Background

Changes in the intracellular Ca²⁺ concentration ([Ca²⁺]) regulate many intra- and inter-cellular responses (reviewed in Lu and Means, 1993; Berridge et al., 2000; Orrenius et al., 2003; Zampese and Pizzo, 2012). In resting conditions, cells are able to maintain a large [Ca²⁺] gradient between the cytosol ([Ca²⁺]_c~100 nM) and the extracellular medium ([Ca²⁺]e~1.5-2 mM) thanks to the activity of a complex toolkit of antiporters and pumps (reviewed in Zampese and Pizzo, 2012). Upon different stimuli that trigger the release of Ca²⁺ from the intracellular stores and/or the entry of Ca²⁺ from the extracellular space, [Ca²⁺]_c can rapidly increase up to 1-3 µM. This [Ca²⁺]_c increase initiates and modulates a variety of cellular processes that operate over a wide time range, from the microsecond scale of neurotransmitter release, to the min and hours time scale of gene transcription (reviewed in Berridge et al., 2003; Clapham, 2007). This signal versatility is achieved by a sophisticated cellular machinery, composed by a variety of channels, pumps, exchangers and Ca²⁺ binding proteins localized in all cell subcompartments, that allow a precise regulation of the amplitude, duration and localization of Ca²⁺ signals (reviewed in Berridge et al.,2000; Giorgi et al., 2018).

A balance between the ‘on’ reactions that allow Ca²⁺ influx into the cytosol (from internal stores or from the extracellular milieu) and the ‘off’ reactions that remove Ca²⁺ from the cytosol (reviewed in Berridge et al., 2003), determines the dynamics of the Ca²⁺ signals and the specificity of the cellular response. In this contribution, we focus on changes in [Ca²⁺]_c or in the mitochondrial matrix ([Ca²⁺]_m) caused by either the release of Ca²⁺ from endo/sarcoplasmic reticulum (ER/SR), or by the influx of Ca²⁺ across the plasma membrane (PM) through the so called “store-operated Ca²⁺ entry” (SOCE) pathway (for a recent review, see Putney, 2017). The ER/SR is the main intracellular Ca²⁺ store, where the luminal [Ca²⁺] ([Ca²⁺]_ER) can reach values as high as 0.2-2 mM. The molecular actors that allow the maintenance of such a high [Ca²⁺]_ER are: (i) the Sarco/Endoplasmic Reticulum Ca²⁺ ATPases (SERCAs), responsible for Ca²⁺ accumulation within the lumen (for a recent review, see Primeau et al., 2018); (ii) the low-affinity high-capacity Ca²⁺-buffering proteins, responsible for luminal Ca²⁺ storage (reviewed in Zhou et al., 2013). Ca²⁺ release from the ER/SR depends on the presence in the organelle membranes of the ubiquitously expressed inositol 1,4,5-trisphosphate (IP₃) receptors (IP3Rs; for a recent review, see Prole and Taylor, 2019) and the ryanodine receptors (RyRs) (mainly expressed in excitable cells; for a recent review, see Santulli et al., 2017). Noteworthy, this Ca²⁺ toolkit is shared by another important organelle, the Golgi Apparatus (GA) (Pinton et al.,1998; reviewed in Pizzo et al., 2011). Subcompartments of the GA (medial- and trans-Golgi) also express another Ca²⁺ pump, the Secretory pathway Ca²⁺-ATPase pump type 1 (SPCA1, Pizzo et al., 2011). Evidence has been provided in the last years suggesting that Ca²⁺ can be released also from other cell compartments (secretory granules and lysosomes) possibly though cADPR sensitive channels (for a recent review, see Morgan, 2016), although the details of this process are still debated. There is no doubt, however, that the majority of the Ca²⁺ released during cell activation derives from the ER/SR compartment.

Once Ca²⁺ is released from these stores, the cation diffuses in part from the cytosol into the nucleoplasm through the nuclear pores (Brini et al., 1993; Miyawaki et al., 1997; Bootman et al., 2000; Manjarrés et al., 2008), and in part it can be transiently sequestered into the mitochondrial matrix (Rizzuto et al., 1993; reviewed in Rizzuto and Pozzan, 2006; Filadi and Pozzan, 2015; Pendin et al., 2015; De Stefani et al., 2016). Mitochondria can take up Ca²⁺ thanks to their large membrane potential, negative inside, and the presence in the inner mitochondrial membrane (IMM) of the so-called Mitochondrial Calcium Uniporter (MCU) Complex (MCUC). MCU is the channel forming subunit that mediates Ca²⁺ uptake, while MCUC also contains a number of recently identified regulatory proteins, such as MCUb (an inhibitory subunit), MICU1, MICU2, MICU3 and EMRE (that modulate Ca²⁺ affinity and assembly of the complex) (for a recent review, see De Stefani et al., 2016; Pallafacchina et al., 2018). The efflux of Ca²⁺ from the matrix back to the cytosol, depends on the H⁺/Ca²⁺ (Jiang et al., 2009) and the Na⁺/Ca²⁺ exchanger (mNCX) (Palty et al., 2010). Once Ca²⁺ increases in the mitochondrial matrix, it modulates the activity of a series of enzymes involved in ATP production and, under conditions of massive accumulation of Ca²⁺ (the so-called “Ca²⁺ overload”), the release of caspase cofactors and thus the initiation of the apoptosis process, and in part it can be transiently sequestered into the mitochondrial matrix (for recent reviews, see De Stefani et al., 2016; Rossi et al., 2019).

The [Ca²⁺]_c increases are terminated by plasma membrane (PM) extrusion mechanisms, i.e., the PM Ca²⁺ ATPase (PMCA, reviewed in Stafford et al., 2017), the Na⁺/Ca²⁺ exchanger (NCX, DiPolo and Beaugé, 2006), and by Ca²⁺ re-accumulation in the lumen of intracellular stores.

A unique and intensely studied process of Ca²⁺ influx is the SOCE that is activated by the decrease of the [Ca²⁺]ER. The molecular mechanism of SOCE activation was clarified in the last decade and depends on: i) an ER protein called STIM1 that “senses” the [Ca²⁺]_ER and ii) a PM protein, called ORAI1, that forms Ca²⁺ channels. Upon store depletion, STIM1 changes its distribution in the ER membrane, from diffused to clustered “puncta” that interact with the PM-located ORAI1. The formation of STIM1-ORAI1 complexes activate the channels and thus the influx of Ca²⁺ into the cytosol (reviewed in Putney, 2018). No threshold of ER depletion are known to be necessary for SOCE activation (Hofer et al., 1998). Furthermore, the Ca²⁺ influx induced by SOCE is not selectively taken up by the ER, since it induces a Ca²⁺ transient in both the cytosol and the mitochondria (Giacomello et al., 2010). However, in the point of close proximity between ER and PM, the so-called ER and PM contact sites (reviewed in Saheki and De Camilli, 2017), the uptake of Ca²⁺ by the SERCA is favored.

Given the importance of Ca²⁺ signaling in the modulation of a variety of many different cellular functions (from muscle contraction to exocytosis, from enzyme activity activation to fertilization, from gene transcription to cell death), the development of methodologies to monitor the dynamics in [Ca²⁺] in living cells has been a major topic of interest for physiologists and pathologists. A milestone in the field has been the invention by Roger Y. Tsien and co-workers of a series of intracellularly trappable indicators (reviewed in Tsien, 1981; Tsien et al.,1982) that are still extensively used today by hundreds of laboratories. Today, the available probes have different spectral properties and Ca²⁺-affinity (K_d), but they were designed to stay in the cytosol and are difficult to selectively target to subcellular compartments. The development of Genetically Encoded Ca²⁺ Indicators (GECIs) revolutionized the field of Ca²⁺ signaling investigation, as they allowed the targeting (using specific targeting sequences) of these Ca²⁺ sensors to most cell compartments with very high selectivity. The first targeted GECIs were based on the Ca²⁺ sensitive photoprotein aequorin (Knight et al., 1991) that allowed to monitor the dynamics of [Ca²⁺] in mitochondria, the ER/SR lumen, the GA, the nucleus, under the plasma membrane or within secretory granules (reviewed in Rizzuto and Pozzan 2006, Pendin et al., 2015; Alonso et al., 2017). Chemiluminescence by aequorin requires a prosthetic group, coelenterazine, to function. Once Ca²⁺ binds to aequorin, coelenterazine is irreversibly oxidized to coelenteramide, emitting a photon. The rate of photon emission by aequorin is proportional to the [Ca²⁺]. The major limitations of aequorin are the low amount of photon emission and the consumption of the probe during the experiment, making this tool better suited for relatively short (e.g., tens of min) experiments in cell populations (for review see Pendin et al., 2015).

A further improvement in the methodology for [Ca²⁺] measurements in living cells was introduced in the late 90’s, when GECIs based on green fluorescent protein (GFP) were generated (Miyawaki et al., 1997; Romoser et al., 1997). A further advantage of GECIs is that their expression can be placed under the control of tissue-specific or inducible promoters, controlling their expression both spatially and temporally (for review see Pendin et al., 2017).

Mainly two types of GFP based GECIs are presently available: (i) single fluorophore-based GECIs, in which the binding domain is inserted in the sequence of the fluorescent protein (FP); (ii) Förster/fluorescence resonance energy transfer (FRET)-based GECIs, in which Ca²⁺ binding changes the FRET efficiency between two FPs. Over time, the development of different FPs (reviewed in Rodriguez et al., 2017) allowed the generation of GECIs with different colors, offering the opportunity of monitoring different subcellular compartments at the same time or of combining Ca²⁺ measurements with optogenetics (for review see Wietek and Prigge, 2016). The GECIs have been also extensively engineered to obtain sensors with different K_d, dynamic range (DR, the maximum change in the fluorescence emission or ratio due to Ca²⁺ binding), and kinetic properties.

In this contribution, we will focus on GECIs of the so-called “Cameleon” family (Figure2A) that belong to FRET-based GECIs (Palmer et al., 2006). They are composed by two FPs, a cyan FP (CFP), as donor, and a yellow FP (YFP or its variants citrine and cpV), as acceptor. The two FPs are connected by a Ca²⁺-sensing peptide, calmodulin (CaM) that can bind four Ca²⁺ ions, and the CaM-binding peptide of the myosin light chain kinase M13 (M13) (Miyawaki et al., 1997) that are in turn separated by a short glycine-glycine linker. When Ca²⁺ binds to CaM, a reversible conformational change occurs in CaM that wraps around M13; modifying the distance between the two FPs and thus FRET efficiency between the FPs (Figure 2A). In practical terms: i) the cells expressing the probe are illuminated with a light wavelength that excites the donor FP only; ii) the light emitted by the donor and acceptor FPs is monitored and the ratio between the donor and acceptor fluorescence intensities (emission ratio, R) is calculated; iii) an increase in [Ca²⁺] augments the energy transfer from the donor to the acceptor FP (the donor fluorescence decreases whereas that of the acceptor increases). Thus, changes in [Ca²⁺] can be conveniently monitored as changes in R. Cameleons are the first GFP- and FRET-based GECIs developed and are still among the most commonly used.

Over the last two decades, Cameleons have been widely used and their cDNA sequence has been extensively modified to target it to different organelles, to reduce interference with endogenous CaM, to ameliorate the DR, to tune the K_d, and to improve signal/noise ratio (SNR) (Truong et al., 2001; Nagai et al., 2004; Evanko and Haydon, 2005; Palmer et al., 2006; Horikawa et al., 2010; Greotti et al., 2019). A red-shifted mitochondrial targeted Cameleon has been also generated (Waldeck-Weiermair et al., 2012). Among the different types of Cameleons, the D1, D2, D3 and D4 variants are the more extensively used. They have a circularly permuted Venus (cpV) as acceptor that improves the ratiometric sensitivity (Nagai et al., 2004), CaM and M13 have been mutated to lower interference with the endogenous protein (Palmer et al., 2006). Recently, also the donor has been changed improving the SNR and lowering the pH sensitivity of these sensors (Greotti et al., 2019).

Generally, this ratiometric measurement has the enormous advantages of making FRET-based GECIs independent from changes in the focal plane, of heterogeneity in probe loading/expression, or leakage and of artifacts introduced by sample movements. Thus, the affinity for Ca²⁺ of Cameleons, as other FRET-based probes, can be more easily estimated not only in vitro, i.e., employing the purified protein, but also in situ, i.e., in living cells. Indeed, the in situ-K_d of Cameleons targeted to different subcellular compartments are available (Palmer and Tsien, 2006; Drago et al., 2008; Giacomello et al., 2010; Lissandron et al., 2010; Wong et al., 2013; Greotti et al., 2016 and 2019), allowing the conversion of R into [Ca²⁺]. Other variants of Cameleons are TN-L15, TN-humTnC (Heim and Griesbeck, 2004; Mank et al., 2006) and Twitch (Thestrup et al., 2014), in which the Ca²⁺ responsive element (CaM) has been replaced with chicken skeletal muscle Troponin C (TnC), human cardiac TnC or Opsanus TnC, respectively. Generally, FRET-based sensors are recommended for in vivo experiments or for experimental models that undergo substantial movement during measurements (e.g., cardiomyocytes).

As mentioned, there is another class of GECIs based on a single FP. These sensors are best suited for multiple-organelle imaging and they generally have a larger DR, compared to FRET-based GECIs. Different types of single fluorophore GECIs have been created in the last years. The first generated single FP probes are Camgaroos (Griesbeck et al., 2001) and the Pericam family (flash Pericam, inverse Pericam, and ratiometric Pericam, mitochondria-targeted Pericam; Nagai et al., 2001: Filippin et al., 2005). Nowadays, the most popular among single FP-based sensors are the GCaMPs family and organelle-entrapped protein indicators (CEPIA). These sensors are obtained fusing M13 and CaM to the N- and C-terminus respectively of a circularly permuted EGFP (Nakai et al., 2001). Also for these sensors, many variants with different K_ds and DR are available and some of them have been targeted to ER (Wu et al., 2014; de Juan-Sanz et al., 2017; Henderson et al., 2015) and mitochondria (Patron et al., 2014; Wu et al., 2014; Mammucari et al., 2015). A large DR characterizes CEPIA and GCaMPs, though they are strongly pH sensitive and GCaMPs have low fluorescence at resting Ca²⁺ levels in vivo. CEPIA sensors are especially useful for simultaneous measurement of [Ca²⁺] in mitochondria, ER and cytosolic compartments, since sensors with very different emission wavelengths (and targeted to different organelles) have been generated. Recently, also the CaMP family have been engineered to obtain multicolored XCaMPs (Inoue et al., 2019). Of note, GCaMP6 (Chen et al., 2013) and ratiometric pericams can be used in the ratiometric mode exploiting the isosbestic point of their excitation spectra (Nagai et al., 2001; Filippin et al., 2005; Hill et al., 2014; Patron et al., 2019). Unfortunately, this characteristic is not practically usable in confocal microscopes (single or multiphoton). A FRET-based variant of CEPIA, called GEM-CEPIA, has been also recently generated (Suzuki et al., 2014).

A recent interesting innovation in GECIs is the generation of probes with photoactivable (PAFPs) or photoswitchable FPs (PSFPs). These FPs change their spectral properties in response to a specific wavelength irradiation (reviewed in Rodriguez et al., 2017). PSFPs and PAFPs allow the activation/conversion of a small subset of molecules, providing a tool to perform super-resolved Ca²⁺ imaging. Examples of these sensors are a FRET-based GECIs, named PA-TNXL (Matsuda et al., 2013) and a “Ca²⁺ integrator” named “CaMPARI”. The latter is a non-conventional GECI, since it undergoes an irreversible green to red conversion when UV light is irradiated in the presence of an elevated [Ca²⁺]_c, allowing the identification of cells that experienced increase in [Ca²⁺]_c in vivo (Fosque et al., 2015; Moeyaert et al., 2018).

Last but not least, recently, two new sensors based on Bioluminescence Resonance Energy Transfer (BRET) have been developed. BRET-based Ca²⁺ sensors consist of a bioluminescent molecule that transfer its excited state due to Ca²⁺ binding to the acceptor (typically a YFP) when the two moieties are sufficiently close to allow the resonant transfer of energy. Compared to FRET, BRET sensors do not require donor excitation by an external light source. These probes contain a chimeric GFP-aequorin protein (GAP) (Rodriguez-Garcia et al., 2014) linked to an optimized luciferase reporter, namely Nano-lantern (Suzuki et al., 2016). Although both photoswitchable/photoactivable FP probes and the BRET-based GECI have quite interesting characteristics, their use is still rather limited.

In this contribution, we focus on new Cameleons targeted to mitochondria and the cytosol with improved characteristics (i.e., increased SNR and decreased pH sensitivity, thanks to the donor substitution), and we describe a few typical protocols in which these probes are used to monitor Ca²⁺ influx due to SOCE or the release of Ca²⁺ from ER through IP₃R/RyR stimulation.