. pKa was obtained by fitting the obtained fluorescence intensity with a single Hill equation.

Cell culture

HeLa cells were cultured on collagen-coated plastic dishes (IWAKI, Shizuoka, Japan) in DMEM supplemented with 10% fetal bovine serum, penicillin (100 U•ml^–1) and streptomycin (100 U•ml^–1). For Ca²⁺ imaging, the cells were plated on collagen type-I-coated glass-bottom dishes (MatTek, Ashland, MA, USA) and cultured for 16 h before imaging.

Imaging

Cultured HeLa cells were transfected using Lipofectamine 3000 (Thermo Fisher) 2 or 3 days before imaging. The plasmids used in the current study were: pCMV R-CEPIA3mt, pCMV R-CEPIA4mt, pCMV CEPIA2mt and pcDNA3 Opto-α₁AR-YFP. For cytosolic Ca²⁺ imaging using Cal-520 or fura-2, cells were loaded with 5 µM Cal-520 AM (AAT Bioquest, USA) or 5 µM fura-2 AM (Dojindo, Japan) at room temperature (22–24 °C) for 30 min in culture medium. Before imaging, the loading solution was washed three times and replaced with physiological salt solution (PSS) containing (in mM) 150 NaCl, 4 KCl, 2 CaCl₂, 1 MgCl₂, 5.6 glucose and 25 HEPES (pH 7.4).

Time-lapse images of Cal-520, R-CEPIA3mt and R-CEPIA4mt (hereafter, R-CEPIA3/4mt) were captured using an inverted IX81 microscope (Olympus, Tokyo, Japan) equipped with a ×20 UPlanSApo oil immersion objective [numerical aperture (NA) = 0.75; Olympus], an electron-multiplying cooled-coupled device (EM-CCD) ImagEM camera (Hamamatsu Photonics, Japan), a Lambda 10–3 filter wheel (Sutter Instrument, Navato, CA, USA), an ebx75 xenon lamp (Leica, Wetzlar, Germany) and an EL6000 metal halide lamp (Leica) at a rate of one frame per 1 or 2 s with the following set of excitation and emission filters, respectively: 472 ± 15 nm and 520 ± 17.5 nm for Cal-520; 562 ± 20 nm and 641 ± 37.5 nm for R-CEPIA3/4mt.

For Ca²⁺ imaging in subcellular mitochondrial domains, time lapse or snapshot images of R-CEPIA3/4mt, CEPIA2mt, fura-2 and Opto-α₁AR-YFP were captured using a TCS SP8 confocal microscope (Leica) equipped with a × 63 HC PL APO oil immersion objective (NA = 1.40; Leica) at a rate of one frame per 1–3 s with the following wavelengths [excitation laser (nm); emission spectra (nm)]: R-CEPIA3/4mt (552; 560–750), CEPIA2mt (488; 520–550), fura-2 (405; 430–515) and Opto-α₁AR-YFP [for confirmation of transfection by YFP fluorescence, shown in Fig. 4A (488; 500–550), for light activation (448; 530–555)]. Photobleaching was corrected for by a linear or exponential fit to the fluorescence intensity change before agonist stimulation.

For imaging of subcellular localization of CEPIA, the mitochondria in R-CEPIA3/4mt-expressing cells were stained by a 30-min incubation in culture medium containing 500 nM MitoTracker Green (Thermo Fisher) at 37 °C. Images were captured with a TCS SP8 confocal microscope using a ×63 HC PL APO oil immersion objective at excitation: 488 nm and emission: 500–540 nm for MitoTracker Green and excitation: 552 nm and emission: 560–750 nm for R-CEPIA3/4mt.

Data analysis and statistics

Two-tailed Student’s t-tests were performed to determine the statistical significance using Origin 7 (OriginLab, Northampton, MA, USA). The calculations, processing and analysis of obtained images were performed with ImageJ and ImageJ Fiji (NIH, Bethesda, MD, USA). Graphs and time course traces were produced with Origin 7 and arraigned with Illustrator CC (Adobe, San Jose, CA, USA), respectively.