Allosteric regulation by cytoplasmic Ca²⁺ and IP₃ of the gating of IP₃ receptors in permeabilized guinea-pig vascular smooth muscle cells

Abstract

1. The potentiation by Ca²⁺ of inositol 1,4,5-trisphosphate (IP₃)-induced Ca²⁺ release was studied by measuring luminal Ca²⁺ concentrations of the Ca²⁺ stores using a fluorescent Ca²⁺ indicator, furaptra, in permeabilized smooth muscle cells.

2. Ca²⁺ release at 10 μM IP₃ was potentiated by an increase in the cytoplasmic Ca²⁺ concentration in the presence of 10 mM EGTA. This effect was not due to the pharmacological effect of EGTA, because changes in the EGTA concentration at a constant Ca²⁺ concentration had no effect on the Ca²⁺ release rate.

3. With an increase in the cytoplasmic Ca²⁺ concentration from 30 to 630 nM, the Ca²⁺ release rate at a saturating IP₃ concentration increased 110-fold and the EC₅₀ for IP₃ increased from 0.07 to 1.0 μM. It was also indicated that the relationship between Ca²⁺ concentration and Ca²⁺ release rate was shifted towards higher Ca²⁺ concentrations at higher IP₃ concentrations.

4. These results suggest that IP₃ and submicromolar concentrations of Ca²⁺ allosterically lower the affinity of the IP₃ receptor for each other and are both required for IP₃ receptor activation. These properties enable the IP₃ receptors to detect simultaneous increases in IP₃ and Ca²⁺ concentrations.

Agonist-activated cells often show intermittent increases in intracellular Ca²⁺ concentration, or Ca²⁺ oscillations, which are usually accompanied by intracellular and/or intercellular Ca²⁺ waves. These properties of the Ca²⁺ signalling in stimulated cells resemble the membrane action potentials that exhibit temporal oscillations and propagate as spatial waves. In analogy to the generation of action potentials in excitable cells, it has been thought that the underlying mechanism of the Ca²⁺ waves and oscillations is the presence of a Ca²⁺-based excitable medium, in which an increase in Ca²⁺ concentration activates release of Ca²⁺ from its stores in a regenerative manner. Although the mechanism of these complex Ca²⁺ mobilization patterns is not fully understood, Ca²⁺-mediated feedback regulation of ryanodine receptor (RyR) or IP₃ receptor (IP₃R) activity is likely to be a key component of the excitable medium in many types of cells (Berridge, 1993). Indeed, RyR has been shown to function by Ca²⁺-induced Ca²⁺ release (CICR) and this seems to be the mechanism of Ca²⁺ oscillations and waves in striated muscle cells (Endo, Tanaka & Ogawa, 1970; Takamatsu & Wier, 1990). It has been shown that IP₃R activity is also controlled by changes in the cytoplasmic Ca²⁺ concentration, in a biphasic manner: Ca²⁺ at submicromolar concentrations potentiates IP₃-induced Ca²⁺ release, while Ca²⁺ at higher concentrations inhibits it (Iino, 1990; Bezprozvanny, Watras & Ehrlich, 1991; Finch, Turner & Goldin, 1991; Bootman, Missiaen, Parys, De Smedt & Casteels, 1995). Furthermore, it has been demonstrated that IP₃ induces Ca²⁺ release from the Ca²⁺ stores in permeabilized smooth muscle cells in a regenerative manner, via the Ca²⁺-mediated feedback regulation of IP₃R activity (Iino & Endo, 1992). However, the quantitative aspects of the IP₃R activation are not fully understood and it is not yet known how Ca²⁺ potentiates IP₃R activity.

To elucidate the mechanism of the Ca²⁺-mediated potentiation of IP₃R activity, we analysed the Ca²⁺ dependence of IP₃-induced Ca²⁺ release using a recently developed method for real-time measurement of the luminal Ca²⁺ concentrations of the Ca²⁺ stores (Hirose & Iino, 1994). Our results show that Ca²⁺ potentiates IP₃R activity not by increasing IP₃R affinity for IP₃, but because it is absolutely required for IP₃R activity. This property enables the IP₃R to detect a simultaneous increase in the concentrations of two different intracellular signals: Ca²⁺ and IP₃.

METHODS

Preparation

Guinea-pigs (weight, ∼300 g each) were stunned and exsanguinated. Segments of the portal vein were dissected from the liver and transferred to a dissecting chamber containing physiological salt solution (PSS). The vein was longitudinally cut open, and the endothelium was removed by rubbing with a small piece of paper. Thin smooth muscle bundles (3 mm in length, 150–250 μm in width and 50–60 μm in thickness) were obtained after removal of the adventitial tissue, and were tied to a thin stainless steel wire. We employed furaptra as an fluorescent indicator of Ca²⁺ concentration within the Ca²⁺ stores, because furaptra is a low-affinity Ca²⁺ indicator (∼48 μM at ionic strength of 0.2 M, pH 7; K. Hirose & M. Iino, unpublished observation) and is thus suitable for measurement of the high Ca²⁺ concentrations expected inside the Ca²⁺ stores. The specimens were incubated in PSS containing 20–40 μM furaptra-AM (the acetoxymethyl ester form of furaptra) for 3–5 h at 35°C. The furaptra-loaded fibre bundles were then permeabilized by incubation with 40 μM β-escin in relaxing solution to wash out the furaptra in the cytoplasm. This procedure enables measurement of the concentrations of Ca²⁺ within the intracellular organelles.

Experimental apparatus

The apparatus used for the experiments has been described previously (Hirose & Iino, 1994). Briefly, the specimen tied to the stainless steel wire was inserted into a small glass capillary (i.d. 400 μm) which was attached to the cuvette holder of a fluorescence spectrophotometer (CAF110, JASCO, Tokyo, Japan). The specimen was then illuminated with 340 nm and 375 nm wavelength light alternately and the emitted light was directed to a photomultiplier tube. The ratio of the fluorescence intensity at 340 nm excitation to that at 380 nm excitation was used as an indicator of the luminal Ca²⁺ concentration. One end of the glass capillary was connected to the common output of an electrically driven valve which had sixteen channel inputs.

The other end of the capillary was connected to two peristaltic pumps. One of the pumps was used to change the solution around the specimen rapidly (within 1 s). Continuous slow flow of the solution over the specimen was driven by another peristaltic pump.

Protocol

The solution within the capillary was sequentially exchanged to load and release Ca²⁺ from the Ca²⁺ stores of the specimen. After the fibre bundle was bathed in relaxing solution for 60 s, Ca²⁺ was loaded into the Ca²⁺ stores by application of loading solution, containing 100 nM Ca²⁺ and 0.5 mM MgATP, for 25 s. The Ca²⁺ loading was terminated by the exchange of loading solution with G10RMg0 solution, which contained 10 mM EGTA with neither Ca²⁺ nor MgATP. (All the solutions used in the subsequent steps also contained no MgATP.) The G10RMg0 solution was exchanged with G1RMg0 solution in which the concentration of EGTA was reduced to 1 mM. After 30 or 60 s, the concentration of Ca²⁺ was increased by application of assay solution containing 10 mM EGTA and various amounts of Ca²⁺. IP₃ (30 nM-10 μM) was subsequently applied for 120 s in the same solution. The Ca²⁺ stores were then completely depleted by 120 s application of 10 μM IP₃ in the presence of 300 nM Ca²⁺. After IP₃ and Ca²⁺ were washed out for 60 s, the specimen was again bathed in the relaxing solution to start a new run. The Ca²⁺ loading and Ca²⁺ release sequences were repeated several times for each specimen.

Measurement of IP₃R activity

The IP₃R activity was evaluated in terms of the initial Ca²⁺ release rate as described previously (Hirose & Iino, 1994). The observed changes in ratio of fluorescence intensities were normalized so that 1 and 0 corresponded, respectively, to the values just before the application of IP₃ and after complete depletion by 10 μM IP₃ for 120 s.

The initial part of the normalized time course, where the normalized fluorescence ratio signal remained between 1.0 and 0.8, was fitted by a single exponential function, e^-rt. The rate constant, r, thus estimated was used as an index of the extent of IP₃R activation.

Solutions

The solutions used for the present experiments were essentially the same as those previously described (Iino, 1990; Hirose & Iino, 1994). PSS contained (mM): 150 NaCl, 4 KCl, 1 MgCl₂, 2 CaCl₂, 5.6 glucose and 5 Hepes (pH 7.4). Relaxing solution contained (mM): 116 potassium methanesulphonate, 3.31 ATP, 0.554 magnesium methanesulphonate and 1 EGTA. Loading solution contained (mM): 112 potassium methanesulphonate, 3.32 ATP, 0.564 magnesium methanesulphonate, 1.96 CaEGTA and 8.04 EGTA. G10RMg0 solution contained 112 mM potassium methanesulphonate and 10 mM EGTA. CaG10RMg0 solution contained 112 mM potassium methanesulphonate and 10 mM CaEGTA. G1RMg0 contained 139 mM potassium methanesulphonate and 1 mM EGTA. All solutions except for PSS contained (mM): 20 NaN₃, 1 EGTA and 20 Pipes (pH 7.0). Assay solutions containing Ca²⁺ at various concentrations were prepared by mixing CaG10RMg0 solution and G10RMg0 solution at appropriate ratios. The ratios of the volume of CaG10RMg0 solution to the total volume of the assay solution were: 0.072, 0.196, 0.329, 0.434 and 0.608 for 30, 100, 200, 300 and 630 nM Ca²⁺, respectively.

Chemicals

Furaptra-AM (mag-fura-2 AM) was purchased from Molecular Probes, and IP₃ from Dojin-do (Kumamoto, Japan). All other drugs were purchased either from Sigma or from Wako Chemicals (Tokyo, Japan).

RESULTS

Ca²⁺-mediated potentiation of IP₃-induced Ca²⁺ release

Figure 1A shows the time courses of Ca²⁺ release from the Ca²⁺ stores of permeabilized smooth muscle cells, following application of 10 μM IP₃ at various cytoplasmic Ca²⁺ concentrations in the absence of MgATP. The Ca²⁺ release rate increased as the concentration of Ca²⁺ was increased, within the range used in the experiments (30–630 nM). Figure 1B shows the results compiled from five preparations. It can be seen that the Ca²⁺ release rate at 10 μM IP₃ increased markedly with increasing cytoplasmic Ca²⁺ concentration. This observation is consistent with previous reports based on different methods (Iino, 1990; Bezprozvanny et al. 1991; Finch et al. 1991). In the absence of IP₃, changes in cytoplasmic Ca²⁺ concentration had no effect on the time course of Ca²⁺ release (not shown). Thus, the observed potentiation of Ca²⁺ release by Ca²⁺ did not involve Ca²⁺-induced Ca²⁺ release (CICR) via ryanodine receptors (RyRs). This result is consistent with the results of a previous study showing CICR occurs only when the cytoplasmic Ca²⁺ concentration exceeds 1 μM in smooth muscle cells (Iino, 1989).

Figure 1. IP₃-induced Ca²⁺ release at various cytoplasmic Ca²⁺ concentrations.

A, time courses of Ca²⁺ release at the indicated cytoplasmic Ca²⁺ concentrations. Application of 10 μM IP₃ was started at the time point indicated by the arrow. B, the initial rate of Ca²⁺ release was plotted against the cytoplasmic Ca²⁺ concentration (mean ±s.e.m., n = 5).

We did not measure the Ca²⁺ release rate at higher cytoplasmic Ca²⁺ concentrations (> 1 μM), at which Ca²⁺ has been reported to inhibit IP₃-induced Ca²⁺ release (Iino, 1990). This is because application of high concentrations of Ca²⁺ (> 1 μM) alone increased the fluorescence intensity ratio of furaptra even in the absence of MgATP. The change in the fluorescence intensity ratio was not due to passive influx of Ca²⁺ into the Ca²⁺ stores, because the increase in the fluorescence intensity ratio was reversed upon removal of Ca²⁺ in the absence of IP₃. Therefore, the change in the furaptra fluorescence intensity ratio at high concentrations of Ca²⁺ (> 1 μM) is likely to be due to the binding of Ca²⁺ to compartmentalized furaptra which is accessed only at high cytoplasmic Ca²⁺ concentrations. Although this compartment might be mitochondria, which are known to respond to high concentrations of Ca²⁺ and have been shown to accumulate Ca²⁺-sensitive fluorescent dyes including furaptra (Hoffer, Schlue, Curci & Machen, 1995; Golovina & Blaustein, 1997), we did not study the possibility further. We thus analysed the Ca²⁺ release only at < 1 μM Ca²⁺, ensuring the observed time courses were free from such effects.

Effect of EGTA on Ca²⁺ release rate

Ca²⁺ chelating agents such as BAPTA have been reported to inhibit the binding of IP₃ to its receptor, resulting in attenuation of Ca²⁺ release even at a constant Ca²⁺ concentration (Richardson & Taylor, 1993). This pharmacological effect of Ca²⁺ chelating agents might interfere with the observed potentiation by changing the Ca²⁺ concentration, because the concentration of Ca²⁺-unbound EGTA, which has been reported to inhibit IP₃ binding more strongly than Ca²⁺-bound EGTA, decreased when the Ca²⁺ concentration was raised in the experiments presented in Fig. 1. Although the inhibitory effect of EGTA on IP₃ binding has been reported to be weaker than those of other chelators, we tested the possibility that the potentiation of Ca²⁺ release was due to the pharmacological effects of EGTA.

We found that the potentiating effect of Ca²⁺ (300 nM) did not change when the total EGTA concentration was increased from 10 mM to 14.3 mM or decreased to 4.3 mM so that the concentration of Ca²⁺-unbound or Ca²⁺-bound EGTA, respectively, was the same as that at 100 nM Ca²⁺ with a 10 mM total EGTA concentration (Fig. 2). The results show that the extent of potentiation depends only on the free Ca²⁺ concentration and is independent of the concentration of Ca²⁺-bound and Ca²⁺-unbound EGTA, and indicate that the observed potentiation by Ca²⁺ of IP₃-induced Ca²⁺ release is not the result of the inhibitory effect of EGTA on IP₃ binding.

Figure 2. Extent of Ca²⁺-mediated potentiation of IP₃-induced Ca²⁺ release is independent of EGTA concentration.

The total EGTA concentration at 300 nM Ca²⁺ was changed from 10 mM to either 4.5 mM or 14.3 mM for adjustment of the concentration of Ca²⁺-bound or Ca²⁺-unbound EGTA, respectively, to that at 100 nM Ca²⁺. The IP₃ concentration was 100 nM. Means ±s.e.m., n = 5.

Effect of Ca²⁺ on IP₃ sensitivity of Ca²⁺ release via IP₃R

We then attempted to elucidate the mechanism by which Ca²⁺ potentiates IP₃R activity. We first considered two extreme cases. One possible mechanism is that Ca²⁺ increases the affinity of IP₃R for IP₃. The other is that Ca²⁺ increases the maximal level of IP₃R channel activity without changing the affinity of IP₃R for IP₃.

In the former mechanism, the maximal Ca²⁺ release rate at saturating IP₃ concentrations should be independent of the Ca²⁺ concentration, although the EC₅₀ for IP₃ should decrease with increasing Ca²⁺ concentration. However, in the latter case, the maximum level of IP₃R activation cannot be attained at low Ca²⁺ concentrations even with saturating concentrations of IP₃, and the EC₅₀ for IP₃ should be constant regardless of the Ca²⁺ concentration.

To test the above hypotheses, we compared the IP₃ dependence of Ca²⁺ release at high and low concentrations of Ca²⁺. Ca²⁺ was released at various concentrations of IP₃ at both 300 and 100 nM Ca²⁺ (Fig. 3A). At 300 nM Ca²⁺, the rate of Ca²⁺ release increased markedly with increasing concentration of IP₃, whereas at 100 nM Ca²⁺, the effect of an increase in the concentration of IP₃ was far weaker. Figure 3B shows the relationship between IP₃ concentration and initial rate of Ca²⁺ release at 100 and 300 nM Ca²⁺. We fitted Hill's equation to the data points,

where r_max, K and n represent maximal Ca²⁺ release rate, Ca²⁺ concentration required for half-maximal activation of Ca²⁺ release (EC₅₀), and Hill coefficient, respectively. r_max was ∼6.8 times higher at 300 nM Ca²⁺ than at 100 nM Ca²⁺. Vertical scaling of the IP₃ dependence curve at 100 nM Ca²⁺ (dotted line in Fig. 3B) by 6.8 times shows that the EC₅₀ for IP₃ was higher at 300 nM Ca²⁺ than at 100 nM Ca²⁺.

Figure 3. Ca²⁺ concentration dependence of IP₃-induced Ca²⁺ release.

A, time course of Ca²⁺ release induced by IP₃ at cytoplasmic Ca²⁺ concentrations of 300 nM (left) and 100 nM (right). B, the Ca²⁺ release rate was calculated from experimental results similar to those in A and plotted against the IP₃ concentration. Open and filled circles represent Ca²⁺ release rates at 100 and 300 nM Ca²⁺, respectively. The data were fitted by Hill's equation to give continuous curves. The fitted curve for 100 nM Ca²⁺ was scaled up 6.8-fold (dotted curve). C, the IP₃ concentration required for half-maximal activation of Ca²⁺ release (EC₅₀) was plotted against the cytoplasmic Ca²⁺ concentration. D, the maximal Ca²⁺ release rate was plotted against the Ca²⁺ concentration.

We conducted similar experiments at various cytoplasmic Ca²⁺ concentrations and derived the relationship between the EC₅₀ for IP₃ and the Ca²⁺ concentration shown in Fig. 3C. The EC₅₀ for IP₃ increased from 0.07 to 1.0 μM with increasing cytoplasmic Ca²⁺ concentration from 30 to 630 nM. The maximal Ca²⁺ release rate also increased 110-fold with increasing concentration of Ca²⁺ (Fig. 3D), indicating that the activity of Ca²⁺ release channels is enhanced by Ca²⁺ even at a saturating concentration of IP₃. In the presence of a low cytoplasmic Ca²⁺ concentration (below 100 nM), IP₃ very weakly induced Ca²⁺ release. These results are inconsistent with the two extreme cases that we initially considered.

Both Ca²⁺ and IP₃ are required for maximal IP₃R activation

To obtain a further understanding of how IP₃R is activated, we replotted the Ca²⁺ release rate against the Ca²⁺ and IP₃ concentrations in a three-dimensional graph (Fig. 4A, spheres). It is obvious that the Ca²⁺ release rate increased markedly only when both the Ca²⁺ and the IP₃ concentration were increased. Neither Ca²⁺ nor IP₃ induced Ca²⁺ release in the absence of the other. These results indicate that Ca²⁺ and IP₃ are simultaneously required for IP₃R activation in smooth muscle cells.

Figure 4. Allosteric activation of IP₃ receptors by Ca²⁺ and IP₃.

A, the Ca²⁺ release rate was plotted against the Ca²⁺ and IP₃ concentrations. Experimental data (spheres) are fitted by the equation described in B, as shown by the curved surface. B, a possible mechanism of IP₃ receptor activation. The channel has n Ca²⁺ binding sites and m IP₃ binding sites. K₁ and K₃ represent K_d for Ca²⁺ binding to IP₃-unbound and IP₃-bound forms of the receptor, respectively. K₂ and K₄ represent K_d for IP₃ binding to Ca²⁺-unbound and Ca²⁺-bound forms of the receptor, respectively.

DISCUSSION

In this study, we found that the submicromolar Ca²⁺-mediated potentiation of Ca²⁺ release via IP₃R is not caused by an increase in the affinity of IP₃R for IP₃, and that cytoplasmic Ca²⁺ and IP₃ are simultaneously required for the gating of the IP₃R Ca²⁺ release channel. We showed that the cytoplasmic Ca²⁺ concentration in a submicromolar range determines the level of IP₃R activity at near-saturating IP₃ concentrations and that at the same time the increasing Ca²⁺ concentration lowers the affinity of IP₃R for IP₃ (Fig. 3).

Four-state model of IP₃R activation

We constructed a minimum steady-state model to account for the observed features of IP₃-induced Ca²⁺ release. The essential property of the model is that the binding of either Ca²⁺ or IP₃ alone to the IP₃R does not induce IP₃R activation, but simultaneous binding of both Ca²⁺ and IP₃ to IP₃R leads to IP₃R activation. We speculated that this simultaneous binding is essential, based on our finding that the rate of Ca²⁺ release is extremely low, even at an IP₃ concentration as high as 10 μM, when the Ca²⁺ concentration is low (Figs 3D and 4A). Another important feature of our model is that the binding of Ca²⁺ and IP₃ to IP₃R is allosterically controlled by the binding of IP₃ and Ca²⁺, respectively. We included this in our model, based on our finding that the EC₅₀ for IP₃ increased with increasing Ca²⁺ concentration (Fig. 3C). We noted a similar change in the Ca²⁺ dependence of Ca²⁺ release with an increase in the IP₃ concentration (data not shown). A four-state model that accommodates the above considerations is shown in Fig. 4B. The stoichiometry of IP₃ and Ca²⁺ binding was assumed to be m and n, respectively. The channel is open only in the Ca²⁺- and IP₃-bound state. The Ca²⁺ release rate is thus described as:

where r_max is the maximal Ca²⁺ release rate, K₁ and K₃ are the K_d for Ca²⁺ binding to IP₃-unbound and -bound IP₃R, respectively, and K₂ is the K_d for IP₃ binding to Ca²⁺-unbound IP₃R. We found that the model describes the important features of the experimental data (Fig. 4A, curved surface). Using a least-squares fitting method of a mathematics software package (Mathematica, Worfram Research, Inc., IL, USA) running on a personal computer, we obtained the following parameter values for each reaction: K₁, ∼154 nM; K₂, ∼1.6 μM; K₃, ∼403 nM; K₄ (the K_d for IP₃ binding to Ca²⁺-bound IP₃R), ∼201 nM; m, ∼1; and n, ∼2. According to this model, a Ca²⁺-bound IP₃R has about 8-fold lower affinity for IP₃ than does a Ca²⁺-unbound IP₃R. The affinity of the IP₃R for Ca²⁺ also decreased about 2.5-fold upon binding of IP₃.

Although our model successfully describes the present experimental data, it does not describe all the features of the Ca²⁺ dependence of IP₃R within the entire physiological range of Ca²⁺ concentrations. At micromolar concentrations, Ca²⁺ has an inhibitory effect on IP₃R activity (Iino, 1990; Bezprozvanny et al. 1991; Finch et al. 1991; Bootman et al. 1995). Unfortunately we could not study the Ca²⁺-mediated inhibition due to current experimental limitations. There are several possible ways in which the present model could be modified to include the inhibitory effect of Ca²⁺. One model is that Ca²⁺ binding to an inhibitory site alters the equilibrium of the four states in our model, so lowering the affinity of Ca²⁺ and/or IP₃ for their respective activation sites, resulting in a decreased population of active IP₃Rs (allosteric inhibition). In another model, the binding of Ca²⁺ to the inhibitory site occurs only in the IP₃R that binds Ca²⁺ at activation sites and IP₃ (sequential inhibition). Clearly we need more information to discriminate between these and other possibilities.

Comparison of the present results with those of the IP₃ binding studies

Several biochemical reports regarding Ca²⁺-mediated regulation of IP₃ binding to the IP₃R have been published. Benevolensky et al. demonstrated the presence of high- and low-affinity IP₃ binding sites in sarcoplasmic reticulum vesicles from aortic smooth muscle (Benevolensky, Moraru & Watras, 1994). The K_d of the low-affinity binding sites, accounting for > 90% of the total IP₃ binding sites, depended on the cytoplasmic Ca²⁺ concentration. The K_d was 49.3 nM in the absence of Ca²⁺ and 155 nM in the presence of 2 μM Ca²⁺. Similar Ca²⁺-mediated inhibition of IP₃ binding has been detected in cerebellar IP₃R (Worley, Baraban, Supattapone, Wilson & Snyder, 1987; Hannaert-Merah, Coquil, Combettes, Claret, Mauger & Champeil, 1994). Only a single class of IP₃ binding sites was detected in IP₃R purified from human IP₃R type 1 synthesized in a baculovirus expression vector system, and a Ca²⁺-dependent increase in K_d for IP₃, from 79 nM at 3 nM Ca²⁺ to 312 nM at 1.4 μM Ca²⁺, has been reported (Yoneshima, Miyawaki, Michikawa, Furuichi & Mikoshiba, 1997). There are, however, contradictory results for several cell types, in which an increase in the Ca²⁺ concentration causes a decrease in K_d for IP₃ (Marshall & Taylor, 1994). The effect of Ca²⁺ is complicated in rat basophilic leukaemia (RBL) cells, which have low- and high-affinity IP₃ binding sites. An increase in the Ca²⁺ concentration from 100 to 500 nM results both in the conversion of a portion of the low-affinity sites to high-affinity sites and an increase in the K_d of the low-affinity sites (Watras, Moraru, Costa & Kindman, 1994). The K_d for IP₃ binding to IP₃R type 3 decreased with an increase in Ca²⁺ concentration (Yoneshima et al. 1997). Thus, the mode of Ca²⁺-dependent IP₃ binding may depend on the subtype of the IP₃R and/or cell type-specific post-translational modification of the IP₃R. IP₃R type 1 seems to be dominantly expressed in smooth muscle cells (Chadwick, Saito & Fleischer, 1990; De Smedt et al. 1994; Islam, Yoshida, Koga, Kojima, Kangawa & Imai, 1996). If this is also the case in guinea-pig portal vein, the present results seem to be in general agreement with the results of binding studies, in that the lowering of the sensitivity for IP₃ is accompanied by an increase in the Ca²⁺ concentration. It would be interesting to examine the Ca²⁺ dependence of IP₃-induced Ca²⁺ release in cells expressing only IP₃R type 3.

Physiological significance of the requirement for both Ca²⁺ and IP₃ for IP₃R activation

The properties of the IP₃R revealed in the present study are consistent in many ways with the role of IP₃R as one of the key components for inducing Ca²⁺ oscillations. Firstly, in the presence of low concentrations of Ca²⁺, high concentrations of IP₃ activate IP₃R only weakly. This is consistent with the baseline Ca²⁺ oscillations which are a characteristic pattern of Ca²⁺ oscillations seen in many cell types (Berridge, 1993). In these baseline Ca²⁺ oscillations, after each Ca²⁺ spike the Ca²⁺ concentration returns to the baseline level. During the period between two successive spikes, the Ca²⁺ release rate should remain low in the presence of IP₃. This may be enabled by the low level of IP₃R activity in the presence of a low concentration of Ca²⁺. Secondly, the Ca²⁺ concentration dependence of the Ca²⁺ release rate had a Hill coefficient of ∼2 (also, n≈ 2 in our model fitting), suggesting that IP₃R activation requires simultaneous binding of two Ca²⁺ ions. The strong Ca²⁺ concentration dependence is advantageous for generation of Ca²⁺ spikes. Thirdly, the extent of Ca²⁺-mediated control of IP₃R activity is IP₃ concentration-dependent. In the presence of a low concentration of IP₃, Ca²⁺ alone activates the IP₃R so weakly that Ca²⁺ spikes are not generated. With an increase in the IP₃ concentration, Ca²⁺ becomes effective in inducing Ca²⁺ release. This IP₃ requirement may explain the requirement for receptor activation to start the cytoplasmic Ca²⁺ oscillation.

It seems pertinent here to mention mechanisms other than Ca²⁺-mediated potentiation of IP₃R activity that participate in the generation of Ca²⁺ oscillations. Firstly, the falling phase of the Ca²⁺ oscillation clearly requires inhibitory mechanisms of Ca²⁺ release that are still elusive. One possible mechanism for the termination of each Ca²⁺ oscillation may be the depletion of the Ca²⁺ stores, which results in a decreased Ca²⁺ flux rate and a decrease in the extent of Ca²⁺-mediated positive feedback. Another candidate mechanism may be Ca²⁺-mediated inhibition of IP₃R (Iino, 1990; Bezprozvanny et al. 1991; Finch et al. 1991; Bootman et al. 1995). After Ca²⁺ spikes are generated, the Ca²⁺ concentration around the Ca²⁺ stores may become high enough to terminate Ca²⁺ release via the Ca²⁺-mediated inhibitory mechanism. Secondly, CICR mechanism via RyRs may be involved in the Ca²⁺ oscillations and waves, as is the case in striated muscle cells (Endo et al. 1970; Takamatsu & Wier, 1990). Because activation of CICR has been shown to require a higher Ca²⁺ concentration than activation of IP₃R (Iino, 1989), CICR has the potential to play a role in the amplification of Ca²⁺ release after the IP₃R is activated. However, RyRs are not always expressed in cells in which the production of IP₃ is required for the generation of Ca²⁺ oscillations and waves. Furthermore, CICR in smooth muscle cells that express RyRs contributes little in the generation of Ca²⁺ waves (Iino, Yamazawa, Miyashita, Endo & Kasai, 1993).

In conclusion, the requirement for both Ca²⁺ and IP₃ for IP₃R activation is important for the generation of Ca²⁺ spikes and Ca²⁺ oscillations, in which a multitude of molecular mechanisms are involved.