Type III InsP₃ receptor channel stays open in the presence of increased calcium

Abstract

The inositol 1,4,5-trisphosphate receptor (InsP₃R) is the main calcium(Ca²⁺) release channel in most tissues. Three isoforms have been identified^1–6, but only types I and II InsP₃R have been characterized^7,8. Here we examine the functional properties of the type III InsP₃R because this receptor is restricted to the trigger zone from which Ca²⁺ waves originate^9–11 and it has distinctive InsP₃-binding properties^12,13. We find that type III InsP₃R forms Ca²⁺ channels with single-channel currents that are similar to those of type I InsP₃R; however, the open probability of type III InsP₃R isoform increases monotonically with increased cytoplasmic Ca²⁺ concentration, whereas the type I isoform has a bell-shaped dependence on cytoplasmic Ca²⁺. The properties of type III InsP₃R provide positive feedback as Ca²⁺ is released; the lack of negative feedback allows complete Ca²⁺ release from intracellular stores. Thus, activation of type III InsP₃R in cells that express only this isoform results in a single transient, but global, increase in the concentration of cytosolic Ca²⁺. The bell-shaped Ca²⁺-dependence curve of type I InsP₃R is ideal for supporting Ca²⁺ oscillations, whereas the properties of type III InsP₃R are better suited to signal initiation.

When homogenates from RIN-5F cells, rat hepatocytes, and canine cerebellum were probed by western blot analysis with isoform-specific antibodies, RIN-5F cells contained no detectable type I InsP₃R (Fig. 1a, lane 4) but did express type III InsP₃R (Fig. 1b, lane 4). In contrast, both hepatocytes and cerebellum expressed type I InsP₃R (Fig. 1a, lanes 1 and 3), but no detectable type III InsP₃R (Fig. 1b, lanes 1 and 3). Thus, the InsP₃R of RIN-5F cells is almost entirely the type III isoform, consistent with previous reports¹⁴. Immunocytochemistry demonstrated that type III InsP₃R was diffusely distributed (Fig. 1c) and confirmed that type I InsP₃R expression was minimal (Fig. 1d, e). These findings do not support a previous prediction that type III InsP₃R is preferentially localized near the plasma membrane¹⁵.

Figure 1.

RIN-5F cells preferentially express Type III InsP₃R. a, b, Western blots were probed for types I and III InsP₃R (a and b, respectively). Dog cerebellum and rat hepatocytes were positive controls for type I InsP₃R; HeLa cells were a positive control for type III InsP₃R. Lanes were loaded with 30 μg dog cerebellar microsomes (lane 1), 10 μg HeLa cell lysate (lane 2), 30 μg hepatic microsomes (lane 3), and either 30 μg (a) or 5 μg (b) of RIN-5F microsomes (lane 4). c, e, Cellular distribution of type III (c) and type I (e) InsP₃R in RIN-5F cells. d, Nonspecific binding.

To test whether type III InsP₃R forms a Ca²⁺ channel, we incorporated endoplasmic reticulum vesicles from cultured RIN-5F cells into planar lipid bilayers. Single-channel currents were observed through type III InsP₃R. Like type I InsP₃R, type III InsP₃R required InsP₃ to open (Fig. 2a: compare top three and bottom three traces). The single-channel current of type III InsP₃R measured at 0 mV with 55 mM Ba²⁺ on the lumenal side of the channel was ~2.5 pA, which is similar to that of type I InsP₃R (ref. ⁷). The conductances of type III InsP₃R (88 ± 4 pS; Fig. 2b) and type I InsP₃R (85 ± 3 pS)⁷ were also similar.

Figure 2.

Type III InsP₃R is an InsP₃-gated Ca²⁺ channel. a, InsP₃-gated Ca²⁺ channels from endoplasmic reticulum of RIN-5F cells in planar lipid bilayers. In the absence of InsP₃, channel activity was not observed (top three traces). Addition of 2 μM InsP₃ to the cytoplasmic side induced channel activity (bottom three traces). Channel openings are shown as downward deflections from baseline. Ruthenium red (2 μM) was present to block ryanodine receptors. b, Current–voltage relationship of type III InsP₃R. Inset shows an amplitude histogram at 0 mV for one experiment. Values plotted in the I–V curve represent the mean for three experiments. Standard errors for data points, which range from 0.02 to 0.05 pA, are too small to be seen.

A central feature of type I InsP₃R is that Ca²⁺ acts as an allosteric regulator when InsP₃ is present. Maximum channel activity occurs when the concentration of free Ca²⁺ is 250 nM; there is a sharp decline in channel activity on either side of the maximum, with complete inhibition when cytosolic Ca²⁺ exceeds 5 μM (Fig. 3b, circles)^16,17. This bell-shaped regulation provides amplification of the initial InsP₃ signal, as well as negative-feedback inhibition of further InsP₃-stimulated Ca²⁺ release. The feedback inhibition is particularly important because it provides autoregulation and is essential for Ca²⁺ oscillations and for the propagation of regenerative intracellular Ca²⁺ waves¹⁸.

Figure 3.

Single channel open probability for type I and type III InsP₃R as a function of Ca²⁺ concentration. a, Channel activity for type III InsP₃R in the presence of 2 μM InsP₃, 0.5 mM ATP, 0.5 mM EGTA, 2 μM ruthenium red, and at 0 mV. Channel openings are shown as downward deflections. b, Single channel open probability of type I InsP₃R (circles) and type III InsP₃R (triangles). Data points for type I InsP₃R were taken from ref.17. Data for three experiments are shown for type III InsP₃R. Individual points with error bars are the mean ± s.e.m. for n = 2 (1 and 10 μM Ca²⁺) or n = 3 (0.01 and 0.1 μM Ca²⁺).

Does Ca²⁺ regulate type III InsP₃R in the same way? We hypothesized that type III InsP₃R should remain open at high cytoplasmic Ca²⁺ concentration ([Ca²⁺]) because InsP₃ binding to type III InsP₃R is not inhibited by elevated Ca²⁺ (refs ^{12, 13}). To test this idea, we altered the cytoplasmic [Ca²⁺] in the presence of a fixed InsP₃ concentration (2 μM). Like type I InsP₃R, type III InsP₃R was progressively activated as the cytoplasmic [Ca²⁺] increased to 250 nM. However, channel activity for type III InsP₃R remained high even when cytoplasmic [Ca²⁺] was raised to 100 μM (Fig. 3a, bottom traces; Fig. 3b, triangles). The mean open time for type III InsP₃R (6.5 ± 0.62 ms) was nearly identical to the reported value for type I InsP₃R (6.4 ± 1.0 ms)⁷ and was independent of the cytoplasmic Ca²⁺ concentration. Thus, cytoplasmic free Ca²⁺ regulates the two isoforms differently; although both isoforms show similar Ca²⁺-dependent activation, Ca²⁺-dependent inhibition is lacking for type III InsP₃R.

To investigate the physiological importance of this pattern of activation in an intact cell, we monitored the cytoplasmic [Ca²⁺] in RIN-5F cells which we stimulated with extracellular ATP that binds to P_2Y receptors and increases [Ca²⁺] through the InsP₃ cascade^19,20. RIN-5F cells responded to ATP stimulation with a single, transient increase in [Ca²⁺] (Fig. 4a). In single cells, stimulation with 100 μM ATP induced a 209 ± 24% increase in Fluo-3 fluorescence relative to baseline; this increase lasted for 15.2 ± 0.8 s (Table 1). The response was similar in Ca²⁺-free medium (Table 1). Similar, but slightly smaller, increases in Fluo-3 fluorescence were seen in cells stimulated with 10 μM or 1 μM ATP. There was no increase in [Ca²⁺] in response to stimulation with 0.1 μM ATP (Fig. 4b), and neither sustained nor repetitive increases in [Ca²⁺] (that is, Ca²⁺ oscillations) were seen in any of the cells. In contrast, hepatocytes, which contain only types I and II InsP₃R, produced Ca²⁺ oscillations after stimulation with 1 μM ATP (Fig. 4c), as shown previously²¹. RIN-5F cells that were serially stimulated with 10 μM and then 100 μM ATP responded only to the initial exposure to 10 μM ATP (n = 8). In addition, thapsigargin (2 μM) had a minimal effect on intracellular [Ca²⁺] when RIN-5F cells were pretreated with ATP (Fig. 4d). This finding provides direct evidence that activation of type III InsP₃R drains internal Ca²⁺ stores in RIN-5F cells. Furthermore, the magnitude of ATP-induced Ca²⁺ spikes is significantly greater in RIN-5F cells than in hepatocytes (Fig. 4d, f). Thus, positive feedback of Ca²⁺ on type III InsP₃R in RIN-5F cells causes rapid, massive, near-complete Ca²⁺ release, and results in more intense Ca²⁺ spikes that are of shorter duration than those that occur through the type I InsP₃R in hepatocytes.

Figure 4.

Ca²⁺ signalling patterns in RIN-5F cells and rat hepatocytes. a, Stimulation of a RIN-5F cell with 100 μM ATP induces a single intracellular Ca²⁺ transient. b, Dose–response curve for RIN-5F cells stimulated with ATP. Values indicate total numbers of cells stimulated. c, Stimulation of an hepatocyte with 1 μM ATP induces Ca²⁺ oscillations. d, ATP (100 μM) induces a single transient increase in Ca²⁺ in RIN-5F populations (peak change in fluorescence ratio, ΔR = 0.20 ± 0.01; n = 6). Subsequent treatment with thapsigargin (TG; 2 μM) has little effect on Ca²⁺ (ΔR = 0.03 ± 0.01; n = 6). e, Thapsigargin alone (2 μM) increases Ca²⁺ in RIN-5F cells (ΔR = 0.08 ± 0.01; n = 6). f, ATP (100 μM) induces a rapid, sustained increase in Ca²⁺ in hepatocyte populations (ΔR = 0.13 ± 0.02; n = 7). g, Thapsigargin (2 μM) increases Ca²⁺ in hepatocytes like ATP alone (ΔR = 0.10 ± 0.02; n = 7).

Table 1.

Ca²⁺ transients in RIN-5F cells after stimulation by ATP

ATP (μM)	Magnitude (% increase)	Duration (s)
100	209 ± 24 (n = 23)	15.2 ± 0.8 (n = 31)
100 (Ca2+-free)	200 ± 13 (n = 25)	14.8 ± 1.2 (n = 25)
10	160 ± 14 (n = 25)	16.8 ± 2.1 (n = 19)
1	121 ± 15 (n = 15)	13.0 ± 2.1 (n = 8)
0.1	0 (n = 32)	0 (n = 32)

RIN-5F cells were stimulated with ATP and changes in cytoplasmic Ca²⁺ in individual cells were measured by Fluo-3 fluorescence using confocal line scanning microscopy. The magnitude (per cent increase over baseline) and duration of the Ca²⁺ transients are expressed as mean ± s.e.m. (n).

We next compared subcellular Ca²⁺ release from type III InsP₃R in RIN-5F cells with subcellular Ca²⁺ release from type I InsP₃R in SKHep1 cells (a hepatoma cell line that expresses type I but not type III InsP₃R; our unpublished observation). Small amounts of InsP₃ were released in both cell types by flash photolysis of microinjected caged InsP₃. Release of InsP₃ in the RIN-5F cells always resulted in all-or-none global Ca²⁺ signalling, although increases in [Ca²⁺] began in focal subcellular regions before spreading to encompass the entire cell (Fig. 5a–d). In contrast, non-propagating increases in [Ca²⁺] could sometimes be elicited in SKHep1 cells by photorelease of minimal amounts of InsP₃ (Fig. 5e–g, i), similar to responses previously seen in pancreatic acinar and HeLa cells and Xenopus oocytes^10,11,22,23. Photorelease of larger amounts of InsP₃ induced a global Ca²⁺ response in SKHepI cells (Fig. 5h, j). These findings suggest that subcellular Ca²⁺ release via type III InsP₃R results in a positive-feedback cycle that leads to all-or-none Ca²⁺ signalling that spreads throughout the cell, whereas release through type I InsP₃R with low concentratons of InsP₃ results in localized, non-propagating increases in [Ca²⁺].

Figure 5.

Subcellular Ca²⁺ release differs between RIN-5F and SKHep1 cells. a, confocal image of a RIN-5F cell. b, Pseudocolour image of the cell loaded with Fluo-3. The same pseudocolour scale was used for b, c, and f–h. c, Line scan collected along the line indicated in b after flash photolysis (λ) of caged InsP₃. Ca²⁺ increases throughout the cell after photorelease of InsP₃ (representative of 13 experiments with flash duration 50–100 ms). No Ca²⁺ increase was detected in 6 experiments with flash duration <50 ms. Spatial resolution, 0.26 μm per pixel; temporal resolution, 6 ms per pixel. d, Release of caged InsP₃ results in a global increase in Ca²⁺. Trace duration in d, i and j is 2 s. e, f, Confocal (e) and pseudocolour (f) images of an SKHep1 cell. g, h, Confocal line scans of the cell during 30 and 60 ms flashes to photolyse caged InsP₃. A response is detected in only the left side of the cell after a short flash (g), and throughout the cell after a long flash (h). Responses were absent or minimal in only part of an SKHep1 cell in 7 experiments (flash duration, 3–50 ms), whereas global Ca²⁺ increases were seen in 12 experiments (flash duration, 5–60 ms). i, An increase occurs on the left side of the cell in g (1), but not on the right (2). j, After more photorelease of InsP₃ (h), similar Ca²⁺ increases occur at the same two subcellular locations.

The localized, subcellular Ca²⁺ signals in SKHep1 cells lasted from one to several seconds, a result that is typical for the duration of subcellular Ca²⁺ signals described in other mammalian cells^10,11,24. Localized increases in [Ca²⁺] of shorter duration are routinely seen in Xenopus oocytes, in which Ca²⁺ puffs typically last for 300–1,000 ms (ref. ²³). Although individual Ca²⁺ puffs are restricted to a small fraction of the total volume of an oocyte²³, they occupy a region that is larger than an entire mammalian cell. We believe that the localized elevation in [Ca²⁺] in SKHep1 cells is a reasonable mammalian cell equivalent of the localized Ca²⁺ signals described in oocytes.

Ca²⁺ concentrations in RIN-5F cells returned nearly to baseline after stimulation with ATP (Fig. 4a, d), although depletion of intracellular Ca²⁺ stores in other cells activates a robust store-operated Ca²⁺ current (I_soc)²⁵. Our findings suggest that I_soc is either small or absent in RIN-5F cells. We therefore stimulated RIN-5F cells with thapsigargin (2 μM) to deplete Ca²⁺ stores, and then removed Ca²⁺ from the medium. Thapsigargin induced a small but sustained increase in intracellular [Ca²⁺] (Fig. 4e); this increase was abolished by subsequent addition of extracellular EGTA. When RIN-5F cells were stimulated with thapsigargin in the absence of extracellular Ca²⁺, there was no sustained increase in intracellular [Ca²⁺]. Addition of thapsigargin to hepatocytes, in which I_soc has been described²⁵, induced a larger increase in intracellular [Ca²⁺] (Fig. 4g); the level of Ca²⁺ also returned to baseline in response to extracellular EGTA. These findings indicate that RIN-5F cells contain less I_soc than hepatocytes, which is interesting in light of a suggestion that the primary role of type III InsP₃R is to initiate I_soc (ref. ¹⁵) rather than Ca²⁺ transients. This hypothesis was based upon heterologous expression studies in Xenopus oocytes in which the protein seems to be targeted to the cell surface and may differ from the distribution found in RIN-5F cells (Fig. 1b) and other mammalian cells^5,9.

The high degree of homology among InsP₃R isoforms indicates that many of their functional properties ought to be comparable. We found that properties such as activation by InsP₃, the magnitude of the single-channel current, and activation by concentrations of Ca²⁺ less than 250 nM were quite similar. In addition, the sustained activity of type III InsP₃R at raised Ca²⁺ concentrations is similar to that reported for type I InsP₃R in the presence of high concentrations of InsP₃ (180 μM)¹⁷. In both cases, there is a lack of Ca²⁺-dependent inhibition, and the channel remains open even when the cytoplasmic Ca²⁺ exceeds 50 μM. In the presence of low concentrations of InsP₃ (<5 μM), however, the two isoforms have fundamentally distinct responses to cytoplasmic [Ca²⁺] greater than 250 nM (Fig. 3). If type III InsP₃R does not show Ca²⁺-dependent inhibition, what closes the channel? The primary level of regulation may be the control of InsP₃ generation and degradation because type III, like type I, InsP₃R only opens when InsP₃ is present. Further regulation of intracellular Ca²⁺ release by type III InsP₃R can occur by locally depleting the intracellular Ca²⁺ stores. Other mechanisms shown to be important for type I InsP₃R, such as phosphorylation and regulation by associated proteins²⁶, may also modulate the activity of type III InsP₃R.

Multiple isoforms of the InsP₃R are expressed in a variety of cell types¹⁴, but the physiological significance of this was unclear. The functional differences between type I and type III isoforms of the InsP₃R now indicate that each has a special role in the cell. Type I InsP₃R, with both Ca²⁺-dependent activation and inhibition, is well suited for establishing Ca²⁺ oscillations^16,21,27, where the frequency of Ca²⁺ transients can be modulated when InsP₃ concentrations are increased^17,27. In contrast, type III InsP₃R, by remaining open in the presence of high [Ca²⁺] (Fig. 3), initiates a rapid, large, and almost total release of Ca²⁺ from intracellular stores as long as InsP₃ is present (Fig. 4a, d). Thus, type III InsP₃R alone will not support a regenerative response, but its properties make it well suited to initiate intracellular Ca²⁺ signals. In support of this, type III InsP₃R has been localized to the apical region of epithelial cells, the trigger zone from which intracellular Ca²⁺ waves originate^9–11. Thus, the single-channel properties of type III InsP₃R are well adapted to its role as the starting gate for Ca²⁺ signals in the cell.

Methods

Western blots and immunocytochemistry

Immunoblots were probed with antibodies against type I (C-19; custom produced by Research Genetics) or type III (Transduction Lab) InsP₃R. Blots were visualized using ECL (Kirkegaard & Perry). For immunocytochemistry, the same primary antibodies were used to probe for type I and type III InsP₃R in fixed cells. The secondary antibody contained a fluorescent label (FITC) that was observed by confocal microscopy. Nonspecific staining was determined by using only the secondary antibody.

Single-channel recordings

Endoplasmic reticulum vesicles from RIN-5F cells were prepared using the protocol for cerebellum described previously⁷. Vesicles were fused into planar lipid bilayers composed of phosphatidylethanolamine and phosphatidylserine (3:1, w/w; Avanti Polar Lipids) so that the cis and trans chambers corresponded to the cytosol and lumen of the endoplasmic reticulum respectively. Cytoplasmic bilayer solutions contained 110 mM Tris and 250 mM HEPES (pH 7.35), and lumenal solutions contained 55 mM Ba(OH₂) and 250 mM HEPES (pH 7.35). The trans chamber was held at virtual ground and the transmembrane voltage was maintained at 0 mV. Single-channel currents were recorded under voltage-clamp conditions using a patch-clamp amplifier (BC-525B, Warner Instruments) and stored on VHS tape (Instrutech). Data were filtered at 1 kHz and digitized at 4 kHz for computer analysis using pClamp 6.0.3 (Axon Instruments). For the I–V curve, the membrane potential was clamped at values between 5 and −30 mV. The amplitude of channel openings at each voltage was determined by fitting the data (100–1,600 openings) with a gaussian function. Single-channel conductance was determined by linear regression. An extrapolation to 0 pA to estimate the reversal potential is not included because barium (55 mM Ba(OH)₂) was in the trans chamber only, and current through the InsP₃-gated channel can only flow in one direction from trans to cis. Current in the opposite direction was not detected. If more positive voltages were included, the current–voltage relationship would begin to curve, and the slope conductance would be underestimated. These are the same experimental conditions as used previously for type I InsP₃R (ref. ⁷). For the Ca²⁺ dependence curve, calibrated amounts of CaCl₂ were added to the cytoplasmic solution to obtain the desired free Ca²⁺ concentration. We estimated the number of active channels in each bilayer using a statistical model that is dependent on the maximum number of channels observed simultaneously, and then calculated the open probability for a single channel using this value^7,17. Open probability data for type I InsP₃R were fitted using the ‘2-InsP₃/2-Ca²⁺’ model¹⁷, whereas open probability data for type III InsP₃R were fitted according to Michaelis–Menten kinetics with K_m = 0.3 and P_max = 5.4: P_o = (P_max ·[Ca²⁺])/(K_m + [Ca²⁺].

Cytoplasmic Ca²⁺ measurements

Cytoplasmic Ca²⁺ was measured in single RIN-5F cells or rat hepatocytes using confocal line scanning microscopy²⁸ or in cell populations using ratio spectrofluorimetry²⁹. For single-cell studies, RIN-5F cells or hepatocytes³⁰ were plated onto glass coverslips, incubated at 37 °C, and loaded with Fluo-3/AM (6 μM). Coverslips containing the cells were transferred to a perfusion chamber on the stage of a BioRad MRC-600 confocal microscope and observed using a 20× objective³⁰. Increases in [Ca²⁺] are expressed as a percentage of baseline fluorescence³⁰. Cells were examined first under control conditions, then in the presence of 0.1–100 μM ATP. In selected experiments, RIN-5F cells were stimulated with ATP in Ca²⁺-free medium containing 1 mM EGTA; similar results were obtained in the presence and absence of extracellular Ca²⁺. For population studies, cells were loaded with Fura-2/AM (10 μM), then maintained at 37 °C in a cuvette. Cells were excited at 340 and 370 nm; fluorescence emission was detected at 485 nm using a PTI DeltaRAM system (these excitation and emission wavelengths were chosen to optimize Fura-2 ratio measurements with this system). Ratios were determined after background subtraction²⁹.

Subcellular Ca²⁺ release

RIN-5F or SKHep1 cells were placed in a perfusion chamber on the stage of a BioRad confocal microscope, then individual cells were pressure-microinjected with a solution containing 1 mM caged InsP₃, 1 mM Fluo-3, 1 mM HEPES, and 150 mM KCl. Cells were given 5–10 min to recover from injection, then InsP₃ was photoreleased using a custom-built system that couples a mercury lamp to a 1-mm quartz fibreoptic cable through a high-speed shutter and filterwheel while cells were observed using confocal line scanning microscopy²⁸. SKHep1 cells were kindly provided by D. Spray.