Illuminating the hidden world of calcium ions in plants with a universe of indicators

Abstract

The tools available to carry out in vivo analysis of Ca²⁺ dynamics in plants are powerful and mature technologies that still require the proper controls.

Introduction

Calcium (Ca²⁺) is a well-known second messenger in both unicellular and multicellular organisms (Berridge et al., 2000; Carafoli and Krebs, 2016). In plants, apart from its role as a structural component (White and Broadley, 2003), calcium plays a role in signaling events in response to a multitude of developmental and environmental stimuli (Kudla et al., 2010; Edel et al., 2017; Kudla et al., 2018). Biotic and abiotic challenges affect the cellular Ca²⁺ homeostasis by triggering transient changes of Ca²⁺ concentrations in the cytosol as well as in subcellular compartments (McAinsh and Pittman, 2009; Stael et al., 2012; Costa et al., 2018; Pirayesh et al., 2021; Resentini et al., 2021b).

The basis of calcium’s role as a signaling component lies in its peculiar chemistry and the existence of a large electrochemical gradient across the cell’s membranes, maintained by the activity of the proton- and calcium-ATPases (H⁺-ATPases and Ca²⁺-ATPases; Palmgren, 2001; Demidchik et al., 2018; Klejchova et al., 2021). The main evolutionary reason for evolving mechanisms that generate and maintain this large gradient is based on the need to keep the cytosolic Ca²⁺ concentrations ([Ca²⁺]_cyt) low, to prevent the precipitation of organic and inorganic molecules (e.g. phosphates including adenosine triphosphate (ATP)) (Clapham, 2007). Under resting conditions, the [Ca²⁺]_cyt is in the range of hundreds of nanomolar (100–200 nM) whereas in the external spaces and subcellular compartments it can reach up to millimolar ranges (Stael et al., 2012). This steep concentration gradient implies that the opening of a limited number of calcium permeable channels, located on the different cellular membranes, is sufficient to rapidly increase the [Ca²⁺]_cyt with a 10-fold increase compared with the resting concentration (Demidchik et al., 2018).

Within the plant cell, the change in [Ca²⁺]_cyt is sensed by Ca²⁺ binding proteins acting as primary responders (e.g. calcium dependent kinases) or sensor relays. Among the latter Calmodulin (CaM), CaM-like proteins (CMLs) and calcineurin B-like proteins (CBLs) all bind cytosolic calcium, which triggers a conformational change of the proteins enabling them to interact with different targets modulating their activities (Kudla et al., 2018; Tian et al., 2020). When the Ca²⁺ sensors are stimulated they become primed to regulate downstream processes, which include ion fluxes, enzymatic activities, transcription, etc. (DeFalco et al., 2010; Kudla et al., 2018; Tian et al., 2020). Importantly, after the perception of a stimulus and the occurrence of the Ca²⁺ transient, resting [Ca²⁺]_cyt needs to be quickly reestablished to prevent cell death (Clapham, 2007). The molecular mechanisms that are responsible for the recovery of the resting [Ca²⁺]_cyt are Ca²⁺ buffers and Ca²⁺ active transporters, such as Ca²⁺-ATPases and Ca²⁺/cation exchangers (CAX) that are localized in the plasma membrane and membranes of intracellular compartments (Corso et al., 2018; Costa et al., 2018; Demidchik et al., 2018; Hilleary et al., 2020; Ishka et al., 2021; Resentini et al., 2021b).

The intertwined and coordinated activities of influx and efflux Ca²⁺ transport systems in the plasma membrane and internal stores (Stael et al., 2012; Costa et al., 2018; Pirayesh et al., 2021; Resentini et al., 2021b) jointly shape the characteristic cellular Ca²⁺ dynamics, also called Ca²⁺ signatures (Sanders et al., 2002).

The possibility of visualizing and studying Ca²⁺ signatures is based on the exploitation of “calcium imaging techniques” that, thanks to continuous improvement, have allowed us to study Ca²⁺ dynamics with increasing resolution and sensitivity with minimal invasiveness in different organisms, including plants. An important boost for the progression of the calcium imaging techniques is the possibility to extend the use of valuable and innovative tools across the different kingdoms of life. In such a scenario, one driver of innovation is neuroscience studies. In neurons, action potentials (APs) or the activation of ionotropic glutamate receptors trigger large and rapid changes in [Ca²⁺]_cyt (Tian et al., 2009). This has led neuroscientists to develop innovative technologies to study Ca²⁺ dynamics, which are historically based on the simultaneous and continuous improvement of Ca²⁺ indicators, and the development and the implementation of the appropriate microscopy instrumentation.

In this update, we will briefly retrace the history of Ca²⁺ sensitive indicators (Fig. 1) focusing on their use in plants, providing important clues about how to choose among them, and also taking into consideration the most appropriate imaging techniques. We will explain when it is recommended to use ratiometric biosensors and when instead it is reasonable and more convenient to use intensiometric ones. The intention of this update is to guide the readers into the recent developments in the area of Ca²⁺ biosensing in plants and make them enthusiastic about this amazing and imaginative field of research.

Figure 1.

History of major achievements for the in vivo study of Ca²⁺ biosensing in plants.

Synthetic dyes for measurements of cytosolic calcium dynamics

A breakthrough in the field of Ca²⁺ imaging was the development of sensitive fluorescent Ca²⁺ indicators (or dyes) and buffers by Tsien (1980). These indicators were the result of the “fusion” between Ca²⁺-selective chelators like ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA) or 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) with a fluorescent chromophore (Fig. 2).

Figure 2.

Examples of two commonly used Ca²⁺ chelators and calcium-sensitive dyes derived from them.

In plants, the first attempts to study Ca²⁺ dynamics in vivo were based on the use of Ca²⁺ sensitive dyes (e.g. Fura-2, Fura-2 dextran, calcium green dextran, and Indo-1) which were instrumental in measuring [Ca²⁺] in aleurone protoplasts, guard cells in response to ABA, in growing pollen tubes or in root hairs in response to nodulation factors (Bush and Jones, 1987; McAinsh et al., 1990; Gilroy et al., 1991; McAinsh et al., 1995; Ehrhardt et al., 1996; Holdaway-Clarke et al., 1997). The use of those dyes allowed the measurement, for the first time, of Ca²⁺ variation within isolated mitochondria from land plants, giving the basic knowledge of the molecular mechanism for Ca²⁺ uptake in these organelles (Zottini and Zannoni, 1993). Whereas the use of Ca²⁺-sensitive dyes allowed the making of fundamental discoveries, they have some limitations which primarily include (i) their requirement to be loaded or manually injected into the cells, possibly leading to an unequal amount of dye within different cells; (ii) when accumulated into the cells they can be compartmentalized or sequestered into vacuoles (Bush and Jones, 1990); and (iii) an excessive loading can affect cytosolic Ca²⁺ availability due to their Ca²⁺ chelator-backbone (Bush and Jones, 1990; Table 1). Nevertheless, a major advantage of Ca²⁺-sensitive dyes is the ability to perform Ca²⁺ imaging analyses without the need to generate transgenic plants (see the next section), as transformation protocols are not available for every species (Table 1). If a reliable Ca²⁺-sensitive dye for plant cells were to be available, it could be used for example to perform the first screen of mutants, instead of performing a tedious transgenic plants selection (Fichman and Mittler, 2021). Nevertheless, since the use of dyes has still more disadvantages than advantages, in the last 20 years, plant scientists have moved to the use of genetically encoded Ca²⁺ indicators (GECIs) (Perez Koldenkova and Nagai, 2013) which opened another dimension for the quantitative in vivo imaging of Ca²⁺ dynamics.

Table 1.

Summary of advantages and disadvantages/limitations of the different available Ca²⁺ indicators

Type of indicator	Advantages	Disadvantages/limitations
Synthetic dyes	(i) Ca2+ imaging analyses without the need to generate transgenic plants.	(i) To be loaded or manually injected into the cells; (ii) can be compartmentalized or sequestered into vacuoles; (iii) an excessive loading can bring to cytosolic Ca2+ buffering.
Aequorin	(i) Does not need to be excited with fluorescent light; (ii) long-term experiments; and (iii) it allows a good estimation of Ca2+ concentrations.	(i) Low quantum yield efficiency; (ii) requirement for the external provision of coelenterazine; and (iii) non-linearity of its rate of light emission.
Ratiometric GECIs (e.g. Cameleon, YC-Nano)	(i) Reliability; (ii) reduced artifacts; (iii) strong independence from the expression levels; and (iv) good for genetic backgrounds comparison of resting Ca2+.	(i) Large size of the indicators; (ii) tailored strategies for their targeting; (iii) possible silencing issues; and (iv) more sophisticated and expensive imaging equipment is required.
Intensiometric GECIs (e.g. GECOs/GCaMPs)	(i) Need simple microscope setups; (ii) sensitivity to detect subtle changes of Ca2+; (iii) they allow the easy measurement of Ca2+ in adult plants; and (iv) an easier combination of different spectral variants for simultaneous imaging of Ca2+ in different compartments.	(i) Possible artifacts related to their intensiometric nature and (ii) Ca2+ buffering with some high affinity and high Hill coefficient variants.

Genetically encoded calcium indicators

Aequorin

Aequorin is a bioluminescent protein that was discovered and isolated from the jellyfish Aequorea victoria by Shimomura et al. (1962). Shimamura was awarded the Nobel Prize in 2008 together with Martin Chalfie and Roger Tsien for the discovery and exploitation of the green fluorescent protein (GFP). Without the work done by these three visionary scientists this Update, this Focus Issue, and a large field of bio-imaging would not exist.

Aequorin is a 22-kDa holoprotein that contains bound oxygen, a prosthetic group, the coelenterazine, and three Ca²⁺-binding sites. When aequorin binds Ca²⁺ ions (Shimomura, 1995), it undergoes a conformational change, converting itself into a luciferase, which then catalyzes the luminescence reaction of coelenterazine. Coelenterazine is oxidized to coelenteramide with the emission of CO₂ and blue light (at 465 nm) which is caused by the decay of the coelenteramide from an excited state (Ohmiya and Hirano, 1996). This property offered a useful tool for detecting the concentration of Ca²⁺ ions in real time. However, the aequorin quantum yield is very low, requiring the simultaneous reactions of hundreds or thousands of proteins to collect enough photons to a level detectable by photon emission measurements (Mithöfer and Mazars, 2002). Nevertheless, identification of the aequorin gene and of its coding sequence (Prasher et al., 1985) allowed the protein to be expressed recombinantly in cells and tissues from different organisms. In the 1980s, the injection of recombinant aequorin in Chara allowed the detection of a transient increase in light emission when the cell generated an AP (Williamson and Ashley, 1982). A real revolution then occurred in 1991 when Marc Knight and colleagues generated the very first transgenic multicellular organism, Nicotiana plumbaginifolia plants, stably expressing aequorin (Knight et al., 1991; Figure 1). The authors reported that in response to touch, cold-shock, and elicitors, the plants showed clear photon emission determined by cytosolic Ca²⁺ increase. This demonstration opened de facto the modern era of Ca²⁺ imaging in plants. Indeed, the low quantum yield of aequorin hinders high-resolution imaging and, as a matter of fact, its detection was usually carried out with the use of a luminometer, collecting the emitted photons without creating an image (Mithöfer and Mazars, 2002). However, modern ultrasensitive cameras now allow the detecting of the photons emitted by a single whole plant, providing low-resolution images and averaged responses from different tissues or cells (Kiegle et al., 2000; Zhu et al., 2013; Kiep et al., 2015). Therefore, the recent technological advances in camera sensitivity have revitalized the use of aequorin imaging, posing the basis for the design of powerful and successful genetic screenings (Yuan et al., 2014; Jiang et al., 2019; Chen et al., 2020; Wu et al., 2020).

Even if aequorin is the oldest GECI, it still represents a reliable tool to study Ca²⁺ dynamics in plants, in particular, to determine quantitatively the magnitude of the responses. In fact, for every single experiment, the measurement of the light expressed as relative luminescence units (RLUs) can be converted into absolute [Ca²⁺] thanks to a calibration curve that considers the total amount of aequorin molecules present in the sample. This can be done by discharging the reconstituted aequorin with a solution containing Ca²⁺ and ethanol (100 mM CaCl₂, 10% ethanol (v/v)) (Allen et al., 1977; Mithöfer and Mazars, 2002). Specifically, the formula used to perform the conversion is the one reported in Mithöfer and Mazars (2002):

$$[{\text{Ca}}^{2+}]=\{{(L_0/L_{\mathit{max}})}^{1/3}+[\text{KTR}{(L_0/L_{\mathit{max}})}^{1/3}]–1\}/\{\text{KR}–[\text{KR}{(L_0/L_{\mathit{max}})}^{1/3}]\},$$

where L₀ is the aequorin luminescence intensity per second and L_max is the total amount of luminescence present in the sample over the experiment. KR and KTR are the dissociation constants for the first and second Ca²⁺ ions bound by aequorin, respectively (Mithöfer and Mazars, 2002).

As with every technique, aequorin-based Ca²⁺ analysis presents both advantages and disadvantages (Table 1). An advantage of aequorin is that it does not need to be excited with fluorescent light, which helps when the measurement of Ca²⁺ levels for long time intervals is required (Sai and Johnson, 2002; Love et al., 2004; Martí et al., 2013; Martí Ruiz et al., 2020) . Among the disadvantages, besides the already cited low quantum yield efficiency, is its requirement for the external provision of coelenterazine (Knight et al., 1991; Mithöfer and Mazars, 2002) which foresees the incubation of the plant material with the prosthetic group for several hours before starting the experiment. Moreover, after its oxidation, coelenterazine is irreversibly consumed, with the consequent decrease of the active aequorin pool. Another disadvantage of aequorin is the non-linearity of light emission rate as a function of the Ca²⁺ concentration (Robert et al., 2000; Table 1). Within the physiological range of [Ca²⁺]_cyt (10⁻⁷–10⁻⁵ M), the photon emission rate increases by more than 100-fold for a 10-fold change in [Ca²⁺]. In practical terms, this means that for a change in [Ca²⁺] from 0.1 µM (at resting) to 3–4 µM (in response to a stimulus), the rate of aequorin photon emission increases over 1000-fold. As a consequence, the signal coming from 1 activated cell will be the same as that of 1000 cells at resting [Ca²⁺]. If Ca²⁺ is not homogeneous in the population of cells, which is indeed the case when using entire seedlings or leaf disks, the overall aequorin light emission is dominated by the most responding cells subpopulation, leading to an averaging of the response. One way that can be followed to study more specifically the contribution of different cell types is the expression of aequorin under the control of tissue-specific promoters or using enhancer trap lines. By following this strategy, it was possible to discover the oscillatory dynamics of Ca²⁺ signaling in root cells (Kiegle et al., 2000), circadian gating of cold-induced Ca²⁺ oscillations in guard cells (Dodd et al., 2006), and cell-type and stimulus-specific Ca²⁺ oscillations (Martí et al., 2013).

Aequorin being a GECI could also be targeted to different subcellular compartments, such as the tonoplast (Knight et al., 1996), the nucleus (van Der Luit et al., 1999), the Golgi apparatus (Ordenes et al., 2012), mitochondria (Logan and Knight, 2003), plastids/chloroplasts (Johnson et al., 1995; Mehlmer et al., 2012; Sello et al., 2016), chloroplast subcompartments like the outer and inner envelope membranes (Mehlmer et al., 2012), and the thylakoid lumen and membrane (Sello et al., 2018; Table 2). Aequorin was also targeted to the apoplastic space (Gao et al., 2004).

Table 2.

Summary of available fluorescent-based GECIs used in plants

Name	Version	Type	Peaks of excitation/emission (nm)	In vitro Kd for Ca2+a	Subcellular localization	References
Cameleon	YC2.1	Ratio EYFP/ECFP	Ex 440/Em 480/530	0.8 μM/2 μM	Cytosol and nucleus	Allen et al. (1999); Miyawaki et al. (1999)
	YC3.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Cytosol and nucleus	Nagai et al. (2004); Mori et al. (2006)
	NES-YC3.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Cytosol	Krebs et al. (2012)
	NLS-YC3.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Nucleus	Krebs et al. (2012)
	NUP-YC3.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Nucleus	Costa et al. (2017)
	4mt-YC3.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Mitochondria	Loro et al. (2012)
	PM-YC3.6-LTI6b	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Plasma membrane	Krebs et al. (2012)
	2Bam4-YC3.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	250 nM	Chloroplasts and plastids	Loro et al. (2016)
	YC-Nano 65	Ratio cpVenus/ECFP	Ex 440/Em 480/530	65 nM	Cytosol and nucleus	Horikawa et al. (2010); Choi et al. (2014)
		Ratio cpVenus/ECFP
	SP-YC4.6-ER	Ratio cpVenus/ECFP	Ex 440/Em 480/530	58 nM/14.4 μM	Endoplasmic reticulum	Nagai et al. (2004); Iwano et al. (2009)
	2Bam4-YC4.6	Ratio cpVenus/ECFP	Ex 440/Em 480/530	58 nM/14.4 μM	Chloroplasts and plastids	Loro et al. (2016)
	4mt-D3cpv	Ratio cpVenus/ECFP	Ex 440/Em 480/530	600 nM	Mitochondria	Loro et al. (2013)
	D3cpv-KVK-SKL	Ratio cpVenus/ECFP	Ex 440/Em 480/530	600 nM	Peroxisomes	Palmer et al. (2006); Costa et al. (2010)
	TP-D3cpv	Ratio cpVenus/ECFP	Ex 440/Em 480/530	600 nM	Tonoplast	Krebs et al. (2012)
	CRT-D4ER	Ratio citrine/ECFP	Ex 440/Em 480/530	195 μM	Endoplasmic reticulum	Palmer et al. (2006); Bonza et al. (2013)
Twitch	Twitch 3	Ratio cpCit174/ECFP	Ex 440/Em 480/530	250 nM	Cytosol and nucleus	Thestrup et al. (2014); Waadt et al. (2017)
	Twitch 2B	Ratio cpVenus/mCerulean3	Ex 440/Em 480/530	200 nM	Cytosol and nucleus	Thestrup et al. (2014); Waadt et al. (2017)
CerTN-L15	CerTN-L15	Ratio citrine/cerulean	Ex 440/Em 480/530	1.2 μM	Cytosol and nucleus	Heim et al. (2007); Denninger et al. (2014)
GECOs	R-GECO1	Intensiometric mApple	Ex 561/Em 600	482 nM	Cytosol and nucleus	Zhao et al. (2011); Ngo et al. (2014); Keinath et al. (2015)
	NR-GECO1	Intensiometric mApple	Ex 561/Em 600	482 nM	Nuclear	Zhao et al. (2011); Kelner et al. (2018)
	NR-GECO1.2	Intensiometric mApple	Ex 561/Em 600	1.2 μM	Nuclear	Wu et al. (2013); Kelner et al. (2018); Leitão et al. (2019)
	CG-GECO1	Intensiometric cpGFP	Ex 488/Em 515	749 nM	Cytosol	Zhao et al. (2011); Kelner et al. (2018)
	CG-GECO1.2	Intensiometric cpGFP	Ex 405/Em 515	1.15 μM	Cytosol	Zhao et al. (2011); Kelner et al. (2018); Leitão et al. (2019)
	GEM-GECO1	Intensiometric cpGFP	Ex 405/Em 515	340 nM	Cytosol and nucleus	Zhao et al. (2011); Waadt et al. (2017)
	B-GECO1-mCherry	Ratio cpGFP/mCherry	Ex 405/561/Em 480/600	NA	Cytosol and nucleus	Waadt et al. (2017)
	G-GECO1.1-mCherry	Ratio cpGFP/mCherry	Ex 488/561/Em 515/600	NA	Cytosol and nucleus	Waadt et al. (2017)
	R-GECO1-mTurquoise	Ratio mApple/mTurqouise	Ex 405/561/Em 480/600	NA	Cytosol and nucleus	Waadt et al. (2017)
GCaMPs	GCaMP3	Intensiometric cpGFP	Ex 488/Em 515	542 nM	Cytosol and nucleus	Tian et al. (2009); Nguyen et al. (2018)
	GCaMP5	Intensiometric cpGFP	Ex 488/Em 515	NA	Cytosol and nucleus	Akerboom et al. (2012); Diao et al. (2018)
	GCaMP6f	Intensiometric cpGFP	Ex 488/Em 515	375 nM	Cytosol and nucleus	Chen et al. (2013); Waadt et al. (2017)
	GCaMP6s	Intensiometric cpGFP	Ex 488/Em 515	144 nM	Cytosol and nucleus	Liu et al. (2017); Shao et al. (2020)
	NES-GCaMP6m	Intensiometric cpGFP	Ex 488/Em 515	167 nM	Cytosol	Luo et al. (2020)
	NLS-GCaMP6m	Intensiometric cpGFP	Ex 488/Em 515	167 nM	Nucleus	Luo et al. (2020)
	ER-GCaMP6-210	Intensiometric cpGFP	Ex 488/Em 515	210 μM	Endoplasmic reticulum	de Juan-Sanz et al. (2017); Resentini et al. (2021a)
	GCaMP6f-mcherry	Ratio cpGFP/mCherry	Ex 488/561/Em 480/600	NA	Cytosol and nucleus	Waadt et al. (2017)
	MatryoshCaMP6s	Ratio cpGFP/LSSmOrange	Ex 440/Em 515/600	197 nM	Cytosol and nucleus	Ast et al. (2017)
Case	Case12	Intensiometric cpGFP	Ex 488/Em 515	1 μM	Cytosol and nucleus	Souslova et al. (2007); Zhu et al. (2013)
CEPIA	CRT1a-R-CEPIAer	Intensiometric cpGFP	Ex 561/Em 600	565 μM	Enoplasmic reticulum	Suzuki et al. (2014); Luo et al. (2020)
Aequorin	Aequorin	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Cytosol and nucleus	Knight et al. (1991); Brini et al. (1995)
	Aequorin	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Nucleus	van Der Luit et al. (1999)
	Aequorin	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Chloroplast stroma	Johnson et al. (1995)
	Aequorin	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Mitochondria	Logan and Knight (2003)
	Aequorin	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Golgi	Ordenes et al. (2012)
	Aequorin	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Vacuole/tonoplast	Knight et al. (1996)
YFP-aequorin	CYA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Cytosol	Mehlmer et al. (2012)
	NYA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Nucleus	Mehlmer et al. (2012)
	YA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Plasma membrane	Mehlmer et al. (2012)
	CHYA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Chloroplast/plastid stroma	Mehlmer et al. (2012); Sello et al. (2016)
	MYA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Mitochondria	Mehlmer et al. (2012)
	OEYA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Chloroplast outer envelope	Mehlmer et al. (2012); Sello et al. (2016)
	IEYA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Chloroplast inner envelope	Mehlmer et al. (2012); Sello et al. (2016)
	TL-YA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Chloroplast thylakoid lumen	Sello et al. (2018)
	TM-YA	Bioluminescence	No Ex-/Em 465	7.2–13 μM	Chloroplast thylakoid membrane	Sello et al. (2018)
GFP5-aequorin	pchitGFP5:AQ	Bioluminescence	No Ex-/Em 465	NA	Apoplast	Gao et al. (2004)
GFP-aequorin	G5A	Bioluminescence resonance energy transfer	No Ex-/Em 515	NA	Cytosol and nucleus	Baubet et al. (2000); Xiong et al. (2014)

^aThe in vitro K_d for Ca²⁺ of the different sensors are those reported in the original works.

In A. victoria, the photons emitted by aequorin excite the GFP which causes the jellyfish to emit fluorescence (Prasher et al., 1992). In laboratory experiments, this property has been exploited to generate a bioluminescence resonance energy transfer (BRET)-based GFP-aequorin reporter (i.e. G5A; Baubet et al., 2000; Rogers et al., 2005). This BRET-based sensor has overcome one of the major limitations of aequorin, that is its low amount of emitted light. This allows easier detection of [Ca²⁺] variations using a camera by permitting the possibility to reduce the exposure time compared with aequorin imaging. However, this sensor still requires the exogenous administration of coelenterazine. Nevertheless, the G5A sensor expression in Arabidopsis (Arabidopsis thaliana) plants by Xiong et al. (2014) allowed visualization for the first time of long-distance Ca²⁺ waves propagating from the roots to shoot upon salt treatment.

FRET-based fluorescent sensors

To overcome aequorin limitations, in the late 1990s, scientists started to exploit fluorescent proteins (FPs) (i.e. GFP) and its spectral variants to develop another generation of GECIs. The first ever fluorescence-based GECI, named Yellow Cameleon (YC), was developed by Roger Tsien in 1997 (Miyawaki et al., 1997) quickly followed by the development of the YC2.1 version (Miyawaki et al., 1999) which was shortly after expressed in plants (Allen et al., 1999; Figure 1). This sensor is based on the Ca²⁺-induced protein–protein interaction between CaM and the CaM-binding peptide M13, a fragment from the myosin light chain kinase (Miyawaki et al., 1997). Both components are fused by a flexible linker and are sandwiched between a FP pair with partial overlap between the emission spectrum of the donor and the absorption spectrum of the acceptor. Upon Ca²⁺ binding, CaM interacts with M13, bringing the donor and acceptor FPs into sufficiently close proximity to allow for Förster resonance energy transfer (FRET) between them through a nonradiative dipole–dipole coupling. This phenomenon can be measured via fluorescence microscopy. At the level of the acceptor molecule, it results in the emission of fluorescence upon excitation of the donor, allowing the estimation of Ca²⁺-induced FRET by the ratio of donor and acceptor emission upon excitation of the donor using relatively simple microscopy setups (Miyawaki et al., 1997; Rudolf et al., 2003).

A wide variety of FRET-based GECI has been developed based on FRET pairs with different spectral properties and in combination with a variety of Ca²⁺ regulated protein–protein interactions (Greenwald et al., 2018; https://biosensordb.ucsd.edu/biosensorDB/bsSearch.php). The spectral range of the FRET pairs is usually cyan/yellow, green/red occupying a wide spectrum, reducing the options for dual imaging with other sensors or reporters.

In plants, the most popular FRET-based GECIs are based on cyan and yellow FRET pairs linked together by the Ca²⁺-binding protein CaM and the CaM-binding peptide M13, as in the original YC configuration by Roger Tsien. FRET, and thus [Ca²⁺] increases, can be conveniently measured by the increase in the ratio between the emission intensity of Enhanced Yellow FP (EYFP) and Enhanced Cyan FP (ECFP) upon ECFP excitation. The Cameleon indicators can go back and forth from the bound and unbound Ca²⁺ state. However, this is limited by the bleaching of the sensor by the excitation light especially when high magnifications are used, which may cause photo-oxidative cellular stress (Laissue et al., 2017). One of the most important properties of Cameleon is its pure ratiometric nature: a single wavelength excitation and a dual emission (Miyawaki et al., 1997). The advantages of having a ratiometric sensor will be further discussed in the next section. The Cameleon’s reliability brought, in the first years of the 2000s, an “explosion” of different variants, with different Ca²⁺ affinities, different linkers, and different FP pairs (Table 2;Palmer and Tsien, 2006). In 2004, the Miyawaki group developed the Cameleon YC3.60 (often called YC3.6) where the EYFP was replaced with a circularly permuted variant of the Venus FP (cpVenus) (Nagai et al., 2002). The use of the cpVenus as a FRET acceptor greatly increased the energy transfer efficiency from the donor (ECFP), which in practical terms allowed the very reliable performance of in vivo measurements. In fact, in living cells, the simultaneous decrease in ECFP and increase in cpVenus fluorescence emissions, due to FRET, were almost identical to the in vitro analyses (Nagai et al., 2004). This property is of great relevance because it permits ascertaining with high confidence that a change of cpVenus/ECFP ratio, even if small, corresponds to actual FRET levels, thus, a real [Ca²⁺] change. This latter aspect gains importance when a single cell or single organelle imaging is performed. Moreover, whereas both ECFP and cpVenus might show a pH sensitivity, the FRET ratio is almost unaffected, at least in a narrow change of physiological cytosolic pHs (around pH 7–7.5; Nagai et al., 2004; Behera et al., 2018).

The YC3.6 was expressed in Arabidopsis under the control of the guard cell and pollen-specific promoters, pGC1 and pLat52, respectively, which allowed detecting spontaneous Ca²⁺ oscillations in these two cell types with an improved FRET efficiency (Mori et al., 2006; Yang et al., 2008; Iwano et al., 2009) in comparison to the first YC versions used in plants, the YC2.1 (Allen et al., 1999) and the YC3.1 (Michard et al., 2008; Iwano et al., 2009). The YC3.6 was then expressed under the control of the CaMV35S promoter, making it possible in Arabidopsis to study, at high spatial and temporal resolution, root hair tip Ca²⁺ oscillations (Monshausen et al., 2008). The next step in the exploitation of YC3.6 was obtained by Schumacher’s group which placed the sensor with a cytosolic or nuclear localization signal (NES-YC3.6 and NLS-YC3.6) under the control of the pUBQ10 promoter (Grefen et al., 2010; Krebs et al., 2012) that offered a homogeneous expression of the indicator in both Arabidopsis wild-type and mutant backgrounds (Wagner et al., 2015; Teardo et al., 2017; Behera et al., 2018, Corso et al., 2018; Hazak et al., 2019; Wang et al., 2020) as well as in rice (Oryza sativa; Behera et al., 2015). YC3.6 was also successfully expressed in the moss Physcomitrium patens (formerly Physcomitrella patens), allowing the visualization of systemic Ca²⁺ wave propagations in the absence of vascular tissues (Storti et al., 2018). Another important breakthrough was the expression of YC3.6 under the control of the synergid-specific promoter, pMYB98, which allowed the monitoring of Ca²⁺ dynamics in this type of cells during the fertilization process (Hamamura et al., 2014; Ngo et al., 2014). The versatility and reliability of YC3.6 were also demonstrated by its use for the analysis of Ca²⁺ dynamics in different subcellular compartments including mitochondria, chloroplast stroma, and subplasmalemmal space (Table 2;Krebs et al., 2012; Loro et al., 2012, 2016; Storti et al., 2018).

Based on the original work from Nagai et al. (2004) the in vitro K_d of the YC3.6 is 250 nM (Table 2) with a Hill coefficient (Table 3) of 1.7. The knowledge of these two parameters allows the rough conversion of the cpVenus/ECFP ratio into an [Ca²⁺] if the ratio minimum and ratio maximum are experimentally measured (Palmer and Tsien, 2006; Monshausen et al., 2008; Wagner et al., 2015). One formula that can be used to perform the conversion is the one reported in Monshausen et al. (2008):

$$\left[{\text{Ca}}^{\text{2}+}\right]=K_{\text{d}}\left(R–R_{\text{min}}\right)/{\left(R_{\text{max}}–R_{\text{min}}\right)}^{\text{1}/n},$$

where R represents the cpVenus/ECFP ratio measured at any given time during the experiment, n represents the Hill coefficient, and the K_d the in vitro affinity for Ca²⁺ (Table 3). However, since both the K_d and the Hill coefficients are usually measured in vitro and not in vivo, the ratio conversion into concentration values must be taken with caution. As a matter of fact, a recent work reported for the YC3.6 an in vitro K_d of 719 nM and a Hill coefficient of 2.12 (Li et al., 2021), values that are quite distant from those reported in the original work by Nagai et al. (2004) (250 nM and 1.7, respectively); therefore, pointing out the need for prudence when the conversion is applied (Palmer and Tsien, 2006). Nonetheless, the YC3.6 is indeed a sensor suitable for the analysis of Ca²⁺ dynamics when a given stimulus can induce an increase that is around and above its K_d value. It is also true that in response to stimuli that induce small Ca²⁺ increases, YC3.6 shows some limitations in comparison to the more recent generation of ultrasensitive GECIs (Keinath et al., 2015; Waadt et al., 2017), being in fact not efficient at detecting very subtle changes of [Ca²⁺] (Keinath et al., 2015). However, the beauty of any genetically encoded sensor is that by following rational and random mutagenesis approaches they can be modified to address specific needs. In 2010, the group of Nagai generated a series of Cameleon variants with higher affinity for Ca²⁺ that were dubbed YC-Nano (Table 2;Horikawa et al., 2010). In particular, the YC-Nano 65 (with an in vitro K_d for Ca²⁺ of 65 nM; Table 2) was efficiently expressed in Arabidopsis and this was instrumental to demonstrate the existence of a long-distance subtle Ca²⁺ wave in seedlings locally challenged with salt stress (Choi et al., 2014). The same sensor was used to compare the cytosolic [Ca²⁺] at resting and in response to wounding between the wild type and fatty acid oxygenation upregulated 2 (fou2) mutant (Lenglet et al., 2017). Very recently, YC-Nano 65 was used to study the response to flg22 in cotyledon and leaf cells in the Arabidopsis wild type, aca4/aca11 and aca1/2/7 autoinhibited Ca²⁺-ATPase mutants (Hilleary et al., 2020; Ishka et al., 2021).

Table 3.

Brief summary of the principal definitions used to describe GECIs properties

Biochemical parameter	Description
K d (µM)	Apparent dissociation constant for Ca2+ of the sensor, at this concentration, half of the indicators are bound with Ca2+. Given that Hill coefficients are usually higher than 1, the Kd delineates the optimal concentration range at which a GECI should be used.
k on (s−1)	Indicates the rate of Ca2+ association shown by the sensor, and thus the speed by which a sensor responds to an increase in Ca2+ levels.
k off (s−1)	Indicates the rate of Ca2+ dissociation shown by the sensor, and thus the speed by which a sensor responds to a decrease in Ca2+ levels.
Hill coefficient	Indicates the cooperativity of the sensor in the Ca2+ binding process. A value greater than 1 indicates that binding of one Ca2+ ion facilitates the binding of another. The CaM-based GECI bind four Ca2+ ions. The closer to 1 the more linear the output of the reporter.
Dynamic range	For non-ratiometric indicators indicate the maximal fluorescence intensity (typically in a Ca2+-bound state) divided by minimal fluorescence intensity (determined in the presence of EGTA). For ratiometric indicators indicate the maximum fluorescence emissions ratio (typically in a Ca2+-bound state) divided by minimal fluorescence emissions ratio (determined in the presence of EGTA).

A side-by-side in vivo comparison with plants expressing the YC3.6 and the YC-Nano 65 has not been published yet, but we can predict that the higher affinity for Ca²⁺ of the YC-Nano could saturate in response to different stimuli. Moreover, the different biochemical properties of YC-Nano 65 and YC3.6, like the rate of association (k_on) and dissociation (k_off) for Ca²⁺ (Table 3), can determine different sensor dynamics which can be attributed in vivo, as different Ca²⁺ dynamics. We do, therefore, suggest not comparing the published data obtained with these two sensors, but rather, to perform preliminary independent experiments to see which one offers the best readout in relation to the applied stimulus or developmental program under investigation.

An important aspect that needs to be considered is that plants expressing the Cameleon YC3.6 in the cytosol do not show any obvious gross phenotypes, pointing out that the sensor per se does not alter the Ca²⁺ homeostasis and the plant physiology in stable mature plants (Waadt et al., 2017). However, possible effects of Ca²⁺ buffering should be considered in every biological process of interest.

Besides Cameleon and its variants (Table 2), there are other FRET-based sensors that have been successfully used in plants. We can cite here the CerTN-L15 (Heim et al., 2007; Denninger et al., 2014) and the Twitch 2B and 3 (Thestrup et al., 2014; Waadt et al., 2017) that are alternative FRET-based Ca²⁺ sensors which instead of having the CaM domain, use troponin C, a protein exclusively found in myocytes, to sense and bind Ca²⁺. The substitution of the CaM domain with troponin C prevents any interference due to the endogenous CaM when present at high concentrations such as in the subplasmalemmal region (Miyawaki et al., 1999; Palmer et al., 2006). Plants expressing these two sensors have been shown to properly report [Ca²⁺] dynamics in synergids and root cells, but their use has so far been limited (Denninger et al., 2014; Waadt et al., 2017). An alternative to the FRET-based sensors might be represented by the use of dimerization-dependent FPs (ddFPs), which is a technology involving the reversible binding of two dark FP monomers to form a fluorescent heterodimeric complex (Alford et al., 2012).

Single FP GECIs

Single fluorophore-based Ca²⁺ indicators GECIs are intensiometric Ca²⁺ sensors, based on a circularly permuted FP (e.g. GFP, YFP, or mApple; Baird et al., 1999; Nakai et al., 2001) fused at its C- and N-termini with the components of a Ca²⁺ sensing module (i.e. CaM domain and the M13 peptide). In the presence of Ca²⁺, this causes a tightening of the interaction between C- and N-termini of the fluorophore, protecting the chromophore from the environment and leading to increased brightness. In simple words, this interaction induces a dramatic alteration of the spectral properties of the FP with a strong increase in the fluorescence emitted, thus, making these GECIs suitable to indicate Ca²⁺ levels in real time (Nakai et al., 2001). Currently, this principle has been also exploited for the development of different types of biosensors, including kinase activity reporters (e.g. ExRAI; Greenwald et al., 2018).

Similarly to Cameleon, for single FP sensors there was a strong development that yielded a family of GECIs with different colors, different Ca²⁺ affinities, different Hill coefficients, less pH sensitivity, and with improved signal-to-noise ratio (Table 2; Zhao et al., 2011). Just to cite an example, from 2001 to 2019, there has been an evolution that has led from the GCaMP to the jGCaMP7 (Figure 1;Nakai et al., 2001; Tian et al., 2009; Akerboom et al., 2012; Chen et al., 2013; Greenwald et al., 2018; Dana et al., 2019). The aim to improve GCaMPs was driven by the necessity to increase the sensitivity and kinetics of the sensors, trying to make them closer, in terms of properties, to the synthetic Ca²⁺ dyes that for the needs of neuroscience are still among the most sensitive and rapid Ca²⁺ indicators. It is worth considering that in plant cells, the kinetics of the sensor response does not represent a big limitation, since in most of the published works the imaging sampling was set at every 2–5 s (e.g. guard cells, pollen tubes and root hair growth, and root tip cells; Allen et al., 1999; Monshausen et al., 2008; Yang et al., 2008; Michard et al., 2011; Candeo et al., 2017; Li et al., 2021; Resentini et al., 2021a). Instead, the ease of use and the sensitivity of single FP GECIs are good properties that have pushed the plant community to move toward their use. The fact that single FP GECIs rely on a single excitation and a single emission makes them particularly suitable to be combined with other fluorescent markers or sensors, and for their use with simple and accessible microscope equipment (Table 1). The first single FP GECI expressed in plant cells were the GFP-based Ca²⁺ indicator Case12 (Zhu et al., 2013) and the red-shifted R-GECO1 (Ngo et al., 2014; Keinath et al., 2015). The latter is a red fluorescent GECI derived from the GCaMP3 (Tian et al., 2009), where the circularly permuted GFP (cpGFP) was substituted with the cpmApple (Zhao et al., 2011). Importantly, R-GECO1 is excited with green light (e.g. 561 nm; Table 2) which besides offering a greater tissue penetration neither stimulates photosynthesis nor photoreceptors (Taiz et al., 2014), making it particularly suited for use in green tissues. In 2015, R-GECO1 was first used in a series of “classical experiments” in Arabidopsis root meristems, treated with external ATP and pathogen elicitors such as flg22 and chitin, and demonstrating the very high sensitivity of this sensor (Keinath et al., 2015). It must be said that the same sensor was also previously expressed in pollen tubes used to fertilize an Arabidopsis line expressing the Cameleon YC3.6 in synergids, thus enabling Ca²⁺ imaging in two different tissues (Ngo et al., 2014). The in vitro K_d of R-GECO1 of 482 nM (Table 2) with a Hill coefficient of 2.06 (Zhao et al., 2011) makes this sensor suitable to efficiently detect cytosolic Ca²⁺ variations, and its constitutive expression under the pUBQ10 promoter in the stable Arabidopsis line does not produce gross and visible phenotypes (Waadt et al., 2017; Resentini et al., 2021a). Instead, when other color variants of the GCaMP3 (e.g. GEM-GECO1, B-GECO1, and G-GECO1.1; Table 2) were stably expressed in Arabidopsis, the plants showed some growth defects (Waadt et al., 2017). This negative effect on plant physiology indicates that these reporters with a higher Hill coefficient for GEM- (2.94) and B-GECO1 (2.64) or a lower Ca²⁺ affinity for the G-GECO1.1 (K_d 618 nM; Zhao et al., 2011) might potentially act as Ca²⁺ buffers, but this needs to be demonstrated. Overall, as a good practice, it is advised to compare the phenotype of a chosen sensor line with wild-type plants in the context of a given process of interest, and never to forget that the sensor could have a potential effect on plant physiology.

The GCaMP3 is a relatively old GCaMP version (Tian et al., 2009), and neuroscientists have now moved to the use of the most recent versions (GCaMP6 and jGCaMP7). Nevertheless, GCaMP3 was only recently expressed in Nicotiana benthamiana, Arabidopsis, and P. patens (Figure 1;DeFalco et al., 2017; Kleist et al., 2017; Vincent et al., 2017; Nguyen et al., 2018; Toyota et al., 2018; Krogman et al., 2020) and allowed the detection of bright long-distance leaf-to-leaf Ca²⁺ waves (Toyota et al., 2018; Nguyen et al., 2018). GCaMP3 has been also expressed under the control of tissue-specific promoters (i.e. phloem and different root tissues; Vincent et al., 2017; Toyota et al., 2018; Krogman et al., 2020).

Similarly to EYFP/ECFP FRET-based sensors, GCaMPs come with some disadvantages, as the blue excitation light can stimulate photosynthesis (Taiz et al., 2014) and can trigger in itself [Ca²⁺]_cyt increase via phototropins (Zhao et al., 2013; Ishka et al., 2021). This could be avoided using R-GECO1, which being excited with green light is possibly more compatible with long-term imaging in the shoot. Other GCaMP versions, such as GCaMP5, GCaMP6f, and GCaMP6s (Table 2;Zhu et al., 2013; Liu et al., 2017; Vincent et al., 2017; Diao et al., 2018; Shao et al., 2020; Suda et al., 2020; Ishka et al., 2021) as well as other green and red variants of GECO1 (CG-GECO1.2 and R-GECO1.2; Kelner et al., 2018; Table 2) have also been used in plant cells.

Such a wealth of available tools might, however, generate some confusion in choosing one version over another. Indeed, the different GCaMP variants have different colors, different K_d for Ca²⁺ which make them suitable for use in different compartments or, similarly to the use of the YC-Nano 65, to report subtle changes of [Ca²⁺]. As an example, when subtle changes of [Ca²⁺] are expected, the high affinity for Ca²⁺ of GCaMP6s (K_d of 144 nM) is more suitable than GCaMP3 (K_d of 542 nM) or GCaMP6f (K_d of 375 nM; Table 2). However, a complete side by side comparison of all different sensors used in plant cells is not available, with only some tested by Waadt et al. (2017). Possibly an interested researcher should preliminarily test them to identify the best one for each experimental condition. We are currently working with both R-GECO1 and GCaMP3 expressing plants, and overall, they provide similar results. Nevertheless, as anticipated above, the most recent generation of single FP GECIs is continuously improving to increase the sensitivity and the signal-to-noise ratio to perform imaging in neuronal tissues, which might also be useful for plant scientists. As a matter of fact, all the single FP GECIs exploit an off/on response with a negligible fluorescent signal in the Ca²⁺ unbound state and strong fluorescence with even relatively small changes in [Ca²⁺]. Such a property represents a big advantage when imaging an entire plant. Indeed, both GCaMP3 and R-GECO1 have demonstrated that they offer enough sensitivity to detect and measure Ca²⁺ wave propagation between leaves in adult Arabidopsis plants in response to wounding or insect chewing (Nguyen et al., 2018; Toyota et al., 2018; Resentini et al., 2021a; Table 1). This result represents a major achievement in the field that was brought to a new dimension when the GCaMP6f was expressed in the carnivorous plant Dionaea muscipula, also known as Venus flytrap (Figure 1; Suda et al., 2020). The Venus flytrap rapidly closes the valves following mechanical stimulation of “sensitive hairs” and a role of Ca²⁺ as a second messenger in the closing mechanism was previously hypothesized (Scherzer et al., 2015; Hedrich and Neher, 2018). By using the GCaMP6f, Suda et al. (2020) demonstrated that the stimulation of the sensory hairs had the effect of inducing a Ca²⁺ wave that propagated throughout the entire leaf blade at a speed greater than 20 mm/s (Suda et al., 2020). Such a propagation rate exceeds at least 20 times the Ca²⁺ response to leaf injury in Arabidopsis (Toyota et al., 2018; Shao et al., 2020), and thus requires GECI with fast k_on-kinetics such as the GCaMP6f.

In conclusion, the use of different single FP GECIs provides reliable data and offers the chance to capture dynamic processes in multicellular adult organisms, opening opportunities to explore new hypotheses. Also, Cameleon sensors allow the performance of large imaging experiments (Beneloujaephajri et al., 2013; Benikhlef et al., 2013; Costa et al., 2017; Behera et al., 2018; Doccula et al., 2018; Hilleary et al., 2020; Ishka et al., 2021), but for their use, a piece of more sophisticated and expensive equipment is required, whereas single FP GECIs need a simple single excitation, single emission fluorescence microscope equipped with a good camera (Table 1; Vincent et al., 2017; Nguyen et al., 2018; Toyota et al., 2018; Resentini et al., 2021a).

A notable disadvantage of the single FP GECI is that the obtained intensities are not only determined by Ca²⁺ levels but are also dependent on local expression levels and can be quenched at low pH.

Organellar calcium dynamics: state of the art

The contribution of organellar Ca²⁺ handling in the regulation of signaling processes has been recently reviewed, so we will redirect interested readers to those papers (Stael et al., 2012; Costa et al., 2018; Pirayesh et al., 2021; Resentini et al., 2021b). Nevertheless, here, we briefly summarize what GECIs are available for the study of Ca²⁺ dynamics in subcellular compartments and which tools, among those recently developed, we foresee for the advancement of the field. Table 2 is an updated version of the table published in Costa et al. (2018) and reports the most available Arabidopsis lines expressing subcellular targeted GECIs.

Aequorin was targeted to the cytosol, nucleus, mitochondria, Golgi apparatus, tonoplast, apoplast, chloroplast, and chloroplasts’ subcompartments (Knight et al., 1991, 1996; Johnson et al., 1995; van Der Luit et al., 1999; Logan and Knight, 2003; Gao et al., 2004; Mehlmer et al., 2012; Ordenes et al., 2012; Sello et al., 2016, 2018). One needs to bear in mind that aequorin can offer only an averaged response of organelles or subcellular compartments from different cells and often from multiple plants. Therefore, in the case of the requirement of single organelles, the use of fluorescent GECIs is required.

Both FRET-based cameleon and single FP sensors were targeted to different subcellular locales, allowing the detection of Ca²⁺ dynamics within different compartments with, in some cases, a single cell or even single organelle resolution (Iwano et al., 2009; Costa et al., 2010, 2013; Krebs et al., 2012; Loro et al., 2012, 2016; Bonza et al., 2013; Kelner et al., 2018; Leitão et al., 2019; Resentini et al., 2021a). A special mention is needed for the simultaneous expression, in Medicago truncatula and Arabidopsis, of the G-GECO1.1 (and 1.2) and R-GECO1.1 (and 1.2) sensors targeted, respectively, to the cytosol and nucleus (Kelner et al., 2018; Leitão et al., 2019). High-resolution imaging of these sensors in Medicago root hairs revealed that in response to Nod-Factors, the nuclear Ca²⁺ increase preceded the cytosolic one (Kelner et al., 2018). Similarly, a nuclear spontaneous Ca²⁺ increase anticipated the cytosolic one in Arabidopsis root tip meristematic cells during growth (Leitão et al., 2019). Moreover, the use of Cameleon YC3.6 targeted to mitochondria and nucleus enabled the simultaneous imaging of Ca²⁺ in these two compartments in Arabidopsis guard cells challenged with osmotic stress (Loro et al., 2012) or root tip cells in response to ATP (Loro et al., 2013).

At the present time, most of the Ca²⁺ analyses in subcellular compartments have been carried out with Cameleon variants, and only limited use of recent generation GECIs has been exploited (reviewed in Costa et al., 2018; Luo et al., 2020). The trend is different in animal cells, where more recent GECIs have been used in different subcellular compartments. Modified GCaMP6, CEPIA, and RCaMP sensors (Greenwald et al., 2018) with different spectral features or lower affinities for Ca²⁺ have shown to be suitable for analyses of Ca²⁺ dynamics in the endoplasmic reticulum (ER) and mitochondria of single neurons in response to APs (Suzuki et al., 2014; Dana et al., 2016; de Juan-Sanz et al., 2017; Ashrafi et al., 2020). Hence, this is the right time to explore the use of these sensors in plants, as they can offer some advantages compared with Cameleon sensors, although there are other limitations. The use of Cameleon sensors for the analysis of Ca²⁺ dynamics in subcellular compartments is primarily justified by its ratiometric nature that helps to reduce artifacts when measuring Ca²⁺ in moving organelles at high magnification or when comparing the steady-state concentration (Table 1;Wagner et al., 2015; Corso et al., 2018). As an example, based on our own experience, the large size of Cameleon (∼73 KDa) may pose some limitations for its proper targeting. In the case of mitochondria, ER, and chloroplasts, the Cameleon targeting required tailored strategies. To prevent a cytosolic mislocalization of the Cameleon (i) the mitochondrial targeting sequence had to be repeated four times (Palmer et al., 2006; Loro et al., 2012), (ii) the targeting to the ER required both a KDEL sequence at the C-terminus end plus a plant-specific calreticulin targeting sequence at the N-terminus end of the sensor (Palmer et al., 2004; Iwano et al., 2009; Bonza et al., 2013), and (iii) the targeting to the chloroplast stroma needed to double the BETA-AMYLASE 4 (Bam4) targeting sequence (Loro et al., 2016). The proper targeting of the Cameleon was indeed obtained, but this was at the expense of different issues, such as a more difficult design of the construct, and silencing. As an example, a strong silencing was relevant when the Cameleon was targeted to chloroplasts (2Bam4-YC3.6). In fact, the generation of stable fluorescent plants could be obtained only in the Arabidopsis RNA-dependent RNA polymerase-6-11 (rdr6-11) mutant background which is compromised in silencing (Peragine et al., 2004; Loro et al., 2016). Nevertheless, with the ER-targeted version of the Cameleon (CRT-D4ER), its expression was also limited to juvenile tissues, with some difficulties found in expressing it in some T-DNA lines (Corso et al., 2018; Shkolnik et al., 2018). We hope that the use of single FP GECIs might reduce silencing issues. A recent publication has reported the use of the ER-localized CRT1a-R-CEPIAer sensor and no issues related to silencing were highlighted (Luo et al., 2020). We have also successfully expressed the ER-GCaMP6-210 sensor localized in the ER (Table 2; de Juan-Sanz et al., 2017; Resentini et al., 2021a) in Arabidopsis wild-type and R-GECO1 backgrounds, offering the possibility to perform the simultaneous Ca²⁺ imaging in the ER and cytoplasm of adult plants (Resentini et al., 2021a). This result was not reached with the CRT-D4ER, not even in the ER alone, where the fluorescence was lost with aging.

We foresee that in the coming years, the use of single FP GECIs for the analyses of organellar Ca²⁺ dynamics should accompany the use of Cameleon, in particular for the simultaneous analysis of Ca²⁺ dynamics in different compartments, by combining sensors with different excitation and emission spectra (Resentini et al., 2021a). Moreover, organellar-targeted single FP GECIs should also be employed to analyze Ca²⁺ dynamics in adult plants, by exploiting their high sensitivity and dynamic range (Table 3).

Ratiometric versus intensiometric GECIs: a practical guide for their choice

In this Update article, we have referred several times to the advantages offered by using ratiometric FRET-based sensors compared with intensiometric ones. In this section, we want to further discuss this statement. Traditionally, ratiometry-based Ca²⁺ recordings rely completely on ratio changes, and these measurements are not influenced by the actual amount of the indicator or by changes in the focusing position of the imaging system (Rudolf et al., 2003). This is particularly true with FRET-based sensors since the fluorescence emission from the acceptor depends on the light absorbed by the donor, and a single excitation is used, requiring a single excitation light path. In practical terms, one of the advantages of using a FRET-based ratiometric sensor, like Cameleon, is that there is the possibility to efficiently compare the steady-state ratios among different genotypes or in the same genetic background in response to long-term treatments. This is true for different subcellular compartments such as cytosol, mitochondria, and ER (Laanemets et al., 2013; Wagner et al., 2015; Lenglet et al., 2017; Yang et al., 2017; Behera et al., 2018; Corso et al., 2018; Doccula et al., 2018; Shkolnik et al., 2018; Fasani et al., 2019; Hilleary et al., 2020; Ishka et al., 2021). Another advantage of ratiometric FRET-based sensors is that the results can be presented as raw acceptor/donor fluorescence emissions ratios (e.g. cpVenus/ECFP) without the need for calibration. Our direct experience with intensiometric sensors has revealed that the fluorescence is already variable at resting conditions, even within the same genetic background grown in the same conditions (Figure 3, B). To obviate the fluorescence variability among samples, when R-GECO1 or GCaMPs are used, the data are often presented as normalized fluorescence and not as absolute values (Vincent et al., 2017; Dindas et al., 2018; Nguyen et al., 2018; Toyota et al., 2018; Resentini et al., 2021a). Usually, this does not represent a problem; however, the need to rely on normalization may represent an issue for a correct interpretation of the results. To support this statement, we here present a set of experimental data that help to better clarify this point.

Figure 3.

Comparison of the Ca²⁺ response in root tip cells expressing YC-Nano 65 or R-GECO1. Seedlings were exposed to a pulse of 10 µM NAA and medium containing either 1.5 mM calcium (CaCl₂) or 0-calcium. The schematic drawing on the right shows the root tip region examined. A, cpVenus/ECFP ratio at rest in root tip cells of YC-Nano 65 seedling in 1.5 mM calcium and 0-calcium. B, R-GECO1 fluorescence at rest in root tip cells of R-GECO1 seedlings in 1.5 mM calcium and 0-calcium. C, Normalized averaged YC-Nano 65 cpVenus/ECFP ratios (R) (ΔR/R₀ = (R−R₀)/R₀) over time in root tip cells in response to NAA as indicated by the black box on the x-axis. D, Normalized averaged R-GECO1 fluorescence (F) (ΔF/F₀ = (F−F₀)/F₀) over time in root tip cells in response to NAA as indicated by the black box on the x-axis. R₀ and F₀ are the prestimulus values of R and F, respectively. The shaded right-side arrow indicates the direction of the [Ca²⁺]_cyt increase. E, Maximal relative amplitude as ΔR_max/R₀ of cpVenus/ECFP ratio triggered by NAA in 1.5 mM calcium and 0-mM calcium conditions. F, Maximal relative amplitude as ΔF_max/F₀ of R-GECO1 fluorescence triggered by NAA in 1.5 mM calcium and 0-calcium conditions. G and H, Same experiments shown in (C and D) without normalization to the pre-stimulus situation. The shaded right-side arrow indicates the direction of the [Ca²⁺]_cyt increase. I, Maximal amplitude of cpVenus/ECFP ratio triggered by NAA application in 1.5 mM calcium and 0-calcium. J, Maximal amplitude of R-GECO1 fluorescence triggered by NAA application in 1.5 mM calcium and 0-calcium. Arbitrary units of fluorescence (AUF). n ≥ 6. Data were plotted as box-and-whisker plots using GraphPad, in which all the experimental points are plotted, and their distribution represented as a box that extends from the 25^th to 75^th percentiles. The line in the middle of the box is plotted at the median. P values were calculated with an unpaired Student’s t test. Error bars = sd. **P ≤ 0.005, ***P ≤ 0.0005 (t test).

Arabidopsis seedlings expressing the Cameleon YC-Nano 65 and the R-GECO1 were germinated and grown in a standard growth half-strength MS medium and at 6-d-old stage were independently incubated for 10 min in solutions with different concentrations of Ca²⁺ (0-calcium and 1.5 mM calcium) and imaged with the wide field microscope setup described in Behera et al. (2018) (Supplemental Materials and Methods). After 10-min incubation, the root tip cells of the seedlings were imaged while keeping them in continuous perfusion before a 3-min pulse treatment with 10 μM of the synthetic auxin 1-naphthylacetic acid (NAA). For each experiment, the cpVenus/ECFP ratio for YC-Nano 65 and single fluorescence changes for R-GECO1 were measured over 20 min. The data were plotted over time as both normalized (ΔR/R₀ and ΔF/F₀) or raw cpVenus/ECFP ratios and R-GECO1 fluorescence (Figure 3, C, D, G, and H). At the end of the experiment, four sets of data could be compared: (i) YC-Nano 65 seedlings treated with 10 μM NAA in 0-calcium; (ii) YC-Nano 65 seedlings treated with 10 μM NAA in 1.5 mM calcium; (iii) R-GECO1 seedlings treated with 10 μM NAA in 0-calcium, and (iv) R-GECO1 seedlings treated with 10 μM NAA in 1.5 mM calcium. The results showed that NAA induced an increase in the cpVenus/ECFP ratio and R-GECO1 fluorescence that corresponds to a [Ca²⁺]_cyt increase (Figure 3, C and D), as previously demonstrated (Behera et al., 2018). Interestingly, by comparing the maximum change of normalized cpVenus/ECFP ratio and R-GECO1 fluorescence, it was clear that the [Ca²⁺]_cyt increase was higher in the medium with 1.5 mM calcium than the one observed with 0-calcium (Figure 3, E and F). Thus, both sensors reported the same result, suggesting that the amplitude of the NAA-induced cytosolic Ca²⁺ increase depends on the availability of Ca²⁺ in the medium. However, comparison of the raw cpVenus/ECFP ratio and non-normalized R-GECO1 fluorescence changes provided a different result (Figure 3, G and H). In fact, the incubation of seedlings in 0-calcium had the opposite effects in the two sensor lines. On the one hand, the YC-Nano 65 showed a higher ratio at resting (Figure 3, A and G) with a relatively smaller differential response to NAA treatment as evidenced by the normalized data (Figure 3, E), despite the raw maximum signal being higher than in 1.5 mM calcium (Figure 3, I). On the other hand, the resting R-GECO1 fluorescence (Figure 3, B) was lower in 0-calcium compared with 1.5 mM calcium with corresponding reductions in NAA response (Figure 3, F and J). In conclusion, both sensors reported a decreased NAA response, when Ca²⁺ was removed from the media, but with different starting points, de facto with opposite results.

Since it has been reported that R-GECO1 may suffer from a pH sensitivity (Zhao et al., 2011; Keinath et al., 2015), we measured the cytosolic pH in root tip cells of Arabidopsis seedlings using the ratiometric pH sensor pH-GFP (Moseyko and Feldman, 2001; Behera et al., 2018) in response to 0-calcium treatment (Figure 4). The experiment clearly showed that shifting from 1.5 mM calcium to 0-calcium led to a cytosolic acidification (Figure 4, A and B). Thus, the incubation with 0-calcium can have multiple effects, including changing the cytosolic pH. Such acidification may quench the R-GECO1 fluorescence, reducing its dynamic range. On the other hand, even considering that the pH can also affect the cpVenus and ECFP fluorescence of the YC-Nano 65 Cameleon (Supplemental Figure S1) the ratio calculation seems to better correct for this potential issue as previously shown for the YC3.6 (Nagai et al., 2004; Behera et al., 2018). The understanding of how 0-calcium treatment affects both cytosolic Ca²⁺ and pH is currently under investigation. Nevertheless, our example illustrates that different treatments can potentially alter the properties of the sensors and warranting a careful comparison among them.

Figure 4.

Cytosolic pH dynamics in root tip cells in response to 0-calcium treatment. A, pH dynamics in pH-GFP root tip cells of the region indicated in the schematic drawing, treated with medium with 0-calcium (from 1.5 mM calcium to 0-calcium) as indicated by the black box on the x-axis. The ratio corresponds to the emission of the pH-GFP sensor when excited with light at 405 nm divided by the emission of the sensor when excited at 488 nm. B, The amplitude of 405/488 ratio in 1.5 mM calcium and 0-calcium. n = ≥5. The shaded left-side arrow indicates the direction of the pH decrease. Data were plotted as box-and-whisker plots using GraphPad, in which all the experimental points are plotted, and their distribution represented as a box that extends from the 25th to 75th percentiles. The line in the middle of the box is plotted at the median. P values were calculated with an unpaired Student’s t test. Error bars = sd. ***P ≤ 0.0005 (t test).

This series of experiments demonstrate that intensiometric sensors can be prone to artifacts and that whereas they show superior sensitivity compared with FRET sensors (Krebs et al., 2012; Keinath et al., 2015; Waadt et al., 2017; Resentini et al., 2021a), they might lead to a wrong interpretation of the results. This fact does not signify that intensiometric sensors are not reliable, but that depending on the type of experiment being done, some additional controls should be carried out. In light of this, the use of the recently developed transcriptionally linked dual sensors, such as the CapHensor which allows the simultaneous analysis of Ca²⁺ and pH dynamics (Li et al., 2021) or the R-GECO1-GSL-E²GFP to simultaneously monitor Ca²⁺, H⁺, and Cl^– (Waadt et al., 2020) could be an option. Of particular note, the development of different dual sensor reporters to simultaneously monitor the dynamics of Ca²⁺ with other ions, analytes, or redox potentials of different redox couples will represent an important tool to reach an integrated picture of the physiological response to different stimuli (Wagner et al., 2019; Waadt et al., 2020).

Despite the reservations made on the pH sensitivity of single FP GECIs, to correct for variation in expression levels, in frame, or nested fusions of single FP GECI with other FPs have been developed (Ast et al., 2017; Waadt et al., 2017, 2020). Specifically, Waadt et al. (2017) generated a construct where the mTurquoise FP (Goedhart et al., 2010) was introduced in a vector harboring the R-GECO1 (Table 2). Since mTurquoise fluorescence emission is independent of a change in [Ca²⁺], this strategy allowed the authors to use it as a reference (Waadt et al., 2017). The ratiometric R-GECO1-mTurquoise allowed the authors to reveal detailed maps of [Ca²⁺] changes in response to auxin and ATP in root tip cells (Waadt et al., 2017). A different approach aimed at generating a ratiometric GCaMP sensor was followed by the Frommer’s group, which adopted a strategy based on the employment of a single FP-cassette that nests a stable reference FP (large Stokes shift LSSmOrange) within the GCaMP6s reporter that was brought to the generation of the MatryoshCaMP6s sensor (Table 2;Ast et al., 2017). The main difference between the two approaches is that the “MatryoshCaMP6s sensor” requires a single excitation with a dual emission, instead of the two excitations and two emissions required for the R-GECO1 and mTurquoise, respectively. Indeed, similarly to Cameleon, the use of a second FP may limit the use of the Ca²⁺ sensor with other markers or biosensors. The availability of the R-GECO1-mTurquoise ratiometric sensor or a similar design for a GCaMPs-mCherry (Waadt et al., 2017) can represent a good compromise for the correction of artifacts, potentially also reducing the pH sensitivity. However, the need for dual excitation and dual emission may still limit their use: whereas the MatryoshCaMP6s has again an increased size that could introduce other limitations. Alternatively, dual excitation single FP GECIs could be used, such as GEX-GECO and REX-GECO (Zhao et al., 2011; Wu et al., 2014). Unfortunately, the latter are relatively dim and need further optimization. Interestingly, a kinase activity sensor based on GCaMP3 displayed two excitation peaks with a similar sensitivity to pH changes between 5.6 and 10 (Mehta et al., 2018), suggesting that the ratiometric excitation imaging of this kinase sensor is relatively insensitive to pH changes. Therefore, it seems possible to engineer an excitation ratiometric GECI for normalization to expression levels and pH changes, based on the bright single FP GCaMP3. A reasonably up-to-date database of currently available genetically encoded sensors can be found at BiosensorDB.ucsd.edu.

Microscopy techniques for plant calcium imaging

Traditionally, Ca²⁺ imaging has been carried out by using standard wide field and confocal fluorescence microscopy, while imaging of aequorin has been usually performed with back-illuminated charge-coupled device (CCD) cameras equipped with a light-tight box. The first Ca²⁺ analyses performed in single cells, like stomatal guard cells, root hairs, and pollen tubes, were carried out with microscopes using high magnification objectives. However, when Ca²⁺-sensitive dyes were the only available sensors it was difficult to image more than one cell per experiment (e.g. McAinsh et al., 1995; Ehrhardt et al., 1996; Holdaway-Clarke et al., 1997; Garcia-Mata et al., 2003). Instead, using transgenic plants, stably expressing GECIs, it was possible to measure Ca²⁺ dynamics both at single cell level (Allen et al., 1999; Monshausen et al., 2008; Loro et al., 2012; Hilleary et al., 2020) and in the entire organs or tissues, by decreasing the magnification (Kiegle et al., 2000; Krebs et al., 2012; Zhu et al., 2013; Kiep et al., 2015; Yuan et al., 2014; Jiang et al., 2019; Chen et al., 2020; Hilleary et al., 2020; Wu et al., 2020; Ishka et al., 2021).

To perform Cameleon-based Ca²⁺ imaging, a wide field microscope with one filter for the excitation and one for the emission are not sufficient, whereas this configuration is suitable for a single FP GECI. The use of Cameleon requires the detection of the emissions from two FPs obtained by performing a quick change of filters with a filter wheel that therefore needs repetition of the excitation step. An alternative configuration exploits a beam splitter coupled to one or two cameras to simultaneously acquire the two fluorescence signals (e.g. cpVenus and ECFP). The use of a beam splitter configuration for Cameleon detection was recently reported with a stereomicroscope, allowing the performance of a FRET-based analysis in entire Arabidopsis cotyledons and young leaves expressing the YC-Nano 65 (Hilleary et al., 2020; Ishka et al., 2021). With the wide field microscope, single cell resolution can be obtained when imaging guard cells from an epidermal strip preparation or in vitro germinated pollen tubes, but the lack of optical sectioning hinders the analysis of Ca²⁺ dynamics at single cell level in an entire organ (Table 4). Nevertheless, the use of tissue-specific promoters to guide the expression of the sensor in a given cell type allows the performing of single cell analysis in intact leaves by means of wide field microscopy (Yang et al., 2008). However, in cases where the sensor is ubiquitously expressed, to obtain single cell resolution an optical sectioning microscope is needed. The most widely used optical sectioning technique is the confocal microscope, which makes use of a pinhole to reject the light coming from sample positions that are not in focus. In this configuration, a single laser (445 or 458 nm to excite the acceptor) and two photomultipliers are sufficient for FRET imaging, allowing many plant-biology laboratories to perform Ca²⁺ imaging experiments at high spatial resolution (Tanaka et al., 2010; Loro et al., 2012; Costa et al., 2013; Krebs and Schumacher, 2013; Choi et al., 2014). When required to perform Ca²⁺ imaging analysis at a high temporal resolution, the laser scanning confocal microscope shows some limitations in terms of acquisition speed (Table 4). The sampling is in the order of 1 s, with a traditional galvo scanner. As an example, with a pixel number of 1024 × 1024, a typical acquisition frequency is 0.5 Hz, whereas with a half resolution of 512 × 512 it is 1 Hz. For a higher sampling rate, a spinning disk confocal is a possible choice, as it essentially parallelizes the confocal pinhole detection (Table 4). However, in this case, the need for simultaneous acquisition of both fluorescent ECFP and cpVenus emissions as well as the need for a dedicated laser (i.e. 445 nm) increases the complexity and the price of the microscope setup (Table 4). To gain access to a high resolution, high speed FRET acquisition, we developed a tailored fluorescent light sheet fluorescence microscope that allowed us to perform single cell imaging with a fast rate of acquisition as well as long-term developmental analyses (Costa et al., 2013; Candeo et al., 2017; Romano Armada et al., 2019; Alfieri et al., 2020; Table 4). To do experiments with plants expressing single FP GECIs, the same microscopes described above can be used with standard configurations for a single excitation and a single emission, de facto expanding the audience of possible users. The more recent generation of single FP GECIs has largely facilitated Ca²⁺ imaging analyses since relatively low-cost cameras with good sensitivity exist, which can provide reliable measurements. This is shown in recent papers demonstrating that stereomicroscopes equipped with standard CCD or complementary metal–oxide–semiconductor (CMOS) cameras and common light sources like fluorescent lamps or LED illumination systems are suitable to perform analyses of Ca²⁺ dynamics in entire adult plants (Vincent et al., 2017; Nguyen et al., 2018; Toyota et al., 2018; Shao et al., 2020; Suda et al., 2020; Resentini et al., 2021a) whereas, traditionally, the Ca²⁺ analyses were mainly confined to root cells of Arabidopsis, rice seedlings or a limited number of leaf epidermal cells (Krebs et al., 2012; Loro et al., 2012; Behera et al., 2015; Keinath et al., 2015; Waadt et al., 2017; Table 4).

Table 4.

Summary of advantages and limitations of different microscopy systems suitable for Ca²⁺ imaging

Type of microscopy	Advantages	Limitations
Wide field microscopy	(i) Suitable for its use with FRET-based and single FP GECI; (ii) single cell resolution allowed with specific sample preparations (e.g. Guard cells in epidermal strips).	(i) Lack of optical sectioning.
Laser scanning confocal microscopy	(i) Optical sectioning allowed; (ii) spectral separation with multiple detectors; (iii) single cell analyses allowed in entire organs; (iv) most of the commercial configurations allow FRET analyses.	(i) Acquisition speed.
Spinning disk confocal microscopy	(i) Fast rate of acquisition; (ii) optical sectioning allowed; (iii) single cell analyses allowed in entire organs; (iv) lower phototoxicity.	(i) High price.
Light sheet fluorescence microscopy	(i) Fast rate of acquisition; (ii) optical sectioning allowed; (iii) single cell analyses allowed in entire organs; (iv) long-term analyses; (v) tailored design.	(i) Low versatility.

Future developments

Here, we take the opportunity to foresee possible future developments. Whereas single FP GECIs may be prone to artifacts we still foresee an extensive use of them by harnessing their unique properties. As an example, having GECIs emitting at different wavelengths allows one to study Ca²⁺ dynamics in different subcellular compartments, within the same cell. This has been largely carried out in animal cells, but the applications in plants are still limited (Greenwald et al., 2018; Kelner et al., 2018; Resentini, 2021a). We anticipate that we could express GCaMP sensors, in mitochondria or in the ER together with the R-GECO1 localized to the cytosol or other combinations. By producing plasmids that harbor both sensors in the same backbone (Waadt et al., 2020; Li et al., 2021) we can transform different (single or multiple) mutants to identify and study the Ca²⁺ transport mechanisms in the different membranes (e.g. Wagner et al., 2015; Corso et al., 2018; Kelner et al., 2018). R-GECO1 or other red shifted variants (e.g. K-GECO1 or R-CaMPs; Shen et al., 2018) can be expressed together with Cameleon or other FRET-based GECIs. To perform simultaneous Ca²⁺ imaging analyses in different tissues, different GECIs can be expressed in the same plant under the control of different tissue-specific promoters (Ngo et al., 2014).

At the same time, optical imaging techniques could play a relevant role in the study of Ca²⁺ dynamics at different spatial scales. Super-resolution microscopy, the field of optical imaging that goes beyond the diffraction limit, is having an increased rate of application in plant biology (Komis et al., 2015; Shaw et al., 2019) even if its use in Ca²⁺ studies is still limited. The use of high-speed spinning disk confocal and light sheet fluorescence microscopy allows one to image the plant at single cell or even at a single organelle level. We expect future adoption of these techniques aimed at further increasing the spatio-temporal resolution of the analyses of Ca²⁺ dynamics. A high-end microscopy approach that is also used to monitor FRET-based sensors involves the measurement of fluorescent lifetimes of the donor molecule. This requires a pulsed laser, coupled to a very fast, high resolution photon counting device to determine the time of fluorescence emission of every molecule after excitation. A great advantage of this approach is that the readout is independent of expression levels and that the combination with dark acceptors allows for multiplex analysis. However, given that many photons need to be counted, there are important limitations to the speed by which the imaging can proceed. Current technologies are starting to achieve the required speed (fast-FRET-FLIM) for imaging fast processes such as Ca²⁺ signaling and will be instrumental for multiplexed imaging of a variety of sensors.

At a larger spatial scale, plants expressing two or more sensors simultaneously, with non-overlapping spectra, can be analyzed with low magnification (Resentini et al., 2021a), for example with a stereomicroscope equipped with an automatized filters wheel and multiple LED light sources. A low magnification approach is particularly well suited for imaging plants in close-to-physiological conditions. However, one must bear in mind that when dealing with large plant organs, one major limitation of optical imaging is the diffusion of light within them. Photoacoustic imaging which combines ultrasonic resolution with high contrast and specificity of light (Xu and Wang, 2006) could solve this problem. In photoacoustic imaging, short laser pulses are used to generate megahertz ultrasound waves in-depth into the tissue, usually referred to as photoacoustic, optoacoustic, or thermoacoustic signals. Dyes, GECIs, and metallochromic Ca²⁺ indicators (Dean-Ben et al., 2017; Liu et al., 2019; Dana et al., 2016; Roberts et al., 2018) have been proposed for photoacoustic imaging and imaging of the neural dynamics has been demonstrated in the mammalian brain, at depths superior to fluorescence microscopy (Gottschalk et al., 2019). Photoacoustic technology has not yet been used, to our knowledge, for volumetric Ca²⁺ imaging in plant biology. However, considering that photoacoustic imaging offers possibilities such as pigment identification (Tserevelakis et al., 2016) and chemoselective imaging (Zeng et al., 2019), it could potentially be a key technology in the years to come. The development of additional imaging technologies should allow plants to be kept in their pots without being touched or perturbed until the day of the experiment.

As a further development, additional genetic screens could be designed. As an example, the mutagenesis of plant lines expressing different GECIs in different cellular locales could be developed. The generated mutants could be screened with a fluorescent plate reader (Wagner et al., 2019) to identify those mutants impaired in the Ca²⁺ accumulation in a given compartment.

Conclusions

We are facing a revolution in terms of imaging that is based on the availability of tools that will permit us to really use this technology to increase our knowledge about how plants cope with and adapt to a changing environment. We are now getting closer to perform imaging experiments in real physiological conditions and not being limited to the use of a specimen mounted on a microscope slide. Imaging technologies are non-destructive and can be used to extract information from the plants at a whole-plant resolution, from one organ or organelle to another within the same plant (Lew et al., 2020).

Advances

• GECIs have been efficiently employed in plants.

• In vivo Ca²⁺ imaging at high spatial and temporal resolution is possible at tissue, cell, and organelle levels.

• Plants expressing GECIs with different affinities for Ca²⁺ are available for cytosolic and organellar analyses.

• Ratiometric GECIs allow a reliable comparison of Ca²⁺ levels in different genetic backgrounds and in long-term treatments.

• Intensiometric GECIs require simple and cost-effective imaging setups.

Outstanding questions

• Can GECIs be used for in vivo analyses of Ca²⁺ dynamics in real physiological conditions?

• Can Ca²⁺ imaging technology in model plants translate to crops?

• Can GECIs be used as environmental biosensors?

• Can we design genetic screening based on recent generation GECIs to identify additional components of Ca²⁺ signaling?

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Single wavelength emissions as arbitrary units of fluorescence of cpVenus (yellow trace) and ECFP (light blue trace) of the YC-Nano 65 used for the ratio calculations of Figure 3, C and G.

 

M.G. carried out the experiments reported in Figures 3 and 4 and prepared Figures 1, 3, and 4. F.R. prepared Tables 1 and 4. A.C. conceived the project and wrote the article with feedback from all authors. A.C. serves as the author responsible for contact and ensures communication.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/General-Instructions) is Alex Costa (alex.costa@unimi.it).