Dynamic visualization of calcium-dependent signaling in cellular microdomains

Abstract

Cells rely on the coordinated action of diverse signaling molecules to sense, interpret, and respond to their highly dynamic external environment. To ensure the specific and robust flow of information, signaling molecules are often spatially organized to form distinct signaling compartments, and our understanding of the molecular mechanisms that guide intracellular signaling hinges on the ability to directly probe signaling events within these cellular microdomains. Ca²⁺ signaling in particular owes much of its functional versatility to this type of exquisite spatial regulation. As discussed below, a number of methods have been developed to investigate the mechanistic and functional implications of microdomains of Ca²⁺ signaling, ranging from the application of Ca²⁺ buffers to the direct and targeted visualization of Ca²⁺ signaling microdomains using genetically encoded fluorescent reporters.

1. Introduction

All living cells must continually sense and respond to changes in their external chemical environment. During intracellular signaling, information regarding the conditions outside the cell is passed along from the cell surface to the appropriate response machinery inside the cell. Yet although cells contain diverse signaling pathways that specifically control the myriad biological processes that are essential to life, the pool of signaling molecules that comprise these pathways is limited, and specificity cannot be intrinsically encoded into individual pathways. Rather, in order to both ensure the specificity and promote the diversity of signaling outcomes, cells must carefully coordinate the actions of signaling molecules as they participate in a dynamic network of highly integrated signaling pathways. One way for cells to achieve this level of coordination is through the spatial compartmentalization of the cell interior into local signaling domains of various sizes, and intracellular Ca²⁺ signaling offers a striking example of this process.

Ca²⁺ signaling regulates many fundamental biological processes, including neurotransmission, muscle contraction, gene expression, cell proliferation, and cell death [1], often regulating multiple cellular processes in parallel. The remarkable versatility of Ca²⁺ as an intracellular messenger stems from the exquisite spatial and temporal regulation of elevations in intracellular Ca²⁺ concentrations, especially through the formation of discrete microdomains of Ca²⁺ signaling. Often, these microdomains involve so-called “elementary” Ca²⁺-release events that result from the opening of individual Ca²⁺ channels [2]. Broadly speaking, however, the term Ca²⁺ microdomain can apply not only to zones of high Ca²⁺ concentration that occur near the mouths of Ca²⁺ channels but to any Ca²⁺ signaling event that is confined to a particular region of the cell – be it the plasma membrane, a specific part of the cytosol, or Ca²⁺-storing organelles – as opposed to the cytoplasm as a whole [3,4].

As with the compartmentalization of other signaling molecules, such as cAMP (see [5]), the existence of Ca²⁺ microdomains was originally proposed to account for experimental observations that lacked a clear mechanistic foundation. Various theoretical studies (e.g., [6,7]) and indirect observations [8] bolstered this idea, but it was not until the development of optical (e.g., fluorescent) techniques to visualize Ca²⁺ dynamics in intact cells that direct experimental evidence of Ca²⁺ microdomains was obtained. Following the adoption of fluorescent probes to study local Ca²⁺ dynamics, the more recent development of genetically encoded fluorescent reporters based on green fluorescent protein (GFP) and related fluorescent proteins has completely revolutionized the study of spatially confined signaling events. In this review, we provide a brief primer on the design and development of genetically encoded fluorescent reporters and discuss the application of these biosensors to the study of spatially compartmentalized signaling events in living cells, using Ca²⁺ signaling as an example.

2. Fluorescent biosensors

Understanding the functional and mechanistic properties of signaling microdomains depends on our ability to assay molecular events specifically within these confined cellular regions. In other words, the more selectively we can target these microdomains, the more precisely we can study local signaling. These efforts have been greatly facilitated by advances in live-cell fluorescence imaging approaches, which have deep ties to the study of Ca²⁺ signaling. For example, the use of fluorescent indicator dyes for live-cell imaging was popularized by the success of Ca²⁺ indicators such as quin-2, fluo-3, and fura-2 (Fig. 1A) [9], and it was the isolation of the Ca²⁺-dependent photoprotein aequorin – itself a useful tool for monitoring Ca²⁺ inside cells (Fig. 1B) [10,11] – that ultimately led to the discovery of Aequoria victoria GFP (reviewed in [12,13]) and the development of genetically encoded fluorescent reporters. These genetically encoded molecular tools can easily be expressed in cells using recombinant DNA technology and can further be targeted throughout the cell via the incorporation of endogenous subcellular targeting motifs or via fusion to endogenous proteins that natively localize to specific regions of the cell. This relatively straightforward targetability enables the specific and detailed investigation of signaling processes occurring within all manner of cellular microdomains and compartments [14,15].

Figure 1. Tools for visualizing signaling in live cells.

(A) Fluorescent Ca²⁺ indicators. (i) The chemical structures of the Ca²⁺ indicators quin-2, fura-2, and fluo-3 in their Ca²⁺-free states. (ii) Fura-2 emits 510-nm light upon excitation at either 340 nm or 380 nm. In its Ca²⁺-free state, fura-2 is preferentially excited at 380 nm, whereas Ca²⁺-bound fura-2 is preferentially excited at 340 nm. The fluorescence intensity at each excitation wavelength is directly proportional to the concentration of Ca²⁺. (B) Aequorin consists of an apo-protein bound to a coelenterazine co-factor (left). The binding of Ca²⁺ (middle) results in the enzymatic conversion of coelenterazine to coelenteramide (right) and the emission of blue (459 nm) light. (C) The FRET-based Ca²⁺ sensor cameleon consists of a tandem fusion of calmodulin (CaM) and the CaM-binding M13 peptide sandwiched between a FRET pair (shown here: CFP and YFP). The binding of Ca²⁺ induces CaM to form a complex with the M13 peptide, thereby causing a conformational rearrangement that increases FRET between the two fluorescent proteins. (D) GCaMP also uses CaM and the M13 peptide as its molecular switch; however, rather than modulating FRET between a pair of fluorescent proteins, in GCaMP, CaM and M13 are inserted into a single fluorescent protein (e.g., GFP), wherein they modulate fluorescence intensity. (E) Genetically encoded fluorescent reporters are a highly versatile set of molecular tools, owing to their highly modular design scheme. The key component is the molecular switch, which can be (i) assembled from protein fragments that function as a “receiver” for the input signal and a “switch” that drives the conformational change or (ii) taken directly from an endogenous protein that undergoes a native conformational change in response to a known stimulus.

Genetically encoded fluorescent reporters come in a variety of forms (reviewed extensively in [16]) that all share a highly generic, modular design in which a sensing unit capable of detecting a specific biochemical signal is coupled to a reporting unit composed of one or more fluorescent proteins. In the most versatile class of biosensors developed thus far, the sensing unit encompasses a protein or protein fragments in the form of a molecular switch that undergoes a conformational change in response to a particular input signal. The conformational change in the molecular switch then alters the distance and/or orientation of a pair of fluorescent proteins capable of undergoing Förster resonance energy transfer (FRET) [17]. The first such FRET-based biosensor, cameleon, was generated by sandwiching the Ca²⁺ sensor calmodulin (CaM) and a CaM-binding peptide from myosin light-chain kinase (M13) between a FRET pair [18,19], again highlighting the close ties between Ca²⁺ signaling and live-cell fluorescence imaging. In the presence of Ca²⁺, CaM binds the M13 peptide and induces a conformational change that alters the FRET signal from cameleon (Fig. 1C). In an alternative design that eschews FRET, CaM and M13 are instead inserted within a single fluorescent protein and used to directly control the fluorescence intensity, as in the popular GCaMP series of Ca²⁺ sensors (Fig. 1D) [20,21].

Molecular switches have proven to be an extremely versatile solution to the problem of monitoring signaling in living cells, as they are readily adaptable to detecting a variety of biochemical events associated with signaling. As demonstrated with cameleon, molecular switches can often be engineered from protein components that will interact and induce a conformational change in response to a given signal. These bipartite switches comprise a receiving domain that is modified by the input signal and a switching domain that interacts with the receiver to drive the conformational change (Fig. 1E). For example, numerous kinase activity reporters have been generated by fusing a consensus substrate sequence and a phosphoamino acid-binding domain (PAABD), which are then sandwiched between a FRET pair [22]. Binding of the phosphorylated substrate by the PAABD leads to a conformational change that produces a FRET change. Similarly, fusing a small GTPase to a binding partner that only interacts with the active, GTP-bound form of the enzyme results in a molecular switch that underlies many GTPase activation sensors [23]. Molecular switches can also be derived from the intrinsic conformational dynamics of native proteins (Fig. 1E). The most commonly used cAMP biosensors, for instance, use the intrinsic conformational change that occurs when cAMP binds exchange protein activated by cAMP (Epac) to drive a FRET change [24,25]. In theory, any protein whose conformation is directly modulated by an upstream signal can be inserted between a FRET pair to construct a biosensor, though this frequently involves trial and error. This generalizable design scheme has inspired the development of a multitude of biosensors capable of detecting a broad spectrum of biochemical processes (e.g., second messenger production, ion and metabolite concentrations, enzyme activity, and enzyme activation), making genetically encoded fluorescent reporters an invaluable asset in the study of intracellular signaling [15,16].

3. Investigating microdomains of Ca²⁺ signaling

Many approaches have been employed over the years to investigate local Ca²⁺ signaling processes within cells. These range from indirect measurements based on Ca²⁺ buffering to highly selective, direct visualization of Ca²⁺ signals using genetically encoded biosensors. As more targeted and versatile methods have become available, the ability to probe Ca²⁺ signaling has also expanded to include not only Ca²⁺ itself but also a number of Ca²⁺-dependent signaling proteins, thus fueling the development of a much more sophisticated understanding of Ca²⁺ signaling microdomains.

3.1 Ca²⁺ microdomains at the plasma membrane

Ca²⁺ influx across the plasma membrane is an essential signaling mechanism in all cell types, especially in electrically excitable cells such as neurons and muscle cells. Hence, early studies of local Ca²⁺ signaling often focused on Ca²⁺ microdomains near the plasma membrane. Plasma membrane Ca²⁺ microdomains were initially proposed to help explain the very rapid release of neurotransmitter in response to Ca²⁺ influx, under the assumption that the exocytotic machinery must be closely coupled to Ca²⁺ channels [26]. The existence of these microdomains could be observed indirectly by loading cells with different chelators to buffer global Ca²⁺ increases [8]. Typically, cells are loaded with either the “slow” chelator EGTA or the “fast” chelator BAPTA, the idea being that given a very short (<100 nm) distance between the Ca²⁺ channel and the target, only BAPTA will be able to capture Ca²⁺ quickly enough to disrupt the microdomain (Fig. 2A) [8,27]. This is a powerful technique that is still used to study Ca²⁺ microdomains in the vicinity of plasma membrane Ca²⁺ channels to this day. For example, both BAPTA and EGTA were recently shown to disrupt transmission in mature hippocampal mossy fiber synapses, suggesting that synaptic vesicles are only loosely coupled to Ca²⁺ channels in these plastic synapses, in contrast to the tight coupling that is often observed in the mature nervous system [28]. In addition, Selway et al. recently used cell-permeable versions of these chelators to demonstrate that Ca²⁺ microdomains formed by L-type voltage-gated Ca²⁺ channels (VGCCs) were sufficient to activate ERK signaling in response to GLP-1 stimulation in MIN6 pancreatic β-cells [29].

Figure 2. Dynamically probing microdomains of Ca²⁺ signaling.

(A) The extent to which different Ca²⁺ chelators are able to disrupt downstream signaling can provide information about how closely linked a given process is to Ca²⁺ microdomains. As shown here, if a Ca²⁺ target (e.g., synaptic vesicle) is located within a Ca²⁺ microdomain (“tight” coupling), EGTA will not be able to chelate Ca²⁺ before it reaches the target. Only a target that is further away (“loose” coupling) will be affected. BAPTA binds Ca²⁺ far more quickly, however, and is able to chelate Ca²⁺ rapidly enough to prevent it from reaching even the tightly coupled target. (B) In some cases, Ca²⁺ gradients can be directly visualized using diffusible probes. For example, in the squid giant synapse (left), electrical stimulation induces localized Ca²⁺ influx in the pre-synaptic fiber (middle). This can be visualized by loading the axon with a diffusible Ca²⁺ indicator; the cytosolic Ca²⁺ gradient appears as a clearly visible gradient in the fluorescent signal from the indicator (right) due to the relatively large size of the cells (as in [33]). (C) Ca²⁺ signaling compartments can also be directly visualized using genetically encoded biosensors (generically illustrated here in the form of cameleon) targeted to various subcellular locations. Targeting involves appending a specific DNA sequence that encodes an endogenous localization signal onto the DNA sequence of the reporter. For example, these sequences can be used to tether a Ca²⁺ biosensor to Ca²⁺ channels (e.g., L-type Ca²⁺ channel [44] or ryanodine receptor [51]) to probe discrete Ca²⁺ release domains. Similarly, Ca²⁺ sensors can be targeted to the interior or exterior of various organelles, such as the ER lumen to monitor store heterogeneity during Ca²⁺ release [55] or to the mitochondrial inner and outer membranes to investigate Ca²⁺ uptake by mitochondria [3,66–68].

Detailed studies of Ca²⁺ microdomains, however, require more direct methods for monitoring local Ca²⁺ signaling events in cells. A number of optical detection methods have been developed to permit the direct visualization of intracellular Ca²⁺ dynamics in living cells, and many of these had already begun to see widespread adoption by the late 1980s [30]. In many cases, these diffusible probes can be introduced into cells and used to directly image discrete concentration changes that are associated with Ca²⁺ microdomains located at the plasma membrane. Indicator dyes such as fura-2 and fluo-3, for example, could often be used to resolve minute gradients of Ca²⁺ influx [31–33]. Along with fluorescent indicators, the bioluminescent protein aequorin, which has been used to study Ca²⁺ signaling in live cells for nearly half a century [34], has also been used to visualize Ca²⁺ microdomains. In particular, injecting cells with a low-sensitivity derivative of aequorin (n-aequorin-J) enabled the detection of discrete Ca²⁺ blips, or quantum emission domains (QEDs) [26,35–37], which are highly localized sites of Ca²⁺ influx that are attributed to the opening of a small number of plasma membrane Ca²⁺ channels. Such elementary Ca²⁺ release events have since taken on many names to reflect the various cellular contexts in which they have been described [2,38]. More recently, the fluorescent Ca²⁺ indicator fluo-4 was used in combination with total internal reflection fluorescence (TIRF) microscopy to visualize elementary Ca²⁺ release events associated with the vanilloid transient receptor potential (TRP) channel TRPV4 in endothelial cells [39], while Sonkusare et al. performed confocal imaging of diffusible GCaMP2 to visualize TRPV4-dependent Ca²⁺ sparklets and investigate the Ca²⁺-dependent regulation of smooth muscle tone [40].

Another essential consideration in the study of signaling microdomains is selectivity. In the preceding examples, Ca²⁺ microdomains were observed by imaging particular cells in which gradients can be clearly resolved (e.g., giant squid axons) (Fig. 2B) or with the aid of selective illumination to distinguish local Ca²⁺ signals from the diffuse background. Alternatively, with n-aequorin-J selectivity is achieved via reduced Ca²⁺ sensitivity compared with unmodified aequorin [26,34,35]. This reduced sensitivity was advantageous for visualizing Ca²⁺ specifically within microdomains, as these confined regions were predicted to contain very high Ca²⁺ concentrations [35,41]; using n-aequorin-J ensured that only these high-concentration Ca²⁺ compartments will be visualized. Even greater selectivity can be achieved by physically targeting a probe to the specific compartment under investigation. For example, expressing aequorin directly in cells as a chimeric fusion with various targeting sequences enabled the selective detection of Ca²⁺ microdomains in the vicinity of the plasma membrane [42], as well as the first direct visualization of the local Ca²⁺ dynamics associated with intracellular organelles (see Section 3.2). Subcellular targeting is also one of the major hallmarks of genetically encoded fluorescent reporters. In an early example, Emmanouilidou et al. targeted the probe yellow cameleon to the secretory vesicle surface and observed a subset of vesicles within 1 °m of the plasma membrane that experienced significantly higher Ca²⁺ elevations than did more distant vesicles or the overall cytoplasm [43]. Genetically encoded fluorescent reporters have even been localized directly to the mouths of Ca²⁺ channels, as was done recently by Tay and colleagues with the VGCC Ca_v2.2 [44], or to distinct plasma membrane subdomains such as caveolae [45], thereby offering an exquisitely detailed look at Ca²⁺ microdomains that cannot be resolved using more traditional approaches.

3.2 Local Ca²⁺ domains within the cell

Resting cytosolic Ca²⁺ concentrations are kept low by the active extrusion of Ca²⁺ across the plasma membrane as well as by the uptake of Ca²⁺ into intracellular organelles, which act as internal Ca²⁺ stores that can be called upon to release Ca²⁺ to regulate various cellular processes [4,46]. Internal Ca²⁺ stores thus represent another domain of local Ca²⁺ signaling, in terms of both localized Ca²⁺ release from organelles and Ca²⁺ handling within organelles, that have long been the subject of intense scrutiny. The endoplasmic/sarcoplasmic reticulum (ER/SR) in particular often comprises the largest intracellular Ca²⁺ store and serves as a major source of cytosolic Ca²⁺ release and signaling in non-excitable cell types as well as in excitable cells, where it plays essential roles in, for example, muscle contraction. Diffusible Ca²⁺ indicators have been used to image microdomains of Ca²⁺ release from the ER in combination with pharmacological agents known to specifically promote ER Ca²⁺ release, such as caffeine [31,47] or histamine [48]. The inositol (1,4,5)-triphosphate (IP₃) receptor (IP₃R) can also be directly activated by the application of IP₃ or a caged derivative [49]. In an elegant example of this technique, Smith and Parker used TIRF imaging to visualize single-channel flux through IP₃Rs in fluo-4-loaded SH-SY5Y neuroblastoma cells upon photo-uncaging of IP₃ [50]. Notably, the cells were also loaded with EGTA to prevent the build-up of Ca²⁺ waves while simultaneously preserving elementary Ca²⁺ blips, a twist on the Ca²⁺-buffering approach discussed in Section 3.1

Targeted approaches using genetically encoded reporters are also apt for probing ER Ca²⁺ release microdomains. For example, Despa et al. recently targeted GCaMP to the junctional cleft in cardiac myocytes in order to directly measure local Ca²⁺ dynamics in this functionally important microdomain [51]. The authors ensured the correct localization of their probe by fusing it to FKBP12.6, which tightly and selectively binds the cardiac ryanodine receptor (RyR2) [52,53]. Yet where targeted approaches truly shine is in examining Ca²⁺ dynamics within organelles, which represents a significant technical challenge to diffusible probes such as fluorescent indicator dyes (Fig. 2C). However, it is important to make sure that the sensitivity of the probe is sufficiently tuned so as to accurately report high Ca²⁺ concentrations within stores. Montero et al. targeted a low-affinity aequorin to the ER lumen to measure Ca²⁺ accumulation, yet even this probe was readily saturated by the high ER Ca²⁺ concentration ([Ca²⁺]_ER). Instead, the authors used Sr²⁺ as a Ca²⁺ surrogate in their measurements [54]. Recently, another low-affinity aequorin mutant was used to study ER Ca²⁺ homeostasis in detail, revealing significant heterogeneity in [Ca²⁺]_ER, as well as high-concentration areas that were unaffected by Ca²⁺-release agents. Buffering the cytosol with either EGTA or BAPTA also revealed that Ca²⁺ microdomains play a crucial role in store refilling [55]. The modified cameleon D1ER also exhibits a lower Ca²⁺-binding affinity that is suitable for probing [Ca²⁺]_ER [56] and was recently used by the Delbono group to show that physiological stimuli do not induce significant SR Ca²⁺ depletion in skeletal muscle fibers [57]. This same group also used a novel Ca²⁺ probe, CatchER, that consists of a directly Ca²⁺-sensitive GFP [58] to show that this residual SR Ca²⁺ is higher in older mice than in younger mice, providing evidence for excitation-contraction uncoupling in aging muscle [59].

Alongside the ER, numerous other organelles are attracting interest regarding their roles in shaping local Ca²⁺ signaling domains. Recent studies have used cameleon-based fluorescent biosensors to study Ca²⁺ dynamics within both the trans- and medial-Golgi [60,61], as well as in lysosomes [62]. This ability to target probes to specific subcellular locations means investigators are largely unrestrained in their efforts to study these domains in minute detail. For example, mitochondrial Ca²⁺ uptake has been shown to modulate cytosolic Ca²⁺ signals [63,64], and the expression of a COX8-aequorin fusion localized to the inner mitochondrial membrane revealed that Ca²⁺ concentrations in the mitochondrial matrix ([Ca²⁺]_m) change rapidly in response to cytosolic Ca²⁺ signals (Fig. 2C) [65]. Further investigations using matrix-targeted aequorin indicated that [Ca²⁺]_m increased in response to IP₃-mediated ER Ca²⁺ release but was unaffected by diffuse Ca²⁺ elevations of a similar magnitude, thus implicating ER Ca²⁺ microdomains in mitochondrial uptake [66]. Labeling the ER and mitochondria with distinctly colored fluorescent proteins highlighted the close physical association between these two organelles, and aequorin targeted to the mitochondrial intermembrane space showed that the opening of IP₃Rs exposes mitochondria to very high local Ca²⁺ concentrations and even hinted at the existence of Ca²⁺ hotspots on the exterior of the mitochondria [67]. Recently, fluorescence imaging of an enhanced cameleon probe (D1cpv) targeted to the cytosolic face of the outer mitochondrial membrane was combined with detailed, pixel-by-pixel analyses to directly visualize these Ca²⁺ hotspots, in which Ca²⁺ concentrations were up to 10 times higher compared with the bulk cytosol [68].

3.3 Compartmentalized signaling by Ca²⁺ targets

Ca²⁺ acts by modulating the activities of numerous Ca²⁺-sensitive enzymes and regulatory proteins that participate in various cellular processes. The Ca²⁺ signaling machinery also engages in extensive interactions with the components of other intracellular signaling pathways. Hence, building a complete picture of the role of cellular microdomains in regulating Ca²⁺ signaling also requires studying the spatial regulation of Ca²⁺-dependent signaling molecules and investigating how Ca²⁺ signaling microdomains impinge on other signaling pathways. It is well known, for example, that Ca²⁺ plays a major role in regulating cAMP-dependent signaling pathways, and vice versa [69–71]. In particular, Ca²⁺ signals can both stimulate and inhibit the production of cAMP by modulating the activities of a subset of adenylyl cyclase (AC) isoforms [72]. These Ca²⁺-sensitive ACs have been shown to respond almost exclusively to Ca²⁺ signals generated during capacitative Ca²⁺ entry (CCE; also known as store-operated Ca²⁺ entry, or SOCE) [73,74], suggesting that these enzymes might be specifically responding to CCE-induced Ca²⁺ microdomains. Nakahashi and colleagues were able to test this idea directly by fusing aequorin to the C-terminus of the Ca²⁺-inhibited AC isoform AC5 and comparing the response from this probe with that of cytosolic aequorin under various Ca²⁺-elevating conditions [75]. Whereas AC5-aequorin was less responsive to general Ca²⁺ release from intracellular stores than cytosolic aequorin, the targeted probe reported much higher Ca²⁺ concentrations in response to the induction of CCE than were detected in the cytosol. More recently, Willoughby et al. performed similar experiments by fusing GCaMP to either AC8, a Ca²⁺-stimulated AC, or AC2, which is Ca²⁺ insensitive [76]. The authors confirmed that AC8 activity is specifically stimulated by CCE and found that AC8-GCaMP specifically sensed Ca²⁺ from CCE, being virtually insensitive to general Ca²⁺ release. On the other hand, the AC2-GCaMP response mirrored that of GCaMP expressed throughout the plasma membrane or in the cytosol. In fact, so close is the coupling between ER stores and cAMP that store depletion alone can activate cAMP signaling, independent of cytosolic Ca²⁺ influx: Using cAMP biosensors (see Section 2), Lefkimmiatis and coworkers recently demonstrated that STIM1, which monitors ER Ca²⁺ stores [77–79], directly activates cAMP production upon store depletion [80].

However, although they illuminate an important facet of intracellular signaling, at present, studies of the spatial regulation of Ca²⁺ targets remain somewhat rare, especially compared with the vast body of work that has been built around the characterization of Ca²⁺ microdomains themselves. In most cases, this stems from the need to develop appropriate tools (e.g., genetically encoded biosensors) for directly probing the compartmentalized activities of the Ca²⁺ signaling machinery. Though this is not to say that spatially regulated signaling by Ca²⁺ targets cannot be investigated by other means. For example, the Ca²⁺/CaM-dependent protein phosphatase calcineurin (CaN) is known to physically interact with L-type VGCCs [81–83], and studies have shown that local activation by L-type channels is essential for CaN signaling. In particular, Nieves-Cintrón et al. found that BAPTA treatment was unable to disrupt the local activation of CaN by Ca²⁺ sparklets (visualized using the Ca²⁺ indicator fluo-5F) generated by Ltype VGCCs [84]. Nevertheless, genetically encoded fluorescent reporters remain among the most powerful and versatile means of assaying the spatial patterns of intracellular signaling. Wu et al. recently described a FRET-based biosensor, based on the interaction between CaN and Ca²⁺/CaM, which they used to monitor the spatial dynamics of CaN activation in murine primary cortical neurons treated with oligomeric amyloid-β (Aβ) [85]. This reporter, which exhibits a FRET increase when Ca²⁺/CaM binds and activates CaN, revealed that Aβ treatment rapidly activates CaN localized in dendritic spines, followed within minutes by CaN activation in the cytosol and hours later in the nucleus.

Our group also recently investigated the spatial regulation of CaN signaling in pancreatic β-cells [86]. Insulin secretion by pancreatic β-cells is known to be controlled by oscillatory changes in cytosolic Ca²⁺ concentrations ([Ca²⁺]_c) [87], which we previously linked to the spatial regulation of cAMP-dependent protein kinase (PKA) signaling in these cells [88]. Therefore, we used a FRET-based CaN activity reporter (CaNAR), based on conformational changes associated with the CaN-dependent dephosphorylation of nuclear factor of activated T-cells (NFAT) [89], to examine whether CaN activity in different subcellular regions is differentially regulated by these Ca²⁺ oscillations. Using improved, subcellularly targeted versions of CaNAR, we identified two distinct subcellular “zones” with unique CaN activity patterns: In response to Ca²⁺ oscillations, CaN activity in the cytosol increased in an integrative fashion, with each Ca²⁺ peak leading to a step-increase in CaN activity, whereas CaN activity measured near the ER surface oscillated in tandem with [Ca²⁺]_c (Fig. 3). To elucidate the mechanism underlying these spatially distinct CaN activity patterns, we first explored the role of kinases in antagonizing CaN. Specifically, we found that inhibiting PKA activity dramatically altered ER CaN activity dynamics, causing them to adopt the step-like, integrative pattern seen in the cytosol [86]. However, increasing PKA to saturating levels did not similarly induce an oscillatory CaN activity pattern in the cytosol, suggesting that additional factors were rendering ER CaN activity more susceptible to PKA.

Figure 3. Spatial regulation of CaN signaling in pancreatic β-cells.

(A) A FRET-based CaN activity reporter (CaNAR), which consists of the N-terminal 297 amino acids from NFAT1 sandwiched between CFP and YFP. Dephosphorylation of the reporter by CaN leads to a conformational change, and thus a FRET change. (B) In MIN6 pancreatic β-cells, the FRET response from CaNAR (black curves) reveals that CaN activity increases in a step-like fashion in the cytosol (left), whereas CaN activity appears to oscillate near the ER surface (right), in relation to cytosolic Ca²⁺ oscillations (red curves). (C) A CaN activation ratiometric indicator (CaNARi), which consists of the catalytic subunit of CaN (CNA) sandwiched between CFP and YFP. CaN activation via Ca²⁺/CaM binding induces a conformational change, leading to altered FRET between CFP and YFP. (D) The FRET response from CaNARi (black curves) reveals that CaN exhibits an oscillatory activation pattern both in the cytosol (left) and at the ER surface (right), in contrast to the CaN activity pattern revealed by CaNAR. (E) As described in [86], these data are consistent with a model in which free Ca²⁺/CaM is less abundant near the ER surface, thereby leading to weaker CaN activation and CaN activity oscillations in this compartment. Adapted from [86].

We therefore used another FRET biosensor, based on the Ca²⁺/CaM-induced conformational change that occurs during CaN activation, to test whether CaN activation differs subcellularly in pancreatic β-cells [86]. Interestingly, although we observed that CaN activation was uniformly oscillatory throughout these cells in response to Ca²⁺ oscillations (Fig. 3), our results did suggest that less CaN was being activated near the ER surface. The activity of CaN is tightly regulated by Ca²⁺/CaM, which serves as a central node in the Ca²⁺ signaling network, mediating diverse cellular processes through its many targets. Free Ca²⁺/CaM that is available to transiently interact with targets, rather than being bound as a dedicated subunit, is considered to be a limiting resource in cells, and given its limited diffusibility, free Ca²⁺/CaM may impart another layer of spatial control over Ca²⁺ signaling [90,91]. We therefore used a FRET-based biosensor that specifically detects free Ca²⁺/CaM [92] to test the hypothesis that limiting amounts of Ca²⁺/CaM at the ER surface in β-cells were contributing to weaker CaN activation and thus causing CaN activity oscillations in this compartment. This probe was previously used to compare free Ca²⁺/CaM levels in the cytosol and nucleus in rat basophilic leukemia cells [93], and differences in free Ca²⁺/CaM levels detected using this biosensor have been confirmed to correspond to physiologically meaningful differences in Ca²⁺/CaM target activity [94]. Using this sensor, we found that free Ca²⁺/CaM was indeed less abundant at the ER compared with the cytosol. Furthermore, overexpressing additional CaM not only rescued this difference but also led to integrative ER CaN activity. Conversely, treating the cells with a CaM antagonist also endowed the cytosol with oscillating CaN activity [86].

Diffusible reaction systems form the backbone of spatially organized signaling [95]. The formation of Ca²⁺ microdomains, for instance, relies on the interplay between influx, efflux, and buffering to restrict the diffusion of Ca²⁺ [3]. Our study revealed that Ca²⁺/CaM can also impose a spatial signal atop the transient but global Ca²⁺ rises that accompany cytosolic Ca²⁺ oscillations. Ca²⁺ transients have similarly been shown to produce large local increases in free Ca²⁺/CaM but only minimal increases at distal locations (e.g., the nucleus) due to the limited diffusibility of Ca²⁺/CaM [93]. In fact, long-range Ca²⁺/CaM signaling was recently shown to require a dedicated carrier protein [96]. Seen in this light, it is clear that the diffusion of free Ca²⁺/CaM adds another dimension to the organization of Ca²⁺ signaling microdomains.

4. Concluding remarks

The idea that signaling pathways are spatially compartmentalized is not new, though it emerged in the absence of tools for directly probing signaling compartments in living cells, and we have essentially spent the last few decades catching up. Ca²⁺ signaling is somewhat unique in this regard, as specialized tools for visualizing Ca²⁺ in living cells were developed quite quickly based on existing in vitro techniques. Thus, Ca²⁺ has long been able to serve as a model for the study of compartmentalized signaling. The advent of genetically encoded fluorescent reporters has leveled the playing field, so to speak, and has allowed researchers to move into new realms and study the compartmentalization of more and more signaling processes. We now have access to a powerful and versatile arsenal of tools that can be used to probe signaling microdomains from a variety of angles, from unraveling the molecular mechanisms and biological roles of known microdomains to identifying and characterizing entirely new signaling domains. Along these lines, Matsuda and colleagues recently developed a “caged” FRET-based Ca²⁺ indicator whose fluorescence can be activated at specific locations and times via UV illumination [97], thus enabling the precise visualization of Ca²⁺ signals in cellular microdomains that are not currently accessible by approaches that rely on the use of targeting sequences. Like the first FRET-based Ca²⁺ sensor, cameleon, such novel designs will likely serve as prototypes for the development of new probes for other signaling molecules, thus continuing to advance the study of signaling microdomains into new frontiers.

• We review the application of fluorescent biosensors to study local calcium signaling in live cells