Spatial-temporal patterning of Ca²⁺ signals by the subcellular distribution of IP₃ and IP₃ receptors

Abstract

The patterning of cytosolic Ca²⁺ signals in space and time underlies their ubiquitous ability to specifically regulate numerous cellular processes. Signals mediated by liberation of Ca²⁺ sequestered in the endoplasmic reticulum (ER) through inositol trisphosphate receptor (IP₃R) channels constitute a hierarchy of events; ranging from openings of individual IP₃ channels, through the concerted openings of several clustered IP₃Rs to generate local Ca²⁺ puffs, to global Ca²⁺ waves and oscillations that engulf the entire cell. Here, we review recent progress in elucidating how this hierarchy is shaped by an interplay between the functional gating properties of IP₃Rs and their spatial distribution within the cell. We focus in particular on the subset of IP₃Rs that are organized in stationary clusters and are endowed with the ability to preferentially liberate Ca²⁺.

Graphical abstract

1. Introduction

Calcium is a versatile, universal, intracellular messenger. Changes in cytosolic free calcium concentrations ([Ca²⁺]_cyt) are finely tuned by the interactions of Ca²⁺ pumps, channels and buffering proteins to serve critical roles in numerous cellular functions; controlling processes as diverse as secretion, contraction, neurotransmitter release and transcription [1,2]. The capacity to precisely and specifically regulate cellular events arises in large part through an exquisite control of the spatial and temporal patterning of [Ca²⁺]_cyt transients [1]. This is exemplified by the second messenger pathway mediated by inositol 1,4,5-trisphosphate (IP₃). IP₃ is generated by cleavage of membrane phosphatidylinositol bisphosphate by phospholipase C following stimulation of cell surface G-protein and tyrosine kinase coupled receptors. IP₃ then diffuses in the cytosol to bind to IP₃ receptor/channels (IP₃Rs) in the membrane of the endoplasmic reticulum (ER) that serve to liberate Ca²⁺ sequestered in the ER lumen by the activity of SERCA pumps [3]. The opening of an IP₃R channel requires not only IP₃, but also the binding of Ca²⁺ to activating sites on the cytosolic face of the tetrameric receptor [4,5]. The latter process endows IP₃-mediated signaling with a regenerative property of Ca²⁺-induced Ca²⁺ release (CICR), because Ca²⁺ ions released from one IP₃R channel may diffuse in the cytosol and promote the opening of neighboring channels. The resulting cytosolic Ca²⁺ signals are further shaped by the clustered arrangement of IP₃Rs in the ER to generate a hierarchy of events (Fig. 1), with increasing amounts of IP₃ progressively evoking Ca²⁺ liberation from individual IP₃Rs [6] (Ca²⁺ blips), local Ca²⁺ signals arising from concerted openings of several clustered IP₃Rs [6–10] (puffs), and global Ca²⁺ elevations that engulf the cell [9,11–13].

Figure 1:

The spatial-temporal patterning of local Ca²⁺ signals mediated by IP₃Rs. (A) Panels show fluorescence of a HeLa cell loaded with the Ca²⁺-indicator Cal-520, the slow calcium buffer EGTA, and the caged IP₃ analog, i-IP₃. The first panel shows basal Cal-520 fluorescence of the cell immediately prior to the UV flash and the subsequent panels show sequential images at half second intervals following photorelease of i-IP₃. Discreet Ca²⁺ puffs can be seen arising at different times and at different sites throughout the cell. Calibration bar depicts Cal-520 fluorescence in arbitrary units from low (black) to high (white); scale bar = 20 µm. (B,C) Traces depict examples of fluorescence signals during blips (B) and puffs (C) measured from 1.5 µm2 regions of interest placed over the centroid of each event. Scale bar indicates the change in fluorescence (∆F) relative to the basal fluorescence (F0) at that site before photorelease of i-IP₃. (D) Cartoon illustrating the hierarchy of IP₃R mediated local Ca²⁺ signals from blips (left) involving the activation of a single IP₃R/channel, to puffs (right) produced by the concerted opening of differing numbers of IP₃Rs within clusters.

Ca²⁺ puffs were first identified [8] and studied [6,8,10,14,15] in Xenopus oocytes. They have subsequently been recognized as a ubiquitous feature of Ca²⁺ signaling in numerous mammalian cell types [7,9,16–20], where they may serve local signaling functions in their own right as well as constituting the ‘building blocks’ for propagation of global Ca²⁺ waves, and acting as the triggers that initiate Ca²⁺ waves [9,12]. High-resolution Ca²⁺ imaging studies have revealed that puffs recur at fixed cellular sites [18,21,22] and involve the concerted openings of a few to a few tens of IP₃R channels that are clustered together within a few hundred nm [23]. The prevailing view is that the Ca²⁺ released by the stochastic opening of one channel may remain isolated to generate a blip, or may activate closely adjacent channels within a cluster by CICR to generate a puff (Fig. 1 B-D). At higher concentrations of IP₃, diffusion of Ca²⁺ over the longer distance scale (a few μm) between clusters may further propagate Ca²⁺ waves through a fire-diffuse-fire mechanism [24].

The mechanisms terminating Ca²⁺ release in the face of self-reinforcing CICR are less well understood, but are thought to involve inhibitory Ca²⁺ binding sites on the cytosolic surface on the IP₃R which leads to channel closure at high [Ca²⁺]_cyt, providing an intrinsic mechanism to terminate channel activation [4,5]. Cytosolic Ca²⁺ therefore serves as both an activator and inhibitor of channel activity, a feature exemplified by the classical bell-shaped curves demonstrating the dependence of [Ca²⁺] on the open probability of IP₃Rs during electrophysiological channel recordings [4,5]. In addition to Ca²⁺-mediated communication between IP₃Rs, observations that tightly grouped IP₃R/channels may close in a synchronized fashion suggests an additional mechanism of allosteric interaction [25], akin to the coupled-closings of ryanodine receptors [26], that ensures robust termination of Ca²⁺ release.

The spatial-temporal patterning of IP₃-mediated cytosolic Ca²⁺ signals is thus determined not only by the functional properties of IP₃Rs, but also by the subcellular spatial distribution of these receptor/channels and by the distribution of IP₃ itself. Recent progress in addressing these topics has been enhanced by advances in microscopy that enable the function, localization and motility of IP₃Rs to be examined at the single-molecule level in intact cells. We first review studies that utilize Ca²⁺ imaging as a functional reporter of IP3R gating and localization, and then consider experiments using fluorescently-tagged IP₃Rs to determine the locations and motility of IP₃R proteins. Finally, we synthesize information derived from these functional and structural approaches in a working model wherein Ca²⁺ puffs are generated by a small number IP₃Rs that are grouped in tight clusters at fixed subcellular sites predominately located near the plasma membrane, while a majority of IP₃Rs are freely motile and appear to be functionally ‘silent’ under conditions that evoke puffs.

2. Imaging IP₃-mediated Ca²⁺ signals down to the single-channel level

High-resolution Ca²⁺ imaging offers a powerful means to study the activity and localization of IP₃R channels in intact cells with minimal perturbation (increased Ca²⁺ buffering by the indicator probe) of the cytoplasmic environment. As we review here, this approach provides millisecond readout of channel kinetics along with sub-micron localization of the Ca²⁺ source. Moreover, it is unbiased to the subtypes of IP₃R generating the signal, and is applicable to natively expressed IP₃Rs without needing any specific labeling of these receptors. On the other hand, drawbacks include the obvious fact that information is obtained only from those IP₃R channels that are open, and that although the location of a single open channel can be determined with high precision, blurring of the fluorescence signal by diffusion precludes the individual localization of clustered channels that open during a puff.

2.1. The ‘optical patch clamp’ and imaging modalities

Ca²⁺ entering the cytosol through an open IP₃R channel provides a reporter of both the gating kinetics of the channel and its location, and has the advantage of inherent amplification of the fluorescence signal because the physiological single channel Ca²⁺ current (~0.1 pA) [27] corresponds to a flux of ~10,000 Ca²⁺ ions per second, ions that may then bind to a large number of probe molecules. Advances in microscopy and fluorescent indicator dyes have made possible the imaging of such single-channel Ca²⁺ signals (‘optical patch-clamp’) with progressively improved resolution [28–30]. Complementing this optical approach, photorelease of a poorly metabolized IP₃ analog (i-IP₃) from a caged precursor that can readily be loaded into cells by incubation as a membrane-permeant ester [31] provides convenient control of concentration and even spatial localization of IP₃ in the cytosol [16,32].

Current generations of electron multiplied CCD and scientific CMOS cameras exhibit high sensitivity (quantum efficiency) together with superior spatial and temporal resolution, and Ca²⁺ indicator dyes that exhibit large changes in fluorescence upon binding Ca²⁺ are available in a wide range of Ca²⁺ affinities and spectral wavelengths. In conjunction with total internal reflection fluorescence (TIRF) microscopy [33], which creates an evanescent field that extends ~100 nm above the glass surface upon which cells are cultured to selective excite fluorescent cytosolic Ca²⁺ indicators adjacent to the cell membrane, these improvements have enabled the imaging of transient, subcellular changes in [Ca²⁺]_cyt with high fidelity [28,29]. As well as its applicability for studying plasmalemmal Ca²⁺ channels, TIRF microscopy is also optimal for studies of IP₃R channels, because the IP₃R clusters that generate puffs are, fortuitously, located in close apposition to the plasma membrane [16,18]. In contrast to spot-scanning techniques such as confocal microscopy the evanescent field permits large cell areas to be continuously illuminated, enabling the positions and activities of many IP₃R clusters to be simultaneously imaged. Resolving the fluorescence signals produced by Ca²⁺ emanating from open IP₃R/channels can be further enhanced by loading cells with the slow Ca²⁺ buffer EGTA, which sharpens the spatial-temporal profile of puffs by limiting the diffusion of Ca²⁺ and inhibiting global elevations in [Ca²⁺]_cyt [34]. When imaging is then performed at high-speed acquisition rates with appropriate fluorescent reporters [35] the Ca²⁺ flux and spatial architecture of IP₃Rs can be interrogated down to the single-channel level in intact cells with a temporal resolution of ~2 ms and a spatial resolution of tens of nm. This has made it possible to determine the numbers and gating properties of the IP₃Rs underlying puffs, as well as their cellular locations [21,23,36]. In concert with advances in imaging hardware, software developed for analyzing massive data files has facilitated the automated detection, localization and analysis of local Ca²⁺ signals from multiple cells within minutes [37,38].

Light-sheet microscopy provides an alternative method for resolving time-dependent fluorescent signals, by projecting a thin sheet of excitation light through the cell and imaging the emitted fluorescence through an objective lens positioned orthogonal to the path of excitation light. An advantage of this technique is that, in contrast to TIRF microscopy, the light-sheet can be projected through any part of the cell and can be scanned to produce 3D volumetric images. Moreover, light-sheet imaging circumvents the necessity to culture cells on a glass substrate as imposed by TIRF microscopy, thereby permitting the study of cells under conditions that may more accurately reflect their native environment without sacrificing the superior spatial resolutions achieved with camera-based imaging. A recent advancement of this technique by Eric Betzig reduced the thickness of the light-sheet and extended its area by structuring it as a dithered lattice of Bessel beams [39]. Lattice light-sheet (LLS) microscopy has been applied to image Ca²⁺ puffs in SH-SY5Y cells, providing a temporal resolution and signal-to-noise ratio comparable to TIRF imaging, but extending exploration through the cell interior rather than being limited to peripheral regions immediately adjacent to the plasma membrane [40].

2.2. Imaging Ca²⁺ and dissecting Ca²⁺ puffs

TIRF imaging of local Ca²⁺ signals provides sensitivity sufficient to resolve openings of individual IP₃R channels. At some sites the fluorescence signals show a stochastic series of ‘square’ events of roughly constant amplitude (Fig. 1B), closely resembling electrophysiological single-channel patch-clamp records. At other sites generating larger Ca²⁺ puffs, stepwise changes in fluorescence are often apparent (Fig. 1C), which result from successive openings and closings of individual channels [23,36]. The amplitude of the unitary Ca²⁺ signal, expressed as a fluorescence ratio change (ΔF/F0), is remarkably constant (~11%) across different sites in a cell, across different cells and even between cell types [20,23,36,41,42], likely reflecting a consistent localization of IP₃R channels in close proximity to the plasma membrane [16,18,40]. Moreover, measurements of the stepwise changes in fluorescence during puffs show that these remain constant for events involving simultaneous openings of at least 5 channels [36], implying that the Ca²⁺ flux through individual channels is not appreciably reduced by local depletion of Ca²⁺ in the ER lumen even when several clustered channels are open.

2.3. Puffs arise at fixed locations, predominantly adjacent to the plasma membrane

From the earliest imaging of Ca²⁺ puffs it was evident that they arose at seemingly stationary subcellular sites, because recurrent events could be recorded from fixed, small regions of interest [8,10]. This observation was subsequently refined by using image analysis techniques to localize the centroids of puff fluorescence signals with a precision of tens of nm, indicating that the location of repeated puffs at a site remained constant within <100 nm, even over periods of tens of minutes [18,20,43]. Localizations of Ca²⁺ blips arising from openings of individual IP₃R channels remained similarly localized [21]. We had thus concluded that individual, functional IP₃Rs, and the clusters of IP₃Rs that underlie Ca²⁺ puffs, must be anchored to some immotile, likely cytoskeletal element, rather than being freely diffusible in the membrane of the ER [22].

Although visualization of puffs by TIRF microscopy indicates that many sites of Ca²⁺ release must lie close to the plasma membrane [16,18,23], wide-field fluorescence imaging revealed further puffs that were not apparent with TIRF illumination [16], leaving open the question of whether those puffs arose in the interior of the cell, or near the plasma membrane on the ‘far’ side of the cell. A direct answer to this question was provided by lattice light-sheet [40] and spinning-disc confocal imaging techniques [18] that provide an optical section through the cell interior with time resolution sufficient to resolve puffs. In SH-SY5Y [40] and HeLa cells [18] puffs were found to arise almost exclusively in close proximity to the plasma membrane. This may not be a universal cellular architecture, however, as Ca²⁺ puffs in HeLa cells were previously reported to preferentially arise around the nucleus [44], and more recent findings in wild-type and gene-edited HEK-293 cells describe ~20% of puffs occurring in the cell interior [42]. Indeed, in polarized epithelial cell types such as hepatocytes [45,46] and pancreatic and parotid acinar cells [47,48] IP₃Rs are localized to specific subcellular domains within the cell interior and may not be capable of producing Ca²⁺ puffs near the plasma membrane [49].

2.4. Selective activation of puff sites by cell surface receptors?

Most studies of Ca²⁺ puffs have utilized stimulation by photorelease of IP₃ or a poorly-metabolized IP₃ analog, i-IP₃, to circumvent complications from the upstream pathway of GPCR-mediated IP₃ production and to provide better control of the magnitude, timing and spatial distribution of cytosolic [IP₃] elevations [16]. IP₃ has generally been photoreleased by uniformly illuminating the cell, so that all puff sites would experience a uniform [IP₃]. In light of findings of hindered cytoplasmic diffusion of IP₃ [32] that raises the question of whether the puff sites activated by this procedure correspond to those natively activated by cell surface receptors that might be envisaged to deliver IP₃ selectively to specific sites. Two recent studies addressed this issue, by comparing the locations of puff sites activated in the same cell following sequential activation by extracellular application of GPCR-linked agonists and then by UV photolysis of caged i-IP₃ [20,43]. Both studies found that the two stimuli evoked puffs at the same sites, and concluded that there was no evidence that endogenous signaling pathways selectively deliver IP₃ to specific sites not accessible to photoreleased i-IP₃.

2.5. Are puff sites preformed or dynamically assembled?

Based on patch clamp recordings of excised nuclei from DT40 cells stably expressing type 1 or type 3 IP₃Rs that showed IP₃Rs to be diffusible within the lipid membrane and to rapidly (a few seconds) cluster together when [IP₃] rises, it was proposed that puff sites may be dynamically assembled in response to IP₃ under physiological conditions [50]. This claim was refuted with analyses of IP₃R channel recordings by Vais H. et al. [51] and was further challenged by imaging studies on the grounds that Ca²⁺ puffs could be evoked within <200 ms following photorelease of IP₃, a time that model simulations suggested was too brief to allow for assembly of clusters containing several IP₃Rs [22], and because stimulation of IP₃ signaling did not result in any rapid changes in motility or distribution of fluorescently tagged IP₃Rs [52]. Most recently, genetic engineering of an enhanced green fluorescent protein (eGFP) tag to native type 1 IP₃Rs has shown these to be distributed in stable punctae containing several IP₃Rs. Immobile punctae tethered close to ER-PM junctions were associated with functional puff sites, and negligible interchange of IP₃Rs with mobile punctae was observed over many minutes [18].

2.6. Counting active IP₃Rs

Ca²⁺ imaging provides a means to estimate the number of IP₃Rs present in a cell that generate puffs. By utilizing the unitary, single IP₃R channel Ca²⁺ signal as a yardstick, the number of channels simultaneously open at the peak of a puff can be found by directly counting stepwise changes or, more readily, by dividing the peak fluorescence signal by the unitary signal. Taking this measure from the largest puff in a sequence of events then provides a minimum estimate of the number of functional IP₃Rs at the cluster underlying the puff site. Under conditions optimized to activate the maximum number of available puff sites within a cell, a rough approximation of the total number of functional IP₃Rs involved in the generation of Ca²⁺ puffs can then be calculated by taking the sum of the greatest number of IP₃Rs at each site. This is likely an underestimation however, since TIRF imaging is restricted to an ~100 nm optical section adjacent to the plasma membrane of the bottom of the cell under study, entirely neglecting or at least understating puffs originating at the top of the cell or deep within the cell interior. Although many cell types are reported to exhibit puffs almost exclusively near the plasma membrane [16,18,20], recent lattice light sheet imaging of HEK-293 cells revealed ~20% of puffs originated within the cell interior [42]. Taking all of this into account, approximations of the number of functional IP₃Rs participating in the generation of Ca²⁺ puffs range from a few hundred up to ~1000 per cell [20,22,23].

2.7. All three IP₃R isoforms mediate puffs

Vertebrates express three main IP₃R isoforms - types 1 [53], 2 [54], and 3 [55] - that are transcribed from three separate genes and co-translationally oligomerize to form tetrameric channels. These isoforms share similar monomeric molecular masses of ~300 kDa but show only 60–80% homology in their primary structures [56]. This diversity is further amplified by alternative splicing forms of these three isoforms [57–59] and because different isoforms can multimerize to form heterotetrameric channels [60–62]. Because most cell types express two or three IP₃R types [60,61,63,64], as well as a poorly defined complement of splice variants, the IP₃R population in native cells comprises a complex mixture of channel configurations with a staggering number of possible combinations [56].

Because of this diversity, the identities of the specific IP₃R isoforms underlying Ca²⁺ puffs had been difficult to resolve because all imaging studies in intact cells were performed in native cell types expressing their endogenous complement of IP₃Rs. Recently, through the use of CRISPR/Cas9 genetic engineering, the Yule lab created HEK-293 cell lines endogenously expressing single IP₃R isoforms [65]. This has facilitated investigations into the roles of IP₃R types 1, 2, and 3 in the generation of local Ca²⁺ signals, and their respective channel properties within the native cellular environment of intact cells where crucial factors including the ER membrane composition, cytosolic and ER luminal milieus are retained [27]. Using these commercially available cell lines, two studies [41,42] determined that IP₃R types 1, 2 and 3 were all capable of producing Ca²⁺ puffs in response to UV flash photolysis of i-IP₃. Puffs evoked in cells expressing each isoform remained at fixed subcellular sites, implying all three IP₃R isoforms can bind to some, as yet unidentified anchoring element/s that define the location and organization of functional IP₃R clusters. Each isoform produced puffs of similar amplitudes, single-channel Ca²⁺ fluxes and spatial distributions, although with subtle differences in puff kinetics between the different isoforms [41,42]. In contrast to their differing affinities to IP₃ observed in single-channel recordings [66], radio-ligand binding studies [54,67,68], and ER Ca²⁺ efflux assays [69], Ca²⁺ puffs displayed similar dependences on i-IP₃ concentration irrespective of the IP₃R isoform expressed [42]. Collectively, these findings demonstrate a substantial redundancy between the different IP₃R isoforms responsible for generating local Ca²⁺ signals in terms of both their channel properties and spatial distributions, suggesting that environmental and accessory factors, potentially in an isoform-specific manner, contribute to shaping unique cellular responses among different cell types.

2.8. Inferring the architecture of IP₃R clusters from Ca²⁺ imaging data

The density and spacing between IP₃Rs within the clusters underlying puff sites have important consequences for the generation of local Ca²⁺ signals, both in terms of the effectiveness of Ca²⁺ diffusion to evoke CICR from neighboring channels [36,70,71] and in light of possible allosteric interactions between IP₃Rs [25]. The ‘nanoarchitecture’ of the functional IP₃Rs in a cluster cannot be directly visualized from the fluorescence signal during puffs because diffusion of Ca²⁺-bound indicator as well as the microscope point-spread signal introduce blurring greater than the expected dimensions of the cluster [70]. Nevertheless, the fluorescence signal reflects the weighted center-of-mass of those IP₃R channels open at a given time, from which useful information can be derived. Shuai et al. [72] modeled the spread of fluorescence signal expected to be generated by clusters of different dimensions, and obtained the best correspondence with experimental data suggesting that brighter puffs in Xenopus oocytes involved the openings of ~25 IP₃R channels, clustered within a diameter of 300–800 nm. Wiltgen et al. [21] further employed super-resolution of fluorescence signals in SH-SY5Y cells to track rapid stepwise changes in centroid position of fluorescence during individual puffs. On the assumption that these apparent movements result from asynchronous gating of different IP₃Rs, they concluded that the clusters extend over diameters of ~400 nm. Consistent with these findings, all-neighbor distance measurements of Ca²⁺ puff centroid locations in HEK WT cells and in HEK cells exclusively expressing single IP₃R types 1, 2, or 3 exhibited prominent peaks at distances <500 nm, indicative of successive puffs recurring in close proximity, at fixed locations [42].

3. Assaying, localizing and tracking IP₃R proteins

Although Ca²⁺ imaging has yielded valuable information on the subcellular distribution and localization of IP₃Rs, that technique has major limitations in that inactive IP₃R channels are invisible, and the diffusion-blurred Ca²⁺ fluorescence signal precludes mapping of multiple open channels within the clusters that underlie puffs. Accordingly, many studies have utilized a complimentary approach of fluorescently labeling IP₃R proteins, which can then be imaged at levels from macroscopic ensembles using whole-cell epifluorescence microscopy, to visualization of individual molecules by super-resolution microscopy. In most instances this has been done by overexpressing IP₃Rs tagged with a fluorescent protein, an approach that may not faithfully replicate the organization of endogenous IP₃Rs. Moreover, IP₃Rs were visualized independently of Ca²⁺ imaging, so that only inferential conclusions could be drawn regarding the relationship between the structural organization of IP₃Rs and their functional activity to liberate Ca²⁺. Most recently, a study by Thillaipan et al. [18] mitigated these limitations, describing imaging of native type 1 IP₃Rs genetically engineered with an eGFP tag, and for the first time presenting near-simultaneous imaging of IP₃R localization together with functional imaging of Ca²⁺ puffs.

3.1. Ca²⁺ puffs are generated by a small population of immotile IP₃Rs

Radioligand binding assays indicate that typical mammalian cells contain on the order of ~7,500 tetrameric IP₃Rs, corresponding to a cytoplasmic concentration of roughly 100 nM [73,74]. This number is considerably greater than the numbers of IP₃R channels (<1000) estimated to be active in generating puffs (see section 2.6). There is also a marked disparity between immunostaining patterns that show IP₃Rs densely distributed throughout the entirety of the ER in mammalian cells [75–77] and observations that Ca²⁺ release arises at just a few discrete puff and blip sites [11,16,19,22,23,78,79].

Another difference arises in the immotility of IP₃Rs implied by the fixed locations of Ca²⁺ puff sites, as compared to studies of fluorescence recovery after photobleaching (FRAP) of GFP-tagged IP₃R proteins indicating that a substantial proportion of IP3Rs are freely diffusible within the ER membrane [59,75–77,80–84], and may aggregate into clusters following sustained (minutes) activation of IP₃ signaling and/or cytosolic Ca²⁺ elevation [59,75–77,84].

Subsequent experiments have examined the motility and distribution of IP₃Rs at the single-molecule level, employing super-resolution imaging techniques to localize and track individual proteins tagged with fluorescent proteins [18,52]. Type 1 IP₃Rs tagged with mEOS and overexpressed in COS-7 cells distributed in a reticular pattern matching the ER network. Consistent with the earlier FRAP studies, a majority of the IP₃Rs (69%) were freely motile within the reticulum (diffusion coefficient D ~0.1 µm² s⁻¹), whereas the remaining IP₃Rs were relatively immotile (D < 0.008) [52]. An ensuing study employed gene editing to tag all type 1 IP₃Rs in HeLa cells with eGFP, revealing that the IP₃Rs are distributed in punctae, each containing a few to several tetrameric IP₃Rs [18]. Super-resolved tracking of fluorescent punctae indicated that most (~70%) punctae were motile, primarily exhibiting Brownian diffusive motion, whereas the remainder appeared immotile.

To account for these apparent discrepancies in numbers, distribution and motility of IP₃Rs as monitored by functional Ca²⁺ imaging versus protein localization, we proposed that Ca²⁺ puffs arise from small numbers of IP₃Rs in immobilized clusters, whereas the majority of IP₃Rs are motile, but are functionally unresponsive (‘silent’) or perhaps liberate Ca²⁺ only in the presence of high concentrations of IP₃ that lead to sustained elevations of global cytosolic Ca²⁺ [22].

3.2. Silent IP₃Rs and hindered cytosolic diffusion of IP₃.

What consequence might the silent majority of IP₃Rs receptors then have for cell signaling? The concentration of IP₃Rs in a typical cell (~100 nM) [85] is larger than estimates of the dissociation constant for IP₃-binding [54,68], and considerably higher than some measures of the effective concentrations of IP₃ (<100 pM) required to evoke Ca²⁺ release [86]. Binding to silent receptors may thus be expected to buffer cytosolic concentrations of IP₃, and hinder its diffusion [32,85].

Different from this, it has been widely believed that IP₃ diffuses readily in the cytosol, with a diffusion coefficient not much slower than free diffusion in water. However, that conclusion was based on measurements of diffusion of radiolabeled IP₃ in bulk extracts of cytoplasm from Xenopus oocytes [24], in which the concentration of IP₃Rs was likely be much lower than in typical cells because IP₃Rs are predominantly restricted to a thin, circumferential shell underneath the plasma membrane [87]. This issue was recently reexamined by employing the latencies of Ca²⁺ puffs at different sites along the length of SH-SY5Y cells as endogenous reporters of local IP₃ concentration following photorelease of a poorly metabolized IP₃ analog at one end of the cell [32]. Fitting of the experimental data by numerical simulations yielded a diffusion coefficient (≤10 µm2 s⁻¹) about 30-fold slower than that previously reported. Thus, in contrast to the long-accepted conclusion of Allbritton et al. [24] that IP3 rapidly diffuses to act as a global messenger throughout cells of typical size, its hindered diffusion by binding to motile, functionally inactive receptors indicates that IP₃ is better considered a local messenger, with a range of action restricted to a few µm depending on its rate of degradation [32].

3.3. Spatial architecture of IP₃ R clusters

As already noted, the density and spacing of IP₃Rs within puff sites has important consequences for how they are able to interact with each other via CICR and allosteric mechanisms. Although Ca²⁺ imaging studies indicate that IP₃Rs are spread over a few hundred nm, the resolution is insufficient to discern the nanoarchitecture of IP₃R distribution within clusters. On the other hand, super-resolution imaging techniques based on localizing the stochastic blinking of individual fluorescent proteins (PALM; [88]) or antibody-conjugated organic dyes (dSTORM; [89]) provide resolution of a few tens of nm, approaching the diameter of the IP₃R channel itself (~25 nm).

Smith et al., [52] tracked individual mEos- tagged type 1 IP₃Rs in live COS-7 cells, and identified clusters of immotile IP₃Rs situated close to the plasma membrane. The spatial distribution of IP₃Rs within these clusters varied widely, but an overlay of 70 clusters showed IP₃Rs to be distributed, on average, as an elongated oval (~400 nm on the long axis and ~200 nm on the short axis), findings that might be expected from their location along strands of the ER. Furthermore, the packing density of mEos-IP₃R molecules within clusters was highly heterogeneous. Some tight clusters exhibited IP₃R localizations within <250 nm of each other, whilst other clusters showed a more sparse distribution.

A recent study [18], utilizing genetically engineering to tag native type 1 IP₃Rs in HeLa cells with eGFP, supports the essence of this model, while prompting important revisions. Firstly, live cell TIRF imaging of Ca²⁺ puffs in conjunction with near-simultaneous imaging of eGFP directly revealed examples of immotile IP₃R clusters underlying puff sites; although a more detailed statistical analysis is needed to substantiate this correlation and establish whether all immotile clusters are functionally active. Second, in contrast to the previous assumption that tracks of motile IP₃Rs represent individual IP₃R channels [52], a surprising finding was that motile IP₃Rs were located within punctae that moved en-mass, rather than as individual proteins [18]. By rapidly fixing cells after live TIRF imaging of eGFP punctae, STORM imaging then enabled super-resolution localization of IP₃Rs within previously identified motile and immotile punctae. Both motile and immotile punctae were found to be composed of IP₃Rs distributed across a typical area of ~0.02 μm², with some receptors in a cluster being closely associated whereas others were separated by distances appreciably greater than the ~25 nm size of the IP₃R itself. The authors thus proposed that IP₃Rs are corralled within mobile and immotile puncta by scaffolds that maintain a loose affiliation between IP₃Rs, without requiring their direct contact [18].

Estimates of the numbers of IP₃R channels (tetramers) within immotile clusters from single-step photobleaching and STORM imaging [18] and from single-particle tracking of over-expressed mEos-IP₃R [52] are subject to considerable uncertainties, but these approaches all point to clusters containing from a few to at most a few tens of IP₃Rs. This is in good agreement with estimates from analysis of Ca²⁺ signals, which suggest that around 2–10 IP₃R channels may open during puffs [20,36]. Thus, it seems that most, if not all IP₃Rs in the stationary clusters that underlie puff sites are functional, although the wide variability in amplitudes of successive puffs at a site indicates that only a subset of IP₃Rs may stochastically respond during any given event.

4. Summary: how the spatial organization of IP₃Rs shapes subcellular Ca²⁺ signals

Advances in imaging technology have greatly enhanced our understanding of the mechanisms underlying the generation and patterning of IP₃-mediated cellular Ca²⁺ signals. For technical reasons our knowledge derives largely from studies on immortalized cell lines (e.g. SH-SY5Y [16,19,21–23,25,32,36,40,43,90], HeLa [7,9,11,18–20,78,79], and HEK-293 cells [20,41,91]) that are readily accessible to high-resolution techniques such as TIRF microscopy. However, key attributes of Ca²⁺ puffs have been corroborated in cultures of diverse primary cell types (e.g. astrocytes [22], fibroblasts [92], endothelial cells [19], smooth muscle [93] and hippocampal neurons [94]), and cells within intact tissues (e.g. vascular endothelial cells [95] and hippocampal astrocytes [96]). Thus, the hierarchical organization of IP₃-mediated Ca²⁺ signals as local and global events appears ubiquitous across numerous cell types, although this architecture may be fine-tuned to suit the particular functions of specific cell types.

Fig. 2 presents a working model of how the organization and functioning of IP₃Rs may underlie the generation of local Ca²⁺ puffs at fixed subcellular sites. This is based on an earlier proposal [28], but incorporates new findings that we review here. IP₃Rs are depicted grouped within clusters; as determined for type 1 IP₃Rs [18] but yet to be confirmed for the other two IP₃R subtypes. Consistent with ensemble bleaching studies [59,75–77,80–84] and single-molecule tracking of IP₃Rs [52] a majority (~70%) of these clusters are mobile within the ER membrane, with the remainder remaining immotile over periods of at least several minutes. Little or no exchange of IP₃Rs is apparent between motile and immotile clusters, so the stationary clusters can be considered as stable entities in terms of both their localization and composition of IP₃Rs. The stationary nature of Ca²⁺ puff sites, together with the multi-channel composition of puffs, have long been taken as indicating that puffs arise from fixed clusters containing small numbers of IP₃Rs [14,23], and a recent paper [18] has correlated a mapping of puff centroids onto immotile IP₃R punctae located near the cell membrane.

Figure 2:

Cartoon illustrating a working model of the organization and function of IP₃R clusters. Two populations IP₃R clusters are depicted: 1) A motile population of IP₃Rs (orange) that move within the endoplasmic reticulum (ER) membrane in preformed clusters with individual channels ‘chained’ together by an as yet unidentified scaffolding structure; and, 2) A stationary population of IP₃Rs clusters which are further held in place near the plasma membrane by ‘riveting’ to some cytoskeletal element. These immotile IP₃Rs are licensed to preferentially respond to low [IP₃], leading to the generation of Ca²⁺ puffs at sites predominately localized close to the plasma membrane (PM).

Key remaining questions concern the nature of the scaffolding that assembles IP₃R in clusters and anchors them at functional puff sites. This architecture is crucial for the patterning of IP₃-mediated Ca²⁺ signaling, both in determining how IP₃Rs may interact among themselves and in placing localized Ca²⁺ signals in relation to other cellular constituents. The recent finding [18] that motile IP₃Rs do not diffuse individually throughout the ER membrane, but rather move en-mass already grouped in punctae, points to two different levels of organization; a pre-formed scaffold that defines the assembly and structure of clusters, and a second structural element, presumably associated to the cytoskeleton, that locks the clusters in place at active puff sites. These two structural elements appear to play different roles in regulating Ca²⁺ liberation through IP₃Rs. The close proximity of IP₃Rs within clusters enhances their interaction by CICR. Moreover, super-resolution imaging revealed an inhomogeneous distribution of IP₃Rs within individual clusters [18] with nm distances separating some IP₃Rs within ‘sub-clusters’ potentially allowing for direct protein-protein interactions that couple their gating behavior [25]. However, although IP₃R clustering leads to a greater overall channel open probability than if the same number of IP₃Rs were widely distributed [36,97] that cannot explain why puffs are almost exclusively localized to stationary punctae near the plasma membrane, because the larger number of ‘silent’, motile clusters have similar architecture. Thus, some unknown feature of the site to which immotile clusters bind, or the environment in which they are located, appears to ‘license’ those IP₃Rs so that they are preferentially able to respond at low [IP₃] [18,22]. That then leaves the question of what, if any, functional role the motile clusters of IP₃Rs may play in Ca²⁺ liberation. Do they respond when strong IP₃ signaling evokes prolonged global Ca²⁺ waves and oscillations, or can the stationary sites that underlie puffs provide sufficient Ca²⁺ to elevate bulk [Ca²⁺]_cyt throughout the cell?

Abbreviations

[Ca²⁺]_cyt

cytosolic free calcium concentration

CICR

Ca²⁺-induced Ca²⁺ release

eGFP

enhanced green fluorescent protein

GPCR

G-protein-coupled receptor

IP₃

inositol 1,4,5-trisphosphate

IP₃R

inositol trisphosphate receptor

LLS

Lattice light-sheet

TIRF

total internal reflection fluorescence