Buffer kinetics shape the spatiotemporal patterns of IP₃-evoked Ca²⁺ signals

Abstract

Ca²⁺ liberation through inositol 1,4,5-trisphosphate receptors (IP₃Rs) plays a universal role in cell regulation, and specificity of cell signalling is achieved through the spatiotemporal patterning of Ca²⁺ signals. IP₃Rs display Ca²⁺-induced Ca²⁺ release (CICR), but are grouped in clusters so that regenerative Ca²⁺ signals may remain localized to individual clusters, or propagate globally between clusters by successive cycles of Ca²⁺ diffusion and CICR. We used confocal microscopy and photoreleased IP₃ in Xenopus oocytes to study how these properties are modulated by mobile cytosolic Ca²⁺ buffers. EGTA (a buffer with slow ‘on-rate’) speeded Ca²⁺ signals and ‘balkanized’ Ca²⁺ waves by dissociating them into local signals. In contrast, BAPTA (a fast buffer with similar affinity) slowed Ca²⁺ responses and promoted ‘globalization’ of spatially uniform Ca²⁺ signals. These actions are likely to arise through differential effects on Ca²⁺ feedback within and between IP₃R clusters, because Ca²⁺ signals evoked by influx through voltage-gated channels were little affected. We propose that cell-specific expression of Ca²⁺-binding proteins with distinct kinetics may shape the time course and spatial distribution of IP₃-evoked Ca²⁺ signals for specific physiological roles.

Cytosolic Ca²⁺ ions regulate numerous aspects of cellular activity in virtually all cell types (Berridge et al. 2000). This versatility is made possible by the diverse mechanisms by which Ca²⁺ signals are generated and transmitted to act over very different time and distance scales (Berridge et al. 2000; Marchant & Parker, 2000). Endogenous Ca²⁺ binding proteins play a key role in determining the magnitude, kinetics and spatial distribution of cytosolic Ca²⁺ signals, and typically buffer most (> 90 %) of the total Ca²⁺ ions in the cytosol (Berridge et al. 2000). As well as simply reducing the availability of free cytosolic Ca²⁺ ions, buffers also affect their diffusional mobility. Immobile buffers reduce the effective diffusion coefficient for Ca²⁺, whereas mobile buffers can act as a ‘shuttle’ to speed Ca²⁺ diffusion in the presence of immobile buffers (Stern, 1992; Roberts, 1994). It is thus likely that cells utilize buffers to shape Ca²⁺ signals for their specific functions; a notion supported by observations that different cell types - particularly sub-populations of neurons - selectively express mobile Ca²⁺ binding proteins (e.g. parvalbumin and calretinin) with differing properties (Andressen et al. 1993). However, the functional consequences of selective buffer expression remain unclear (Neher, 2000; Schwaller et al. 2002).

Most experimental and theoretical investigations of cellular Ca²⁺ buffers have concentrated on signals arising from a fixed ‘pulse’ of Ca²⁺, e.g. Ca²⁺ entry through voltage- or ligand-gated membrane channels (e.g. Lee et al. 2000a,b). A more complex case involves Ca²⁺ liberation from intracellular stores through IP₃Rs. The opening of these channels shows dual positive and negative regulation by cytosolic Ca²⁺ (Iino, 1990; Finch et al. 1991; Bezprozvanny et al. 1991; Mak et al. 1998), leading to a mechanism of CICR whereby small elevations of cytosolic [Ca²⁺] promote channel opening to liberate further Ca²⁺. Buffers are thus expected to exert a powerful effect on signals arising through intracellular Ca²⁺ liberation, as they will affect not only the fate of Ca²⁺ ions already deposited into the cytosol, but also influence the subsequent liberation of Ca²⁺ by modulating feedback onto IP₃Rs.

IP₃-evoked Ca²⁺ signals display complex spatial and temporal patterning. Stimulation by relatively high [IP₃] typically gives rise to repetitive, propagating Ca²⁺ waves (Lechleiter & Clapham, 1992). Lower [IP₃] evokes transient, localized Ca²⁺ elevations, christened ‘Ca²⁺ puffs’, which have been visualized in oocytes and other cell types (Yao et al. 1995; Bootman et al. 1997). Puffs are believed to arise because localized CICR causes the concerted opening of channels within localized clusters of a few tens of IP₃Rs, which act as functionally discrete Ca²⁺ release units (Callamaras et al. 1998a,b; Swillens et al. 1999). Puffs may serve both as local Ca²⁺ signals in their own right, and as ‘building blocks’ from which Ca²⁺ waves are generated. When [IP₃] is sufficiently high, Ca²⁺ diffusing from one release site triggers CICR at neighbouring sites a few microns away, giving rise to saltatory waves propagating from site to site (Yao et al. 1995; Bootman et al. 1997; Berridge, 1997; Callamaras et al. 1998a,b). Thus, Ca²⁺ diffusion and feedback onto the IP₃R act over two very different distance scales: rapid (microsecond) interactions over nanometre scales between individual IP₃Rs within a cluster to generate local Ca²⁺ puffs, and slower (tens of milliseconds) interactions over micrometre scales between release sites to generate global Ca²⁺ waves.

In this paper, we studied how Ca²⁺ buffers with differing kinetics modulate Ca²⁺ interactions over these widely divergent distance and time scales. Our objectives were to: (1) explore how cells might employ buffers to shape Ca²⁺ signals for their specific purposes, and (2) use buffers as a tool to derive mechanistic information about Ca²⁺ interactions at scales otherwise inaccessible to direct visualization.

We utilized Xenopus oocytes as a model cell system to image the perturbations of Ca²⁺ signalling resulting from intracellular injections of two mobile buffers, EGTA and BAPTA. The oocyte has advantages that intracellular Ca²⁺ liberation is mediated solely through type 1 IP₃Rs (Parys et al. 1992); its large size greatly facilitates intracellular injections; and it is among the best characterized cells for Ca²⁺ signalling. EGTA and BAPTA were selected for study because these well-characterized exogenous buffers display similar affinities yet have very different Ca²⁺ binding kinetics (Tsien, 1980). EGTA has an on-rate (≈3-10 μm⁻¹ s⁻¹: Table 1) sufficiently fast to bind Ca²⁺ ions diffusing over the micrometre distances between Ca²⁺ release sites (Roberts et al. 1994; Horne & Meyer, 1997; Song et al. 1998; Kidd et al. 1999), but too slow to influence short range feedback of Ca²⁺ between individual IP₃Rs at individual puff sites. On the other hand, BAPTA has a much faster on-rate (100–1000 μm⁻¹ s⁻¹: Table 1) and, whereas it is expected to disrupt local Ca²⁺ feedback within IP₃R clusters, it may actually enhance coupling between release sites by acting as a ‘shuttle buffer’ (Stern, 1992; Roberts, 1994). We show here that injection of the slow buffer EGTA into the oocyte promotes a ‘balkanization’ of Ca²⁺ liberation as transient, autonomous local release events, whereas the fast buffer BAPTA results in a ‘globalization’ of spatially diffuse Ca²⁺ signals. Moreover, these spatial changes are accompanied by profound changes in kinetics of Ca²⁺ signals, suggesting that Ca²⁺ interactions between release sites play a large part in setting the duration of Ca²⁺ spikes.

Table 1.

Summary of kinetic parameters and diffusion distances for binding and unbinding of Ca²⁺ to EGTA and BAPTA

	EGTA	BAPTA
Apparent Kd (nM) (at pH 7.2)	150	160
Kon (μM−1 s−1)	3–10	100–1000
Koff (s−1)	0.5–1.5	16–160
τdwell (ms)	700–2000	6–60
dshuttle (μm)	28–50	3–9
τcapture (mS) (for B = 270μm)	4–10	0.04–0.4
dcapture (μm) (for B = 270μm)	0.7–1	0.07–0.2

Apparent affinities (Morris et al. 1999) and lower and upper estimates for on-rates (K_on) (Naraghi, 1997; Nagerl et al. 2000), were used to derive off-rates (K_off = K_dK_on) for EGTA and BAPTA. Dwell time (τ_dwell) reflects how long Ca²⁺ will remain bound to each buffer, where τ_dwell = 1/K_off. The corresponding mean distances over which the Ca²⁺-buffer complex will diffuse before releasing bound Ca²⁺ (d_shuttle) were estimated as √ (6D_Cabufferτ_dwell), assuming a diffusion coefficient (D_Cabuffer) of 200 μm² s⁻¹. Mean capture times before a Ca²⁺ ion binds to EGTA or BAPTA were calculated as τ_capture = 1/ (K_onB) (Stern, 1992; Roberts, 1994), where B is the concentration of Ca²⁺-free binding sites on the buffer, assuming that Ca²⁺ ions in the cytosol are bound to immobile endogenous buffers for 90% of the time. The corresponding mean distances that a Ca²⁺ ion would diffuse before capture (d_capture) were calculated as d_capture = √ (D_Caτ_capture), assuming an apparent diffusion coefficient (D_Ca) of 20 μm² s⁻¹ for Ca²⁺ in the presence of immobile endogenous buffers.

METHODS

Preparation of Xenopus oocytes

Xenopus laevis were anaesthetized by immersion in 0.17 % MS-222 for 15 min and killed by decapitation in adherence with protocols approved by the UC Irvine Institutional Animal Care and Use Committee. Oocytes (stage V-VI) were manually plucked and collagenase-treated (0.5 mg ml⁻¹ for 30 min) before storage in Barth's solution (composition (mm): NaCl, 88; KCl, 1; NaHCO₃, 2.4; MgSO₄, 0.82; Ca(NO₃)₂, 0.33; CaCl₂, 0.41; Hepes, 5; pH 7.4) containing 0.1 mg ml⁻¹ gentamicin at 17 °C for 1–7 days before use.

Microinjection of oocytes

Intracellular microinjections were performed using a Drummond microinjector. Oocytes were first loaded with Oregon Green 488 BAPTA 1 (OG-1) together with caged IP₃ (d-myo-inositol 1,4,5-trisphosphate, P₄₍₅₎-(1-(2-nitrophenyl)ethyl) ester) to respective final intracellular concentrations of 48 and 8 μm; assuming 1 μl cytosolic volume. After allowing 30 min for intracellular distribution, control Ca²⁺ responses were imaged. A specified volume of buffer solution was then injected through a fresh micropipette and the pipette was immediately removed to minimize leakage. Ca²⁺ responses were then imaged after allowing ≈20 min for intracellular equilibration of buffer concentration, and sequential injections and recordings were made in this way to examine the effects of stepwise increases in buffer concentration. Injection solutions contained 5 mm of either EGTA or BAPTA together with 2.5 mm CaCl₂ (pH 7.25, with KOH), such that the buffers were half-saturated with Ca²⁺. To obtain high final intracellular concentrations (> 270 μm), stock buffer concentrations were increased to 10 mm (including 5 mm CaCl₂) and buffer was administered by a series of repeated 10 nl injections. Control experiments, where high cytosolic buffer concentrations were attained following single injections of EGTA or BAPTA, gave similar results to those where the concentration was raised in a stepwise manner. Moreover, responses evoked in control oocytes (not injected with buffer) remained consistent over periods up to 3 h.

Confocal laser scanning microscopy

Confocal Ca²⁺ images were obtained using a custom-built line-scan confocal scanner interfaced to an Olympus IX70 inverted microscope (Parker et al. 1997). Recordings were made at room temperature, imaging in the animal hemisphere of oocytes bathed in normal Ringer solution (composition (mm): NaCl₂, 120; KCl, 2; CaCl₂, 1.8; Hepes, 5; pH 7.3). The laser spot of a 488 nm argon ion laser was focused with a ×40 oil immersion objective (NA 1.35) and scanned every 16 ms along a 50 μm line, of which representative trimmed sections are shown in the figures. Emitted fluorescence was detected at wavelengths > 510 nm through a confocal pinhole providing lateral and axial resolutions of about 0.3 and 0.5 μm, respectively. The scan line was focused at the level of the pigment granules and images were collected through a coverglass forming the base of the recording chamber. Fluorescence signals are expressed as ratios (F/F₀ or ΔF/F₀) of the fluorescence (F) at each pixel relative to the mean resting fluorescence (F₀) at that pixel prior to stimulation. IP₃ was photo-released from a caged precursor by delivering flashes of UV light, focused uniformly throughout a 200 μm spot surrounding the image scan line (Callamaras & Parker, 1998). The amount of photo-released IP₃ was controlled (in a linear manner) by varying flash duration. Each flash consumes only a negligible fraction of the caged IP₃ (Callamaras & Parker, 1998), thus it was possible to acquire numerous consistent responses using repeated flashes. Intervals of > 60 s were allowed between recordings to allow IP₃Rs to recover from desensitization and for cytosolic [Ca²⁺] to recover to basal levels. Images were constructed and processed using custom routines written in the IDL programming environment (Research Systems, Boulder, CO, USA), and measurements were exported to Microcal Origin 6.0 (OriginLab, Northampton, MA, USA) for analysis and graphing.

Expression of N-type Ca²⁺ channels

Plasmids containing cDNA clones coding for the N-type Ca²⁺ channel α_1B-d and β₃ subunits (Lin et al. 1997) were linearized and transcribed in vitro. Equal quantities of each subunit cRNA were mixed to a final concentration of 0.1-1 μg μl⁻¹ and injected (50 nl) into defolliculated oocytes 3–5 days prior to imaging and electrophysiological recording. Membrane potential was held at −60 mV using a two-electrode voltage clamp (Gene Clamp 500, Axon Instruments, Foster City, CA, USA), and stepped to +30 mV to induce opening of N-type Ca²⁺ channels expressed in the plasma membrane. Ca²⁺ imaging was as described above, except that [Ca²⁺] in the Ringer solution was 6 mm.

Reagents

OG-1, caged IP₃ and BAPTA were from Molecular Probes Inc. (Eugene, OR, USA); all other reagents were from Sigma Chemical Co. (St Louis, MO, USA).

RESULTS

Actions of buffers on IP₃-evoked Ca²⁺ signals

We compared the actions of fast (BAPTA) and slow (EGTA) Ca²⁺ buffers on IP₃-evoked intracellular Ca²⁺ signals. Xenopus oocytes were loaded with the indicator OG-1 together with caged IP₃, and confocal line-scan images of Ca²⁺-dependent fluorescence were acquired along a 50 μm scan line focused ≈5 μm inward from the cell surface where Ca²⁺ release sites are concentrated (Callamaras & Parker, 1999). IP₃ was photo-released by UV light flashes, with durations set by an electronic shutter to provide a linear control of the relative amount of IP₃ liberation. To control for variation between oocytes, flash strengths were normalized relative to that evoking a half-maximal response under control conditions (i.e. before adding buffer) in each cell.

Figure 1 illustrates Ca²⁺ images obtained at various buffer concentrations in response to a fixed flash strength, and Fig. 2 shows superimposed fluorescence traces in response to various flash strengths at each buffer concentration. Following acquisition of control records (upper panels in Fig. 1 and Fig. 2), a volume of buffer solution (EGTA, Fig. 1A and Fig. 2A; or BAPTA, Fig. 1B and Fig. 2B) was injected into each oocyte to give a final intracellular concentration of 23 μm (see Methods for details). Microinjection pipettes were removed immediately following injection to prevent leakage, and oocytes were left for ≈20 min to ensure uniform buffer distribution and re-equilibration of Ca²⁺ ions in the cytosol before imaging responses to the same series of UV flashes. This procedure was then repeated, sequentially injecting increasing volumes of buffer solutions to image Ca²⁺ signals at progressively higher buffer concentrations (46–540 μm).

Figure 1. Modulation of IP₃-evoked Ca²⁺ signals by EGTA and BAPTA.

Confocal line-scan images illustrate Ca²⁺ signals evoked by photoreleased IP₃ in the presence of increasing concentrations of buffer. In each image, distance is depicted vertically, time runs from left to right and increasing Oregon Green 488 BAPTA 1 (OG-1) fluorescence ratio (ΔF/F₀: proportional to [Ca²⁺]_free) is represented by ‘warmer’ colours as denoted by the colour bar. Identical photolysis flashes (normalized intensities of 1.5 in A and 1.9 in B) were delivered at the arrows. Traces below each image show fluorescence profiles averaged over 100 pixel (6.6 μm) regions. A, upper panel shows response before loading buffer, and subsequent panels illustrate responses after sequentially loading the same oocyte with EGTA to the final intracellular concentrations stated. B, similar records from a different oocyte showing the effects of increasing concentrations of BAPTA.

Figure 2. Buffer actions at varying [IP₃].

Representative fluorescence profiles (from different oocytes to those in Fig. 1) show superimposed Ca²⁺ transients evoked by increasing photorelease of IP₃ in the absence of exogenous buffer (upper panels) and after loading increasing concentrations of EGTA (A) or BAPTA (B). Traces correspond to different photolysis flash strengths, indicated in normalized units.

The data in Figs. 1 and 2 are representative of observations in 45 oocytes (EGTA, n = 21; BAPTA, n = 24). The results are analysed in the following sections, but the most striking finding was that EGTA (a Ca²⁺ buffer with slow Ca²⁺ binding kinetics) greatly abbreviated IP₃-evoked Ca²⁺ signals, whereas BAPTA (a faster Ca²⁺ buffer) slowed their decay.

Changes in Ca²⁺ signals result from increased buffering, not changes in resting [Ca²⁺] or store content

The effects described in Fig. 1 and Fig. 2 might, in principle, arise either through buffering of Ca²⁺ liberated in response to IP₃, or because the added buffers chelated oocyte Ca²⁺ to cause long-term changes in basal cytosolic free [Ca²⁺] or in the loading of Ca²⁺ stores. Observations that equal amounts of two buffers of similar affinity produced dissimilar effects already suggested that the latter possibilities were not the case. Moreover, experiments were carried out by injecting each buffer already half-saturated with Ca²⁺, so that the free [Ca²⁺] in the injection solutions would have been ≈150 nm, a little higher than basal free [Ca²⁺] in the oocyte cytosol (Yao et al. 1995). As an additional control, we repeated the experiments using solutions of EGTA and BAPTA that contained either no added Ca²⁺ or were saturated (1:1 ratio) with Ca²⁺. The results were essentially identical to those in Fig. 1 and Fig. 2 (data not shown), indicating that the effects of EGTA and BAPTA on Ca²⁺ signals did not result from addition or removal of available Ca²⁺ ions within the oocyte.

Positive and negative modulation of IP₃-evoked Ca²⁺ signals by buffers

Figure 3A and B shows families of curves plotting the mean peak amplitude of fluorescence signals evoked by increasing photorelease of IP₃ at various concentrations of EGTA and BAPTA, respectively. High cytosolic concentrations (> 135 μm) of both buffers appreciably reduced the amplitude of Ca²⁺ signals evoked by strong photorelease of IP₃, with BAPTA being more potent than EGTA. This reduction was expected - in part because the buffers will compete with indicator dye for free Ca²⁺ ions. In contrast, Ca²⁺ signals evoked by weak photolysis flashes were potentiated by EGTA; as shown in Fig. 3C, plotting Ca²⁺ signals evoked by weak (0.6-0.8) and strong (1.9-2.6) flashes at various buffer concentrations relative to control responses without added buffer. Potentiation was strongest at low (< 300 μm) concentrations of EGTA, but responses remained greater than control values at concentrations as high as 2 mm EGTA. The slow binding kinetics of EGTA appear to be important for this effect, because little potentiation was apparent with corresponding concentrations of BAPTA (Fig. 3D).

Figure 3. Buffers both facilitate and depress IP₃-evoked Ca²⁺ signals.

A and B, mean peak amplitude (ΔF/F₀) of Ca²⁺ signals as a function of photolysis flash strength, plotted for various intracellular concentrations of EGTA (A: n = 21 oocytes) and BAPTA (B: n = 24). Curves were fitted using the Hill equation. C and D, changes in peak fluorescence signal as a function of increasing [buffer] plotted as a percentage of that in the same oocyte before loading buffer. Data are pooled for weak (black symbols) and strong (red) photolysis flashes (respective normalized flash strengths 0.6–0.8 and 1.9–2.6) in the presence of EGTA (C, n = 14) and BAPTA (D, n = 17).

Buffers reduce the apparent cooperativity of IP₃-evoked Ca²⁺ liberation

The concentration-response relationships in Figs 3A and B are fitted by Hill curves (y = V_max(x^nH/(k^nH+x^nH): where y = peak fluorescence ratio (ΔF/F₀) at any given [IP₃] (x); V_max = peak ΔF/F₀ at saturating [IP₃]; and n_H is the Hill coefficient, which gives a measure of the apparent cooperativity of IP₃ action in evoking Ca²⁺ liberation. Data from these curves are plotted in Fig. 4, showing the dependence of n_H (Fig. 4A) and V_max (Fig. 4B) on concentration of EGTA or BAPTA.

Figure 4. Ca²⁺ buffers reduce the apparent cooperativity of IP₃ action.

A, Hill coefficients, derived from the curves fitted to data in Fig. 3A and B, are plotted as functions of [EGTA] (open circles) and [BAPTA] (filled squares). B, plot shows V_max (maximal fluorescence signal at infinite [IP₃] derived from Hill-fits) as functions of [EGTA] (open circles) and [BAPTA] (filled squares). Horizontal lines on each graph represent control values (i.e. no exogenous buffer).

The steep dependence of Ca²⁺ liberation on [IP₃] is likely to arise both from the requirement for binding to more than one subunit of the tetrameric receptor for channel opening and because of positive cooperativity resulting from CICR (Meyer et al. 1990; Hirota et al. 1995; Marchant & Taylor, 1997; Callamaras et al. 1998a). In agreement with this dual mechanism, n_H in control conditions (before adding buffer) was ≈6.3, a value greater than expected even if channel opening requires binding to all four subunits. Moreover, progressive increases in buffer concentrations reduced n_H to a minimal value of about 2 (Fig. 4A). A probable interpretation is that Ca²⁺ buffers disrupt the diffusion of Ca²⁺ ions between IP₃Rs that normally leads to CICR, and that this communication contributes appreciably to the apparent cooperativity of IP₃ action. Indeed, the minimal n_H at high buffer concentrations suggests that binding of IP₃ to two (or possibly even one) subunits may be sufficient for channel opening. Finally, moderate concentrations of EGTA caused a greater reduction in n_H than equivalent concentrations of BAPTA (Fig. 4A). Because EGTA impairs Ca²⁺ diffusion between clusters of IP₃Rs rather than between individual IP₃Rs within a cluster (see Discussion), it seems that cluster–cluster interactions predominate in establishing the normally high cooperativity of IP₃ action.

V_max values derived from Hill fits provide a measure of maximal fluorescence at saturating [IP₃], and are plotted as a function of [buffer] in Fig. 4B. V_max declined progressively with increasing concentrations of BAPTA, whereas a biphasic relationship was apparent with EGTA, with an initial potentiation apparent at 23 and 46 μm EGTA.

Differential effects of buffers on kinetics of IP₃-evoked Ca²⁺ signals

As noted earlier, EGTA and BAPTA exerted very different actions on the kinetics of IP₃-evoked Ca²⁺ transients; respectively causing an acceleration or prolongation of their decay. This is illustrated in Fig. 5A, showing superimposed responses to strong photolysis flashes in the presence of increasing [EGTA] or [BAPTA] after normalizing amplitudes to facilitate comparison. Fluorescence decays were fit by mono- or bi-exponential curves (Fig. 5B). In the absence of exogenous buffer, the decay followed a single exponential, with a time constant (τ_slow) of a few seconds that slowed progressively with increasing photorelease of IP₃. In the presence of EGTA the decay phase became markedly biphasic (Fig. 5B), with a prominent fast component (τ_fast) and smaller slower component (τ_slow). In the presence of BAPTA a slow decay was prominent, and some oocytes (13/20) also showed a small fast component (Fig. 5B).

Figure 5. Ca²⁺ transients are speeded by EGTA but prolonged by BAPTA.

A, families of curves illustrate Ca²⁺ transients evoked by a fixed photolysis flash (normalized strength 2.5) in the presence of the indicated concentrations of EGTA (left) and BAPTA (right). Responses are scaled to same peak height to facilitate comparison. B, Ca²⁺ transients evoked by various photolysis flashes of various strengths (normalized strengths indicated) before loading buffer (left); in the same oocyte after loading 135 μm EGTA (middle); and in a different oocyte loaded with 135 μm BAPTA (right). Red curves illustrate fitting of single or double exponentials to derive fast (τ_fast) and slow (τ_slow) decay time constants.

Figure 6 shows measurements of time constants of the fast and slow decay components, and their relative weights in contributing toward the overall decline in [Ca²⁺], plotted as a function of flash strength for various concentrations of EGTA or BAPTA. Several features are evident. (1) In the absence of added buffer, Ca²⁺ transients showed a mono-exponential decay with a time constant slowing from ≈1 s with just-suprathreshold photorelease of IP₃ to nearly 10 s with stronger photo-release (black symbols, Fig. 6A–D). (2) EGTA caused the decay to become markedly biphasic, with an initial fast decline (τ_fast) that contributed > 80 % of the total decay (Fig. 6E) and shortened with increasing [EGTA] (Fig. 6A). The value of τ_fast slowed with increasing [IP₃] at low (< 50 μm) concentrations of EGTA, but was relatively independent of [IP₃] at high (> 135 μm) concentrations of EGTA (Fig. 6A). Following the large, fast decay, the remaining small component declined with a time constant (τ_slow) that lengthened slightly with increasing photorelease of IP₃, and was slower than in the absence of added buffer (Fig. 6C). (3) Addition of BAPTA greatly slowed (τ_slow 10–30 s) the major part of the Ca²⁺ signal (Fig. 6D). In those oocytes that displayed biphasic kinetics, the initial fast component was much smaller than with EGTA (Fig. 6E), and its time constant (τ_fast) was faster than in controls, but slower than with EGTA (Fig. 6B).

Figure 6. EGTA and BAPTA differentially affect the fast and slow decay components of IP₃-evoked Ca²⁺ transients.

A and C, graphs show, respectively, the time constants for the fast (τ_fast) and slow (τ_slow) decay components as a function of photolysis flash strength, in the presence of various concentrations of EGTA. B and D, similar data, with corresponding concentrations of BAPTA. Because the decay of Ca²⁺ transients without added buffer was mono-exponential, control time constants are repeated in the upper and lower panels (black symbols joined by lines). E, plot shows the relative magnitude of the fast decay component (τ_fast: measured as shown in Fig. 5B) at various concentrations of EGTA (red symbols) and BAPTA (black symbols). The amplitudes of the fast and slow components (a_fast and a_slow, respectively) were derived at the time of the peak signal from bi-exponential fits, and the graph shows the percentage contribution of a_fast to the total signal (i.e. a_fast/(a_fast+a_slow) × 100). Open and filled symbols represent signals evoked, respectively, by weak and strong photolysis flashes.

Decay of Ca²⁺ transients is dominated by release kinetics, not sequestration

IP₃-evoked Ca²⁺ signals reflect a balance between Ca²⁺ liberation through IP₃Rs into the cytosol and its subsequent removal by processes including buffering, extrusion, re-sequestration via sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) pumps and diffusion into the enormous interior volume of the oocyte (Parker et al. 1996). The differing effects of EGTA and BAPTA on the kinetics of Ca²⁺ signals (Fig. 6) might thus arise either through modulation of Ca²⁺ liberation or through actions on subsequent cytosolic Ca²⁺ clearance. To discriminate between these possibilities, we examined the effects of buffers on signals arising from Ca²⁺ flux into the cytosol through voltage-gated N-type Ca²⁺ channels expressed in the plasma membrane. Opening of these channels by a depolarizing pulse is expected to induce a fixed ‘pulse’ of Ca²⁺ influx - without complications of CICR as with IP₃Rs - and the decay of the resulting Ca²⁺ transient should thus provide a good measure of cytosolic Ca²⁺ clearance. Moreover, the spatial distribution of Ca²⁺ influx across the plasma membrane is expected to closely mimic Ca²⁺ liberation from IP₃-sensitive sites, as these are distributed as a thin shell close under the membrane (Callamaras & Parker, 1999). Ca²⁺ arising from both sources will thus approximate a 2-dimensional sheet, with subsequent diffusion along a radial axis into the interior of the oocyte.

Oocytes expressing N-type channels were voltage clamped at −60 mV and stepped to +30 mV for 300 ms to induce Ca²⁺ influx with a duration matching the rise time of IP₃-evoked Ca²⁺ signals (Fig. 7). After the fluorescence signal had returned to baseline (10 s), a UV flash was applied to photorelease IP₃ (red arrow). Before loading Ca²⁺ buffers, signals induced by activation of N-type Ca²⁺ channels decayed much more rapidly than those evoked by IP₃, even though they were of similar amplitude (Fig. 7A). This marked disparity indicates that the slow decay of the IP₃-evoked signals did not arise through dye saturation - a conclusion further supported by findings of similar decay rates with the low affinity (K_d≈4 μm) indicator Fluo-4 dextran (not shown).

Figure 7. The decay of IP₃-evoked Ca²⁺ transients is largely determined by the kinetics of Ca²⁺ release, not sequestration.

A, line-scan image shows a Ca²⁺ transient evoked by a 300 ms duration depolarization to +30 mV to activate expressed N-type channels, followed 10 s later by a UV flash to photorelease IP₃. The lower trace shows the fluorescence profile averaged across a 6.6 μm region of the scan line, and superimposed traces on the right compare the voltage-gated signal (black) with the IP₃-evoked signal (red) after normalizing to the same peak height. B, corresponding image and traces recorded under identical conditions in the same oocyte after loading 270 μm EGTA. C, similar records in a different oocyte loaded with 270 μm BAPTA. Control records in this oocyte before loading BAPTA were similar to one in A.

Thus, the decay of IP₃-evoked signals does not simply reflect Ca²⁺ clearance, but instead is determined primarily by a persistent liberation of Ca²⁺ through IP₃Rs lasting for several seconds. Furthermore, the effects of EGTA and BAPTA are most readily explained by their actions on this prolonged Ca²⁺ liberation, rather than on Ca²⁺ clearance from the cytosol. Both EGTA (Fig. 7B) and BAPTA (Fig. 7C) produced only a modest speeding of decay of signals resulting from Ca²⁺ influx through N-type channels, in marked contrast to their respective actions to accelerate or slow IP₃-evoked Ca²⁺ signals.

Biphasic Ca²⁺ liberation through IP₃Rs

Data from experiments similar to those in Fig. 7 were then used to derive the kinetics of Ca²⁺ flux into the cytosol through IP₃R, assuming that subsequent Ca²⁺ clearance follows a first order process with a rate constant (k) corresponding to the decay of Ca²⁺ signals following influx through N-type channels (Fig. 8). Ca²⁺ flux from intracellular stores is thus proportional to d[Ca²⁺]/dt+k[Ca²⁺], where [Ca²⁺] is the free cytosolic Ca²⁺ concentration as signalled by changes in fluorescence (Parker et al. 1996). Consistent with previous observations (Parker et al. 1996), control oocytes showed an initial fast spike of IP₃-evoked Ca²⁺ liberation, followed by a lingering ‘tail’ of Ca²⁺ efflux persisting for several seconds (upper traces, Fig. 8B and D). Although the tail component is small (Fig. 8A and C), over time it contributes one-half or more of the total Ca²⁺ liberated, and is responsible for the prolonged Ca²⁺ transients evoked by IP₃ under normal conditions. EGTA almost abolished the tail component of Ca²⁺ liberation (Fig. 8B), thus accelerating the decay of the fluorescent Ca²⁺ profile (Fig. 8A). Different to this, the tail component was not significantly altered by BAPTA (Fig. 8D), suggesting that prolonged release of Ca²⁺ through the IP₃R may involve cluster–cluster interactions that are selectively disrupted by slow buffers.

Figure 8. EGTA abolishes the ‘tail’ of IP₃-evoked Ca²⁺ liberation that normally prolongs Ca²⁺ transients.

A, fluorescence signals evoked by Ca²⁺ entry through voltage-gated channels (bar) and by photoreleased IP₃ (arrow), before (upper) and after (lower) loading 270 μm EGTA, derived from images similar to Fig. 7A and B. B, rates of Ca²⁺ flux into the cytosol derived from the records in A, as described in the text. Traces were smoothed using 15 point adjacent averaging. Calibration bar corresponds to a rate of increase in fluorescence (d(ΔF/F₀)/dt) of 3 s⁻¹. Inset graphs show the same records with expanded vertical axes to better illustrate the abolition of the persistent ‘tail’ of Ca²⁺ liberation by EGTA. C and D, similar data from a different oocyte, illustrating the lack of action of BAPTA (270 μm) on the tail of Ca²⁺ liberation.

Changes in spatial distribution of Ca²⁺ signals

In addition to their actions on the amplitude and kinetics of IP₃-evoked Ca²⁺ signals, EGTA and BAPTA produced characteristic, and very different, changes in the spatial distribution of Ca²⁺ signals (Fig. 9). Control oocytes (without added buffer) often display transient localized Ca²⁺ signals (puffs) in response to photolysis flashes within a narrow ‘window’ of flash strengths (Yao et al. 1995), but these were not evident in the oocytes illustrated, and only spatially diffuse Ca²⁺ waves were observed (Fig. 9A). However, following injection of EGTA (135 μm), the same oocytes showed discretely localized puffs (Fig. 9B: representative of 14/15 oocytes). With weak stimuli puffs arose at discrete sites following differing latencies. Stronger photorelease of IP₃ caused sites to respond almost simultaneously, giving rise to the abrupt, rapidly decaying Ca²⁺ transient described earlier. Puffs continued after this initial transient and became more frequent, but smaller, with progressively stronger stimuli. In marked contrast, the same concentration (135 μm) of BAPTA abolished Ca²⁺ responses to weak photolysis flashes (10/18 oocytes with normalized flash strengths <1.1), and at higher flash strengths oocytes showed only spatially uniform and slowly decaying Ca²⁺ signals. We never (0/24 oocytes) observed localized Ca²⁺ puffs.

Figure 9. ²⁺ signals into discrete, autonomous units, whereas BAPTA promotes spatially uniform global signals.

A, line-scan images and fluorescence profiles (averaged over 4 μm regions) showing responses to photolysis flashes (red arrows) of increasing strength (indicated in normalized units) before injecting buffer. B, corresponding records in the same oocyte after loading 135 μm EGTA. Two representative fluorescence profiles are illustrated from each image, recorded at different puff sites (arrowed). C, corresponding records in a different oocyte loaded with 135 μm BAPTA. Records in this oocyte before loading BAPTA were similar to those in A.

DISCUSSION

We show that two mobile Ca²⁺ buffers with similar affinities but very different binding kinetics affect the spatiotemporal characteristics of IP₃-evoked Ca²⁺ liberation in contrasting ways. EGTA - a buffer with a slow binding rate - causes Ca²⁺ signals to become more transient and ‘balkanizes’ Ca²⁺ liberation such that individual release sites act autonomously to generate discrete puffs rather than functioning in concert to generate Ca²⁺ waves. On the other hand, BAPTA - a fast buffer - prolongs Ca²⁺ transients and promotes a ‘globalization’ of Ca²⁺ signalling as a spatially uniform elevation without evidence of local release events. These strikingly different actions cannot be explained merely by chelation of Ca²⁺ subsequent to its liberation into the cytosol, because Ca²⁺ signals generated by influx through voltage-gated channels were affected to a much lesser degree by EGTA and BAPTA (Fig. 7). Moreover, they do not arise through changes in resting free [Ca²⁺] or alterations in Ca²⁺ store filling, as similar results were obtained whether buffers were injected either without Ca²⁺ or saturated with Ca²⁺. Instead, EGTA and BAPTA appear to act over different time and distance scales to modulate the processes of Ca²⁺ diffusion and CICR that shape the regenerative nature of IP₃-evoked Ca²⁺ liberation.

Spatiotemporal scales of Ca²⁺ signalling

IP₃Rs are organized in Xenopus oocytes as tight (ca 50 nm diameter) clusters containing several tens of channels, with a mean cluster–cluster spacing of ≈3 μm (Callamaras et al. 1998a,b; Swillens et al. 1999; Shuai, 2002,Jung, 2003). CICR can thus act over two very different spatiotemporal scales - via fast diffusion over nanometre distances between IP₃Rs within a cluster to give local Ca²⁺ puffs, and much more slowly across micrometre distances between clusters to generate propagating saltatory waves (Fig. 10A: Yao et al. 1995; Bootman et al. 1997; Berridge, 1997; Callamaras et al. 1998a,b; Marchant & Parker, 2000). Depending on their kinetics, buffers may selectively act on one or other of these processes.

Figure 10. Cartoon illustrating the distribution of Ca²⁺ ions around IP₃R clusters, and how communication between IP₃Rs may be affected by mobile buffers with differing binding kinetics.

A, without added buffer, Ca²⁺ ions liberated through open IP₃ at a cluster diffuse a few micrometres and, depending on [IP₃], may or may not trigger CICR at adjacent clusters. B, EGTA acts too slowly to disrupt Ca²⁺ diffusion and CICR within clusters, but binds Ca²⁺ ions as they diffuse between clusters and rapidly ‘shuttles’ them away, so as to disrupt cluster–cluster interactions. C, BAPTA binds Ca²⁺ ions sufficiently quickly to disrupt Ca²⁺ communication between IP₃Rs within individual clusters, whilst promoting Ca²⁺ communication between clusters by acting as a rapid Ca²⁺ shuttle.

Key factors determining the interactions of buffers with Ca²⁺ ions are summarized in Table 1, and include: (1) the mean time (τ_capture) and distance (d_capture) that a Ca²⁺ ion will diffuse before it becomes bound to a buffer molecule; (2) the mean time (dwell time; τ_dwell) a Ca²⁺ ion will remain bound to a buffer before dissociating, and the corresponding mean distance (‘shuttle’ distance; d_shutttle) by which the Ca²⁺-buffer complex will diffuse before dissociation. Further, diffusion of Ca²⁺ ions in the cytosol is slowed by binding to endogenous, immobile buffers. Less than 10 % of the total Ca²⁺ ions in the cytosol are free at any given time, and the apparent diffusion coefficient for Ca²⁺ in the oocyte is thereby slowed about 10-fold as compared to free aqueous diffusion (respectively, 20 vs. 200 μm² s⁻¹: Allbritton et al. 1992; Yao et al. 1995). Buffers, such as EGTA and BAPTA can thus act as a ‘shuttle’ to facilitate diffusion (Stern, 1992; Roberts, 1994) because, when bound to a mobile buffer molecule, a Ca²⁺ ion is ‘protected’ from binding to immobile sites and is thus carried appreciable distances at a rate only a little slower than that for free aqueous diffusion.

Buffer actions on inter- and intra-cluster signalling

We present here a qualitative model to account for the actions of EGTA and BAPTA on IP₃-evoked Ca²⁺ signalling, which is summarized in Fig. 10. This is necessarily simplified, and will require mathematical modelling for rigorous verification, but nevertheless offers useful insights.

Dealing first with EGTA, we consider how cytosolic diffusion of Ca²⁺ would be altered in the presence of 270 μm of this buffer. Its slow on-rate implies that a Ca²⁺ ion will diffuse an average distance of about 1 μm before binding to an EGTA molecule (Table 1). EGTA is thus expected to have little effect on short-range Ca²⁺ feedback over nanometre distances between individual IP₃Rs within a cluster, but it will efficiently capture Ca²⁺ ions diffusing over micrometre distances between neighbouring clusters (Roberts, 1994; Horne & Meyer, 1997; Song et al. 1998; Callamaras et al. 1998a,b; Kidd et al. 1999). Once bound Ca²⁺ ions will be unavailable to trigger CICR for prolonged periods (τ_dwell = 0.7-2 s: Falcke, 2003), and Ca²⁺ bound to EGTA will be shuttled long distances (d_shuttle = 28–50 μm) before being ‘dumped’ in the interior of the oocyte where release sites are absent (Callamaras & Parker, 1999). The overall effect will thus be to more sharply restrict Ca²⁺ signals around individual clusters, and reduce free [Ca²⁺] between clusters so as to functionally uncouple them (Fig. 10B). In agreement, EGTA has little effect on puffs (Callamaras & Parker, 2000), yet strongly inhibits Ca²⁺ waves and dissociates global IP₃-evoked Ca²⁺ signals into discrete, localized Ca²⁺ release events (Fig. 9: Horne & Meyer, 1997; Kidd et al. 1999; Callamaras & Parker, 2000).

In contrast, BAPTA has an on-rate about 100 times faster than EGTA, and binds Ca²⁺ ions within a distance (tens of nanometres; Table 1) comparable to the dimensions of an IP₃R cluster, thereby reducing CICR between individual receptors within the cluster. Subsequently, Ca²⁺ bound to BAPTA will rapidly (ca 10 ms) be ‘shuttled’ a few micrometres before dissociating. This is comparable to the inter-cluster spacing, so whereas BAPTA will inhibit intra-cluster feedback by Ca²⁺ it may actually facilitate inter-cluster communication. Moreover, discrete puffs arise only when the Ca²⁺ release time at a site (ca 50 ms: Sun et al. 1998) is short in relation to the inter-cluster diffusion time (Dawson et al. 1999; Shuai & Jung, 2003). This condition is fulfilled in the absence of exogenous buffer, but the ability of BAPTA to rapidly shuttle Ca²⁺ and thereby speed its diffusion effectively ‘smears’ the distribution of Ca²⁺, so that the endoplasmic reticulum (e.r.) approximates a deterministic, spatially continuous Ca²⁺ source, rather than an array of loosely coupled, discrete stochastic sources (Dawson et al. 1999; Shuai & Jung, 2003).

Cooperativity of IP₃-evoked Ca²⁺ release

In the absence of added buffer, Ca²⁺ liberation in the oocyte increases steeply (n_H > 6) with [IP₃]. This is likely to reflect both the requirement for binding of multiple IP₃ molecules to open the homo-tetrameric IP₃R, and positive cooperativity introduced by CICR. Our results suggest that the latter is the more prominent factor. High (millimolar) concentrations of both EGTA and BAPTA reduced n_H to ≈2, suggesting that binding of two (or possibly one) molecules of IP₃ is sufficient for channel opening. Interestingly, moderate concentrations of EGTA (approximately 25 μm) caused a much greater reduction in n_H than the same concentrations of BAPTA, further implying that inter-, rather than intra-cluster interactions are the main factor contributing to the high cooperativity under physiological conditions.

Potentiation of Ca²⁺ signalling by low [EGTA]

Ca²⁺ signals evoked by strong photorelease of IP₃ became progressively smaller with increasing concentrations of EGTA and BAPTA, as might be expected because the buffers will compete with the indicator dye for available Ca²⁺ as well as reducing Ca²⁺ liberation by impairing CICR. On the other hand, EGTA (but not BAPTA) actually potentiated responses evoked by weak photorelease of IP₃ by as much as 100 % (Fig. 3C). This surprising finding is consistent with a recent demonstration that slow, mobile Ca²⁺ buffers (parvalbumin and EGTA) promote puff activity in Xenopus oocytes, probably by sensitizing IP₃R to resting levels of IP₃ (John et al. 2001). Potentiation of Ca²⁺ excitability by expression of specific endogenous buffers may, therefore, play an important role in regulating cellular signalling. The underlying mechanism is unclear (John et al. 2001), but effects on cluster–cluster interactions are implicated because potentiation is most prominent with slow buffers, appears not to result from changes in resting free [Ca²⁺], and does not correlate with binding of Ca²⁺-free forms of different buffers to IP₃Rs (Richardson & Taylor, 1993; John et al. 2001).

Cluster–cluster interactions prolong Ca²⁺ liberation

Ca²⁺ signals during local Ca²⁺ puffs persist for only a few hundred milliseconds (Yao et al. 1995; Sun et al. 1998), whereas Ca²⁺ elevations during global waves decay over several seconds (e.g. Fig. 2). We originally thought that this difference might reflect differences in Ca²⁺ clearance, and that the kinetics of Ca²⁺ liberation were similar during these local and global events. Specifically, it seemed that Ca²⁺ liberated at a ‘point’ source (as during a puff) would rapidly dissipate by 3-dimensional diffuse into the vast volume of the oocyte, whereas the spatially extensive Ca²⁺ release during a wave would be cleared primarily by slower processes such as sequestration back into the e.r. (Yao et al. 1995). However, the experiment of Fig. 7 indicates that this is not a full explanation, because IP₃-evoked Ca²⁺ signals decayed much more slowly than equivalent signals evoked by extracellular Ca²⁺ influx through voltage-gated channels. Instead, we conclude that IP₃-evoked waves are prolonged largely because Ca²⁺ efflux from the e.r. is biphasic. That is to say, a large, but rapidly terminating flux during the rising phase of the wave is followed by a ‘tail’ of Ca²⁺ efflux which, though small, is sufficiently prolonged that it greatly slows the decline of cytosolic [Ca²⁺] (Fig. 8A). The actions of EGTA and BAPTA on global Ca²⁺ signals to respectively speed and slow their decay can then be understood in terms of their relative effects on the fast and slow components of Ca²⁺ efflux. EGTA almost abolishes the slow component while having little effect on the fast component, whereas BAPTA selectively attenuates the fast component and prolongs the slow component (Fig. 8B and C).

Termination of Ca²⁺ flux through IP₃Rs thus appears to involve two distinct processes. Flux at individual clusters acting in isolation (at low [IP₃], or after uncoupling with EGTA) ceases abruptly and stochastically within ≈100 ms, and shows an ‘adaptive’ property such that it can be reactivated by further elevation of [IP₃] (Callamaras & Parker, 2000). On the other hand, clusters recruited during a wave show an initial large Ca²⁺ efflux with a time course similar to that of puffs, and then abruptly transition to a state in which they continue to ‘leak’ Ca²⁺ at a low rate for several seconds. The mechanisms underlying this transition remain unknown, but the striking action of EGTA to attenuate the slow release suggests that it involves diffusional spread of Ca²⁺ between clusters, such that each continues to maintain others in an activated state. Moreover, the importance of the slow release component in shaping Ca²⁺ transients further emphasizes the role of buffers in cell signalling. Although the slow Ca²⁺ flux is small, its persistence means that the total amount of Ca²⁺ liberated is greater than that during the fast component; and attenuation of the slow component by EGTA shortens the duration of Ca²⁺ spikes as much as 10-fold.

Physiological roles of buffers in cellular Ca²⁺ signalling

Our results point to the importance of buffering kinetics in shaping the spatial and temporal patterning of Ca²⁺ signals generated by intracellular Ca²⁺ liberation through Ca²⁺-sensitive release channels - effects that are more complex than for signals arising from a fixed ‘pulse’ of Ca²⁺, as with Ca²⁺ entry through voltage-gated channels (e.g. Lee et al. 2000). For convenience we utilized the synthetic buffers EGTA and BAPTA, which are not normally present within cells. However, preliminary experiments with endogenous mobile Ca²⁺-binding proteins (Dargan et al. 2003) point out the physiological relevance of our results. In particular, parvalbumin - a mobile Ca²⁺ buffer with slow Ca²⁺ binding kinetics (Lee et al. 2000b; Schmidt et al. 2003) - balkanized Ca²⁺ signals and accelerated their decay similar to EGTA; whereas the fast buffer calretinin (Edmonds et al. 2000) produced BAPTA-like effects. Although there are obvious differences between small-molecule and protein buffers (such as differing mobility and possibility of protein-protein interactions), the emergent theme is that the kinetics of a buffer are the key parameters defining its actions in shaping IP₃-mediated Ca²⁺ signals. Moreover, this theme is likely to extend to mammalian systems because the clustered distribution of IP₃Rs in the Xenopus oocyte is mirrored in diverse mammalian cells (Bootman et al. 1997; Simpson et al. 1997). We finally note that ryanodine receptors (RyRs), the other major class of intracellular Ca²⁺ release channel, also communicate via CICR (Berridge, 1997), so that their signalling may similarly be susceptible to shaping by mobile Ca²⁺ buffers.