Regulation of Ca²⁺ Release by InsP₃ in Single Guinea Pig Hepatocytes and Rat Purkinje Neurons

Abstract

The repetitive spiking of free cytosolic [Ca²⁺] ([Ca²⁺]_i) during hormonal activation of hepatocytes depends on the activation and subsequent inactivation of InsP₃-evoked Ca²⁺ release. The kinetics of both processes were studied with flash photolytic release of InsP₃ and time resolved measurements of [Ca²⁺]_i in single cells. InsP₃ evoked Ca²⁺ flux into the cytosol was measured as d[Ca²⁺]_i/dt, and the kinetics of Ca²⁺ release compared between hepatocytes and cerebellar Purkinje neurons. In hepatocytes release occurs at InsP₃ concentrations greater than 0.1–0.2 μM. A comparison with photolytic release of metabolically stable 5-thio-InsP₃ suggests that metabolism of InsP₃ is important in determining the minimal concentration needed to produce Ca²⁺ release. A distinct latency or delay of several hundred milliseconds after release of low InsP₃ concentrations decreased to a minimum of 20–30 ms at high concentrations and is reduced to zero by prior increase of [Ca²⁺]_i, suggesting a cooperative action of Ca²⁺ in InsP₃ receptor activation. InsP₃-evoked flux and peak [Ca²⁺]_i increased with InsP₃ concentration up to 5–10 μM, with large variation from cell to cell at each InsP₃ concentration. The duration of InsP₃-evoked flux, measured as 10–90% risetime, showed a good reciprocal correlation with d[Ca²⁺]_i/dt and much less cell to cell variation than the dependence of flux on InsP₃ concentration, suggesting that the rate of termination of the Ca²⁺ flux depends on the free Ca²⁺ flux itself. Comparing this data between hepatocytes and Purkinje neurons shows a similar reciprocal correlation for both, in hepatocytes in the range of low Ca²⁺ flux, up to 50 μM · s⁻¹ and in Purkinje neurons at high flux up to 1,400 μM · s⁻¹. Experiments in which [Ca²⁺]_i was controlled at resting or elevated levels support a mechanism in which InsP₃-evoked Ca²⁺ flux is inhibited by Ca²⁺ inactivation of closed receptor/channels due to Ca²⁺ accumulation local to the release sites. Hepatocytes have a much smaller, more prolonged InsP₃-evoked Ca²⁺ flux than Purkinje neurons. Evidence suggests that these differences in kinetics can be explained by the much lower InsP₃ receptor density in hepatocytes than Purkinje neurons, rather than differences in receptor isoform, and, more generally, that high InsP₃ receptor density promotes fast rising, rapidly inactivating InsP₃-evoked [Ca²⁺]_i transients.

introduction

Calcium release from intracellular stores by InsP₃ mediates the activation of liver metabolism by several hormones, resulting, for example, in glycogenolysis and bile secretion. In guinea pig and rabbit liver, the rise of [Ca²⁺]_i produces a large increase in K⁺ and Cl⁻ permeability through Ca²⁺-activated conductances in the plasma membrane. The increase of [Ca²⁺]_i typically occurs as a sequence of periodic spikes of high concentration lasting several seconds (Woods et al., 1987; Field and Jenkinson, 1987) with frequency dependent on the degree of stimulation. The mechanisms generating this pattern of Ca²⁺ release appear to depend to some degree on local regulation by InsP₃ receptor activation itself, probably via local feedback through Ca²⁺ released into the cytosol. Evidence supporting this idea is the generation of periodic spiking of the Ca²⁺ concentration in single hepatocytes by InsP₃ alone perfused directly into the cytosol from a patch pipette (Capiod et al., 1987; Ogden et al., 1990) or by stable 5-thio-InsP₃ released by photolysis (Wootton et al., 1995), although there are additional factors secondary to receptor activation that modify this process during hormone action (Cobbold et al., 1991).

The regulation of the InsP₃ receptor by cytosolic Ca²⁺ in several tissues has been shown to be facilitatory at low concentrations close to resting levels (Iino, 1990; Bezprozvanny et al., 1991; Finch et al., 1991; Iino and Endo, 1992) and inhibitory at high concentrations of 1 μM or greater (Berridge, 1988; Payne et al., 1988; Iino, 1990; Parker and Ivorra, 1990; Ogden et al., 1990; Bezprozvanny et al., 1991; Finch et al., 1991; Iino and Endo, 1992; Combettes et al., 1993; Khodakhah and Ogden, 1995). The facilitatory action is less pronounced or apparently absent in some studies (see Combettes et al., 1994, Khodakhah and Ogden, 1995), but the inhibitory action of high [Ca²⁺]_i is clearly demonstrated in all studies cited. In hepatocytes, there is one report of facilitation by cytosolic free [Ca²⁺] (Marshall and Taylor, 1993) and several studies showing inhibition of Ca²⁺ release (Ogden et al., 1990; Combettes et al.,1993) and Ca²⁺-induced desensitization of InsP₃ receptors (Pietri et al., 1990). Additionally, inactivation processes that do not depend on [Ca²⁺]_i have been described in Xenopus oocytes (Ilyin and Parker, 1994) and in rat hepatocytes (Hajnozcky and Thomas, 1994).

To study the mechanisms of activation and local regulation of InsP₃ receptors in single hepatocytes, it is necessary to gain access to the receptors in the cytosol and release InsP₃ in a controlled fashion and to measure the kinetics of the resulting Ca²⁺ release. Generally, kinetic studies of InsP₃-activated Ca²⁺ fluxes have been made with fast perfusion of cell suspensions after permeabilizing the surface membrane, but endogenous regulatory processes and differences between cells would not be detected with these methods. In a previous single cell study, InsP₃ was released by flash photolysis of caged InsP₃ in the cytosol of single hepatocytes, and the activation by Ca²⁺ of the endogenous apamin-sensitive Ca²⁺-activated K⁺ conductance was used as a monitor of free [Ca²⁺] (Ogden et al., 1990). This showed that InsP₃ produced activation at concentrations of 0.4 μM or higher, that there is delay of up to several hundred milliseconds at low InsP₃ concentrations, and that there is a subsequent prolonged inactivation of the Ca²⁺ release process, requiring ∼1 min to recover, produced by the elevation of cytosolic free [Ca²⁺]. However, the Ca²⁺-activated K⁺ conductance of guinea pig hepatocytes has a strongly cooperative dependence on [Ca²⁺] (Capiod and Ogden, 1989a ) and is unsuitable for quantitative kinetic investigation of Ca²⁺ flux, although it provides a useful indicator of free [Ca²⁺] changes. To overcome this limitation, the experiments described here make use of fluorescent Ca²⁺ indicators to measure particularly the kinetics of activation and inactivation of Ca²⁺ release into the cytosol soon after InsP₃ is liberated by photolysis in single guinea pig hepatocytes. The delays are shorter than those found with the K⁺ conductance, the Ca²⁺ efflux is quantified as d[Ca²⁺]_i/dt and found to be small relative to other tissues, particularly in comparison with cerebellar Purkinje neurons, and the rate and mechanism of termination of Ca²⁺ efflux from stores is investigated and shown to depend on [Ca²⁺]_i.

Evidence from cell to cell and tissue to tissue variation in the kinetics of Ca²⁺ release supports the idea that InsP₃ receptor density is an important factor determining the duration as well as the magnitude of Ca²⁺ flux. Large Ca²⁺ flux which results from high channel density produces a rapid termination of release by inactivating closed InsP₃ channels via local high free [Ca²⁺]. Hepatocytes have a low density of InsP₃ receptor/channels, producing small Ca²⁺ flux and slow termination of release. Comparison is made with similar data from cerebellar Purkinje neurons, where the receptor density is very high and which have high flux (Khodakhah and Ogden, 1995) and rapid rates of termination. The results suggest that InsP₃ receptor density can account for most of the differences in InsP₃ action between hepatocytes and Purkinje neurons and supports the idea that the density of open InsP₃ receptor/channels determines the time course of the rise and fall of free [Ca²⁺] during hormonal signalling in hepatocytes.

methods

Cell Preparation and Solutions

Guinea-pig hepatocytes were isolated by perfusion with collagenase followed by mechanical dispersion (Capiod and Ogden, 1989a , b ). Cells were plated from suspension onto 40-mm collagen-coated quartz coverslips in Williams medium E and kept at 37°C in a hydrated atmosphere of CO₂/5% in air. Recordings were made after 2–6 h in Cl⁻-free solution with gluconate as the major anion. External solution contained (in mM): Na gluconate, 145; K gluconate, 5.6; CaSO₄, 5; MgSO₄, 1.2; HEPES, 8. For experiments in absence of external Ca²⁺, solutions contained (in mM): Na gluconate, 149; K gluconate, 5.6; MgSO₄, 3; EGTA, 0.04; HEPES, 8. Internal Cl⁻-free solutions contained (in mM): K gluconate 153; ATPNa₂, 3; MgSO₄, 3; HEPES, 8; 200 μM fluo-3 or 500 μM furaptra and caged InsP₃ or caged 5-thio-InsP₃. Fura-2–buf-fered internal solutions contained (in mM): K gluconate, 40; ATPNa₂, 3; MgSO₄, 3; fura-2–K₅, 10; CaCl₂, 4 or 7.5; HEPES, 100; KOH, 40. All the solutions were buffered at pH 7.3 and 0.2-μm filtered. Experiments were made at room temperature (∼27°C).

Somata of rat cerebellar Purkinje neurons were isolated from slices of cerebellar cortex from 12–16-d-old Wistar rats by mild proteolysis (Sigma Type 23, 3 mg · ml⁻¹ for 7 min at 37°C; Sigma Chemical Co., St. Louis, MO) and gentle dissociation (see Khodakhah and Ogden, 1995).

Whole-cell Recordings

Standard tight-seal whole-cell recording techniques were used (Hamill et al., 1981). Patch pipettes were made from pyrex glass and coated with a wax comprising parafilm (40%) and mineral oil (60%).

Fluorescence Measurements and Flash Photolysis

The fluorescent indicator used to detect changes in Ca²⁺ concentration was chosen according to the type of measurement made. Fluo-3 has a dissociation constant of K = 0.7 μM in situ, close to the resting [Ca²⁺]_i, has a large fluorescence increase on binding Ca²⁺, and was used to measure delays in the activation of Ca²⁺ release.

Estimates of the time course of [Ca²⁺]_i changes from changes in fluo-3 fluorescence were found to be distorted when compared with data obtained with lower affinity indicators (see Ogden et al., 1995). This results presumably from saturation of the fluo-3 at high local [Ca²⁺] near release sites, averaging fluorescence changes across the cell producing underestimates of the [Ca²⁺] changes at early times after InsP₃ release. Therefore experiments to measure the time course of the rise and decline of [Ca²⁺]_i were made with the low affinity indicator furaptra, (K = 44 μM, Konishi et al., 1991; K = 48 μM, Ogden et al., 1995; K estimated as 100 μM in frog muscle myoplasm, Zhao et al., 1996).

Finally, in some experiments it was desired to buffer cytosolic free [Ca²⁺] with high concentrations of fura-2/calcium mixtures. To obtain precise values of free [Ca²⁺], the concentration of fura-2 was determined from the absorbance of the solution and extinction coefficient e₃₆₂ = 28,700 M⁻¹ cm⁻¹ (Baylor and Hollingworth, 1988), and Ca²⁺ was added to give the desired free [Ca²⁺] assuming K = 0.2 μM (see, e.g., Zhou and Neher, 1993).

Fluo-3 (200 μM), furaptra (500 μM), or fura-2 (10 mM) free acids were introduced into the cell together with caged compounds via the patch pipette, and diffusional equilibration between pipette solution and cell was determined from the fluorescence record; recordings were made at least 5 min after membrane rupture. Microspectrofluorimetry was on a Nikon TMD microscope with 40× 1.3 NA objective. Excitation light (fluo-3, 450–490 nm; furaptra or fura-2, 400–440 nm) was from a xenon arc lamp, and light emitted from a single cell was viewed with long pass filters at 520 nm (fluo-3) or 470 nm (furaptra, fura-2). The area viewed was restricted to a single cell by an adjustable rectangular diaphragm, and emitted light was detected by a photomultiplier operated in photon counting mode. Pulse output was counted by a Tecmar interface, corrected for missed pulses, and stored via software (Photon Technology International, South Brunswick, NJ). Photon counts were converted in parallel to an analogue signal by an integrating amplifier with correction for missed counts (Cairn Research) and stored on FM tape.

Fluorescence changes in fluo-3 were expressed as a ratio with respect to the initial resting fluorescence. Furaptra was used at a single excitation wavelength of 420 nm (in most experiments 400–440 nm) conditions in which the fluorescence is quenched close to zero on binding Ca²⁺. The method for and background to the calculation of free Ca²⁺ from the fluorescence has been described previously (Ogden et al., 1995; see Konishi et al., 1991). Briefly, the intrinsic fluorescence recorded before whole cell recording was taken as the fluorescence at saturating [Ca²⁺]_i, F _Camax, where furaptra fluorescence is quenched. The resting [Ca²⁺]_i in guinea pig hepatocytes has been estimated as 0.2 μM (Burgess et al., 1984) and therefore the fluorescence of furaptra under resting conditions can be taken as an estimate of F _Camin. The free Ca²⁺, Ca _f,was calculated from the fluorescence F by means of the relation:

1

Fura-2 was used at high free concentration, 2.5 or 5 mM, to buffer Ca _f to levels near resting [Ca²⁺], i.e., conditions in which it is the predominant source of intracellular Ca²⁺ binding. Fura-2 was used at a single excitation wavelength of 420 nm, and emission was measured at >470 nm, conditions in which the fluorescence is quenched by high [Ca²⁺] to zero. As for furaptra, the intrinsic fluorescence recorded before fluophore loading was taken as the fluorescence at saturating free [Ca²⁺], F _Camax. The dissociation constant K, the free and bound indicator concentrations, D _f, D _Ca, and the Ca _f are related by:

Total indicator and Ca²⁺ concentrations are:

which on substitution gives:

The dissociation constant for fura-2 binding Ca²⁺ was assumed to be K = 0.2 μM (see Zhou and Neher, 1993), and D _T and Ca _T have values such that Ca _f is << Ca _T, so

2

Ca²⁺ released into the cytosol binds predominantly to fura-2 and is measured as the change of D _Ca. To measure changes of D _Ca an estimate of F _Camin is needed. If Ca _T and D _T added to the cytosol have fluorescence F = F ₀ , and Ca _f << Ca _T as above, rearranging Eqs. 1 and 2 gives:

permits calculation of F _Camin from the initial fluorescence, F ₀, the composition of the fura-2/Ca mixture, and F _Camax, which in these measurements was close to zero. Thus at high fura-2 concentration, Ca²⁺ released onto fura-2 appears as an increase of D _Ca and was calculated from the fura-2 fluorescence, F, as the proportion of F _Camin quenched, by:

The Ca²⁺-dependent K⁺ conductance has a range of activation by cytosolic Ca²⁺ ion concentration of 0.3–1.5 μM and a maximum open probability of 0.9 (Capiod and Ogden, 1989a ). It is not detectably activated at [Ca²⁺]_i less than 0.3 μM and provides a good index of low resting [Ca²⁺]_i. Cells with any degree of activation of Ca²⁺-dependent K⁺ conductance before InsP₃ stimulation were not used. The Ca²⁺-dependent K⁺ conductance was also used as an index of changes of [Ca²⁺]_i during experiments.

Photolysis of caged InsP₃, the P-4 or the P-5 1-(2-nitrophenyl)ethyl esters of InsP₃ (Walker et al., 1989), was produced by a 1-ms pulse from a short arc xenon flashlamp (Rapp and Guth, 1988) focused to produce an image 2–3 mm across at the cell as described previously (Ogden et al., 1990). The output of the lamp was adapted for most experiments to be adjustable in the range producing 6–14% photolysis of caged InsP₃. The coefficient of variation of photolysis of caged ATP was 10%, indicating the degree of reproducibility of photolysis of InsP₃ in each cell. In some experiments a caged stable analogue of InsP₃, caged 5-thio-InsP₃, the S-1 (2-nitrophenyl)ethyl ester of 1-d-myo-inositol 1,4 bisphosphate 5-phosphothiorate. This has a smaller quantum yield than caged InsP₃, 0.57 compared to 0.65, and a slower conversion, 87 against 220 s⁻¹ (Wootton et al., 1995).

The optical artefact, mainly phosphorescence in the objective, arising from the UV pulse was minimized (4–8 ms) by use of quartz coverslips and UV block (suntan) oil between the cell and objective (Carter and Ogden, 1992; Ogden et al., 1993). This artefact constituted the main limitation to resolving the latency of responses at high InsP₃ concentration, the fluorescence changes of fluo-3 and furaptra following photolysis of DM-Nitrophen being complete in less than 4 ms. Photolysis of caged InsP₃ by fluorescence excitation was minimized by shuttering the xenon lamp when not recording.

Data were recorded on FM tape at 200 Hz (fluorescence) or 1 kHz (membrane current) bandwidth, and in software fluorescence was integrated over 10 or 20 ms.

Materials

Chemicals were Analar grade from BDH Chemicals Ltd. (Poole, UK) or Sigma Chemical Co. Collagenase CLS 2 was from Worthington Biochemical Corp., (Freehold, NJ). Fluo-3, furaptra and fura-2 were from Molecular Probes (Eugene, OR). The concentration of fura-2 stock solutions was determined spectroscopically with extinction coefficient at 362 nm of 2.8 × 10⁴ M⁻¹ cm⁻¹ (Baylor and Hollingworth, 1988).

results

Rate of Change of [Ca²⁺]_i as a Measure of InsP₃-evoked Ca²⁺ Flux

Activation of InsP₃ receptors results in flux of Ca²⁺ from stores into the cytosol which appears as an increase of free [Ca²⁺]. The rate of change of free [Ca²⁺], d[Ca²⁺]_i/dt (units moles · s⁻¹ · liters⁻¹), is proportional to the net flux of Ca²⁺ into unit volume of cytosol (Ogden et al., 1990), the proportionality depending on the fraction of Ca²⁺ bound to endogenous buffers and to the Ca²⁺ indicator itself. Measurement of d[Ca²⁺]_i/dt with spatially averaged fluorescence of Ca²⁺ indicators is susceptible to distortion arising from local Ca²⁺ accumulation and the saturating relation between fluorescence and [Ca²⁺]_i. Precise measurement depends on K _Ca of the indicator being several fold higher than the peak local free Ca²⁺ so that spatial summation occurs on the linear part of the fluorescence-[Ca²⁺] relation. Furaptra used here has K _Ca = 48 μM, and at 500 μM binds approximately 10 Ca²⁺ ions for each free Ca²⁺ ion, producing minimal exogenous Ca²⁺ buffering. Peak d[Ca²⁺]_i/dt of each response was estimated from the slope of a straight line fitted to the rising phase.

The characteristic time course of the [Ca²⁺]_i and Ca²⁺-dependent K⁺ conductance of guinea pig hepatocytes after flash photolysis of caged InsP₃ in the cytosol is illustrated by the records shown in Fig 1. After the flash the [Ca²⁺]_i increases after a delay, rises quickly due to net Ca²⁺ flux into the cytosol, then declines slowly due to net loss of Ca²⁺ from the cytosol. The rate of rise of [Ca²⁺]_i due to InsP₃ was up to 52 μM · s⁻¹ and the rate of decline following the peak was −0.2 to −1.3 μM s⁻¹ (range, n = 9). In Purkinje neurons peak flux was much higher (<1,400 μM · s⁻¹), and rates of decline were similar to hepatocytes (Khodakhah and Ogden, 1995). At the peak [Ca²⁺]_i, where d[Ca²⁺]_i/dt = 0, the net flux is zero, and the InsP₃-evoked efflux at this point is similar in magnitude to the flux of Ca²⁺ from the cytosol during the decline, small when compared to the maximum d[Ca²⁺]_i/dt. The flux, d[Ca²⁺]_i/dt, was estimated as the slope of a straight line fitted to the maximum rate of rise. The analysis presented below is of (a) the delay and its dependence on InsP₃ concentration, (b) the rate of rise of [Ca²⁺]_i as a measure of activation of InsP₃ gated Ca²⁺ efflux, (c) the mechanism that terminates InsP₃-gated flux at the peak of the response, and therefore determines the period of high efflux, and (d) the effect of InsP₃ concentration on the duration of Ca²⁺ release.

Figure 1.

InsP₃-evoked increase of [Ca²⁺]_i and Ca²⁺-activated K⁺ current. Single hepatocyte whole cell voltage clamp at 0 mV with pipette solution containing 500 μM furaptra and 10 μM caged InsP₃. 1.2 μM InsP₃ released by 1-ms UV pulse at time indicated by the arrow. Ca²⁺-activated K⁺ current at 0 mV (lower trace) and [Ca²⁺]_i calculated from fluorescence (upper trace) after release of InsP₃. Rate of decline in [Ca²⁺]_i illustrated by the dashed line (left panel). (Right panel) Expanded records illustrating the delays (D1, fluorescence; D2, Ca²⁺-activated K⁺ current), rate of rise (d[Ca²⁺]/dt) and duration (10–90%) of the increase in [Ca²⁺]_i.

The lower record in Fig. 1 shows the K⁺ current recorded at 0 mV membrane potential and measures the activation of Ca²⁺-dependent K⁺ conductance of the plasma membrane by cytosolic Ca²⁺. This conductance has been characterized (activation range 0.3–1.5 μM free [Ca²⁺], Capiod and Ogden, 1989a ) and serves to show that the resting [Ca²⁺]_i is low, and the Ca²⁺ indicator does not affect the response. Because of the steep dependence on free Ca²⁺ (Capiod and Ogden, 1989a ) and saturation at relatively low free [Ca²⁺], the Ca²⁺- dependent K⁺ conductance cannot be used to measure absolute values of d[Ca²⁺]_i/dt for kinetic studies.

Initial Delay of InsP₃- evoked Ca²⁺ Release

Previous work (Ogden et al., 1990) had shown a delay of up to 1 s between releasing low concentrations, <0.5 μM InsP₃, and activation of the plasmalemmal Ca²⁺- dependent K⁺ conductance, and that this delay was reduced to a minimum of ∼120 ms at high (>2 μM) InsP₃ concentration. To measure the delays in [Ca²⁺]_i changes, increases in cytosolic [Ca²⁺] close to the resting level were detected with the high affinity Ca²⁺ indicator fluo-3 to reduce noise near resting [Ca²⁺]_i (K _Ca ∼ 0.4 μM). The delay was estimated as the time from the flash until the fluorescence deviates from the baseline (Fig. 2 A). Delays in the rise of fluo-3 fluorescence declined from a mean of 290 ± 90 ms (mean ± SEM, n = 6) at 0.4 μM InsP₃ to <30 ms at >5 μM InsP₃ (data summarized in Fig. 2 B). The delays in the rise of the Ca²⁺-dependent K⁺ conductance were longer than those of the Ca²⁺ increase detected by fluo-3, from 1,000 ± 170 ms at 0.4 μM InsP₃ declining to 104 ± 16 ms at 10 μM InsP₃ (Ogden et al., 1990). The difference can be explained by the need to raise [Ca²⁺]_i to 0.3 μM to activate the Ca²⁺-dependent K⁺ conductance, higher than resting [Ca²⁺]_i, and the steep activation of the Ca²⁺-dependent K⁺ conductance by free Ca²⁺. The minimum delay in fluo-3 fluorescence at high InsP₃ concentration was often obscured by an optical artifact lasting up to 20 ms due to flashlamp discharge. Even at very high InsP₃ concentrations of 25–100 μM, the delays were >20 ms.

Figure 2.

Initial delays in time course of InsP₃-evoked [Ca²⁺]_i and Ca²⁺-activated K⁺ current increases. (A) Whole cell voltage clamp with 200 μM fluo-3 at 0 mV. The Ca²⁺-induced fluorescence change (ΔF) is expressed as a fraction of fluorescence (F) in the unstimulated cell. (Left) Whole cell Ca²⁺-activated K⁺ current at 0 mV (upper traces) and fluo-3 fluorescence (lower traces) after release of 0.4, 0.8, and 8 μM InsP₃ at the times indicated by the arrows in three different hepatocytes. (Right) Expanded records illustrating the delay and initial rate of rise. Horizontal bar, 2 s (left), 500 ms (right); left vertical bar, I = 200 pA; right vertical bar, ΔF/F = 0.5. (B) Delays between InsP₃ release and initiation of Ca²⁺ flux plotted against InsP₃ concentration (log scale).

Two explanations for the delay in Ca²⁺ release may be cooperativity in the binding of InsP₃ to produce activation, data in RBL cells suggesting n = 4 (Meyer et al., 1988), or a cooperative effect of Ca²⁺ ions at concentrations around the resting level, shown by Iino (1990), Bezprozvanny et al. (1991), and Finch et al. (1991). In the present experiments, when cells were loaded with 500 μM furaptra, the delay to a pulse of InsP₃ was substantially reduced when [Ca²⁺]_i was elevated by Ca²⁺ influx, as illustrated by the records in Fig. 3, which show Ca²⁺ release by 1.2 μM InsP₃ in the same hepatocyte in the presence (middle trace) and absence (upper and lower pre and post controls) of elevated [Ca²⁺]_i evoked by hyperpolarization. In five experiments of this kind, no delay was discernible with prior elevation of [Ca²⁺]_i, whereas control release of the same InsP₃ concentration in the same cells had delays of mean 380 ± 62 ms (mean ± SEM, n = 5). These results agree with the idea that the initial efflux of Ca²⁺ ions produces a localized increase of [Ca²⁺]_i and accelerates activation of the InsP₃ receptor. It should be noted for later reference that although the delays were reduced, the subsequent Ca²⁺ flux induced by InsP₃, measured by d[Ca²⁺]_i/dt, was also reduced in the same cells by the rise in [Ca²⁺]_i (discussed below).

Figure 3.

Effect of raised [Ca²⁺]_i on InsP₃-evoked Ca²⁺ flux and delay. The cell was loaded with 500 μM furaptra, and 1.2 μM InsP₃ was released as indicated by the arrows. The top and bottom traces show the control responses at 0 mV with rates of rises in [Ca²⁺]_i (d[Ca²⁺]_i/dt) of, respectively, 5.6 and 6.2 μM · s⁻¹. An elevation in [Ca²⁺]_i of ∼2 μM obtained by hyperpolarizing the cell to −120 mV (middle trace) resulted in a slower rate of Ca²⁺ release, d[Ca²⁺]/dt of 1.7 μM · s⁻¹, after the release of InsP₃. Flash release of InsP₃ evoked an increase in [Ca²⁺]_i without detectable delay when [Ca²⁺]_i was elevated (middle trace) and with a delay of ∼500 ms in control conditions at 0 mV in the same cell (upper trace). Each traces were separated by a interval of 2 min.

Mean delays at similar InsP₃ concentrations were not significantly shorter with fluo-3 (200 ± 40 ms at 0.8 μM InsP₃, range 80–350 ms, n = 7) than with furaptra (310 ± 40 ms at 1.2 μM InsP₃, range 90–560 ms, n = 11). However, the low affinity Ca²⁺ dye furaptra is less likely to detect small changes in [Ca²⁺]_i than fluo-3, making it more difficult to estimate the deviation of the trace from the baseline.

The cooperativity and narrow range of [Ca²⁺]_i producing Ca²⁺ activation of the K⁺ conductance in guinea pig liver cells (Capiod and Ogden, 1989a ) is most likely responsible for the additional delay, saturation at submaximal InsP₃ concentrations (>0.6 μM), and steep rise of the conductance increase compared to the rise of [Ca²⁺]_i (see Fig. 1).

Amplitude and Rate of Rise of [Ca²⁺]_i

Experiments with fluo-3 as the Ca²⁺ indicator showed evidence of distortion of the kinetics and amplitude of the [Ca²⁺]_i increase averaged over the cell as measured by the fluorescence change (Ogden et al., 1995). The problem was overcome by use of the lower affinity indicator furaptra to extend the range of linear summation of [Ca²⁺]_i (Raju et al., 1989, Konishi et al., 1991) (see methods). The effect of increasing the InsP₃ concentration released by a flash in cells loaded with 500 μM furaptra was to increase the rate of rise and peak amplitude of [Ca²⁺]_i in the range of InsP₃ concentrations 0.2 μM to ∼5 μM. The maximum rate of rise, d[Ca²⁺]_i/dt, provides a measure of the InsP₃-evoked Ca²⁺ flux into unit cytosolic volume and can be used to compare different InsP₃ concentrations in the same cell, and the variation from cell to cell at the same InsP₃ concentration. The effect of increasing concentration within single hepatocytes is shown by the records of [Ca²⁺]_i in Fig. 4. Fig. 4 A shows an increase of peak and rate of rise of [Ca²⁺]_i between pulses of 0.7 and 1.4 μM InsP₃ in the same cell. The d[Ca²⁺]_i/dt (slopes of the fitted lines ± SD) were significantly different (P < 0.01). Fig. 4 B shows [Ca²⁺]_i in response to 7 and 13 μM in another cell with no increase in either parameter at the higher concentration, showing saturation of the response. The pooled data show saturation of the InsP₃-evoked flux, measured as the maximum rate of rise in [Ca²⁺]_i at 5–10 μM InsP₃. The data normalized to 1.2 μM InsP₃ in each of 10 hepatocytes is shown for peak [Ca²⁺]_i in Fig. 5 A and for d[Ca²⁺]_i/dt in Fig. 5 B. Both parameters increased in each cell with InsP₃ concentrations from 0.6 to 2.4 μM, showing that the InsP₃-evoked flux increased in this range of concentrations. An estimate of the Hill coefficient for InsP₃-evoked flux measured by d[Ca²⁺]_i/dt in this range was 1.5, but was poorly defined.

Figure 4.

Concentration dependence of the InsP₃-evoked [Ca²⁺]_i increase. Increase in [Ca²⁺]_i evoked by InsP₃ in two different hepatocytes loaded with 500 μM furaptra at 0 mV. The photolytic releases indicated by the arrows of 0.7 and 1.4 μM InsP₃ in the first hepatocyte are shown in the top four traces (A) and of 7 and 13 μM in the second in the four bottom traces (B). Expanded records illustrating the delays, rates of rise (d[Ca²⁺]_i /dt) and durations (10–90%) of the increase in [Ca²⁺]_i are shown on the right sides of the figure.

Figure 5.

Dependence of Ca²⁺ flux and peak [Ca²⁺]_i on InsP₃ concentration. (A) Peak [Ca²⁺]_i (Δ[Ca]) and (B) rates of rise of free [Ca²⁺]_i (d[Ca]_i/dt) normalized to value at 1.2 μM InsP₃ in each cell and plotted against InsP₃ concentration (log scale). Data from 10 hepatocytes whole cell voltage clamped with 500 μM furaptra at 0 mV. (C) Peak [Ca²⁺]_i and (D) d[Ca²⁺]_i /dt plotted against InsP₃ concentration released by photolysis of caged InsP₃ in 48 guinea-pig hepatocytes. 500 μM furaptra, 0 mV.

Pooled data from 48 hepatocytes showed considerable variability in the amplitude and rate of rise of the response from cell to cell. This is shown by the data plots in Fig. 5, C and D, which show a large variation in both parameters at each InsP₃. The differences in d[Ca²⁺]_i/dt at each InsP₃ concentration from cell to cell may be due to real differences in Ca²⁺ flux because of differing InsP₃ receptor densities, from differing single InsP₃ receptor-channel properties, or the Ca²⁺ flux may be similar but with differences in Ca²⁺ buffering, resulting in lower d[Ca²⁺]_i/dt when buffering is high. The variations in free d[Ca²⁺]_i/dt from cell to cell are analyzed further below.

Termination of InsP₃-evoked Ca²⁺ Efflux

The time course of the InsP₃-evoked rise of [Ca²⁺]_i shown in Figs. 1 and 3 has a well-defined peak where the net flux of Ca²⁺ ions into the cytosol, measured as d[Ca²⁺]_i/dt, is zero. The mechanism terminating Ca²⁺ efflux was investigated previously with twin pulse protocols in hepatocytes (Ogden et al., 1990) and showed that Ca²⁺ release evoked by a second pulse of InsP₃ is inhibited for a period of ∼1 min after a response to the initial pulse, an effect that recovers with a half-time of 10–20 s for the pulse interval. The Ca²⁺-activated K⁺ conductance was used to monitor Ca²⁺ flux by InsP₃ in the earlier study but could not show the time course of the onset of this effect because of the distortion produced by cooperative activation of the K⁺ conductance by Ca²⁺. The onset close to peak [Ca²⁺]_i was tested here with twin pulse experiments monitoring [Ca²⁺]_i with furaptra, and results from three cells are shown in Fig. 6 to illustrate the onset of inhibition. (1) On the left-hand side (Fig. 6 A), the upper record shows Ca²⁺ release evoked by 0.4 μM InsP₃, the peak flux d[Ca²⁺]_i/dt = 25 μM · s⁻¹ is indicated by the line fitted to the fast rise. The middle trace shows that a second pulse of 0.4 μM InsP₃ 500 ms later, before [Ca²⁺]_i had peaked, produced no further activation of Ca²⁺ flux, i.e., no further increase of d[Ca²⁺]_i/dt above 24 μM · s⁻¹ produced by the first pulse. The effect was not due to receptor saturation as 1.4 μM InsP₃ subsequently produced a larger flux (d[Ca²⁺]_i/dt = 46 μM · s⁻¹) in the same cell shown in the lower trace. (2) This effect is shown in another cell where a second photolytic release of 0.6 μM InsP₃ at about 50% of the rising phase of [Ca²⁺]_i showed no change in the slope (Fig. 6 B, upper trace), and a single release of 0.6 μM InsP₃ in the same cell evoked a response of similar amplitude and rate of rise of [Ca²⁺]_i as that established by the first flash with submaximal InsP₃ concentration (Fig. 6 B, bottom trace). It may be noted that InsP₃ sensitivity was suppressed at the time of the second pulse in both Fig. 6, A and B, although receptor/channels were still open as judged by the high steady flux occurring at that time. This suggests that closed receptor/channels were inhibited at the time of the second pulse (see discussion). (3) Finally, in the third example illustrated, twin pulses of 0.6 μM in a cell with low flux showed a significant increase of d[Ca²⁺]_i/dt from 1 μM · s⁻¹ to 3.4 μM · s⁻¹ (slopes of the fitted lines ± SD, P < 0.01) after the second flash with a 1.6-s interval (Fig. 6 C, upper trace). The delay after the first pulse of InsP₃ was determined with the Ca²⁺-dependent K⁺ conductance (not shown). The rate of rise in [Ca²⁺]_i after the second flash can be compared to that evoked by a single release of 1.2 μM InsP₃ (Fig. 6 C, bottom trace), which produced a flux of 3.7 μM · s⁻¹ later in the same cell showing that in this case the effect of two pulses of 0.6 μM were similar to a single pulse of 1.2 μM. The data of Fig. 6, A and B, show that the onset of the inhibition can occur early in the response even though InsP₃ receptor/channels are still activated; a difference in the cell of Fig. 6 C, where no clear inhibition was seen, was the smaller magnitude of the Ca²⁺ flux.

Figure 6.

Time course of onset of loss of InsP₃ sensitivity. [Ca²⁺]_i after photolytic release of InsP₃ at times indicated by the arrows in three different guinea pig hepatocytes loaded with 500 μM furaptra at 0 mV. Traces separated by ∼2 min. Rates of rise of [Ca²⁺]_i (d[Ca]_i/dt) are indicated by lines fitted to data. (A) Top trace shows the control to 0.4 μM InsP₃, middle trace two successive releases of 0.4 μM InsP₃ with a interval of 500 ms showing no change in the rate in [Ca²⁺] release (d[Ca]_i/dt) after the second flash, and the bottom trace the response to photolytic release of 1.4 μM InsP₃. (B) Two successive pulses of 0.6 μM InsP₃ (top trace) show no change in the rate of rise of [Ca²⁺]_i after the second flash (600-ms interval) when compared to a single pulse of 0.6 μM InsP₃ later in the same cell (bottom trace). (C) Two successive pulses of 0.6 μM InsP₃ (top trace) show an increase in the rate of rise of [Ca²⁺]_i after the second flash (1.6-s interval). Bottom trace shows a single release of 1.2 μM InsP₃ in the same cell.

These results and experiments reported earlier with twin pulses of InsP₃ (Ogden et al., 1990) and stable 5-thio-InsP₃ (Wootton et al., 1995) rule out the possibility that the Ca²⁺ release is terminated by InsP₃ breakdown. There is evidence in guinea pig (Ogden et al., 1990) and rat hepatocytes (Combettes et al., 1993) that elevation of [Ca²⁺]_i inhibits InsP₃-evoked Ca²⁺ release. It is well documented that high [Ca²⁺]_i inhibits the InsP₃ receptor of several other cell types (Payne et al., 1988; Iino 1990; Parker and Ivorra, 1990; Bezprozvanny et al., 1991; Finch et al., 1991). Ca²⁺ appears to act on both closed and open InsP₃ channels (Bezprozvanny et al., 1991; Finch et al., 1991; Bezprozvanny and Ehrlich, 1994). One possible mechanism for the abrupt termination is inhibition of InsP₃ channels by the locally elevated [Ca²⁺] released by InsP₃. This hypothesis was tested in two ways, first by clamping the [Ca²⁺]_i with a high concentration of fura-2 and recording the Ca²⁺ flux as the increase in Ca²⁺-bound indicator concentration, second by producing an elevation of [Ca²⁺]_i independently of InsP₃ and testing inhibition of Ca²⁺ release measured by free d[Ca²⁺]_i/dt.

High Ca²⁺ Buffering with Fura-2

An experiment in which [Ca²⁺]_i was controlled is shown in Fig. 7. The stable 5-thio analogue of InsP₃ was released photolytically from caged 5-thio-InsP₃ to avoid complications arising from metabolism of InsP₃. 5-thio-InsP₃ is about 5 times less potent than InsP₃ in guinea pig hepatocytes (Wootton et al., 1995). The cell was perfused from the pipette with 36 μM caged 5-thio InsP₃, 10 mM fura-2, and 7.5 mM total Ca²⁺. There was a small activation of the Ca²⁺-dependent K⁺ conductance suggesting that the [Ca²⁺]_i was elevated above the normal resting level as expected for this mixture. When the fura-2 fluorescence was steady, release of 5-thio InsP₃ produced Ca²⁺ flux into the cytosol, measured as the increase of Ca²⁺ bound to fura-2, at a rate similar to the generation of free Ca²⁺ seen with furaptra as the Ca²⁺ indicator, but with much prolonged duration. Data from two cells are shown in Fig. 7. The upper trace shows release evoked by 2 pulses of 4.5 μM 5-thio-InsP₃ separated by 2 s. Ca²⁺ flux in response to the first pulse is 60 μM · s⁻¹, is increased by release of a further 4.5 μM 5-thio-InsP₃ to 90 μM · s⁻¹ 2 s later, and the total duration of the response is 4 s. The lower record shows in another cell the Ca²⁺ flux produced by a single release of 9 μM 5-thio-InsP₃, with flux also of ∼100 μM · s⁻¹ and duration 2 s. Once the flux had declined to zero at the peak, further release of 5-thio-InsP₃ evoked no further flux. In these experiments no additional activation of the Ca²⁺-dependent K⁺ conductance was seen, showing that [Ca²⁺]_i was buffered. These data show first that the Ca²⁺ flux is much prolonged when [Ca²⁺]_i is prevented from rising by high fura-2 concentrations, that the inhibition of a second pulse of InsP₃, which normally occurs within 0.5 s, is not operating even at 2 s when [Ca²⁺]_i is buffered to near resting levels, and, finally, that further release does not occur at the peak of the response even though [Ca²⁺]_i has not changed and is near resting levels, suggesting that another inhibitory mechanism, possibly simply store depletion, has occurred.

Figure 7.

Total Ca²⁺ increases evoked by photolytic release of 5-thio-InsP₃ in [Ca²⁺]_i clamped cells. The two traces show the effect of photolytic release of 4.5 and 9 μM 5-thio-InsP₃ in two different guinea pig hepatocytes loaded with 10 mM fura-2 and 7.5 mM Ca²⁺. The top trace shows the response to successive release of 4.5 μM 5-thio-InsP₃ with an interval of 2 s. An increase in the rate of total Ca²⁺ release from 60 to 90 μM · s⁻¹ was observed after the second release. A third release of 4.5 μM 5-thio-InsP₃ at the plateau had no effect (not shown). The release of 9 μM 5-thio-InsP₃ raised total Ca²⁺ by 200 μM at a rate of 100 μM · s⁻¹ (bottom trace). A second release of 9 μM 5-thio-InsP₃ at the plateau had no effect. Holding potential 0 mV.

Prior Elevation of [Ca²⁺]_i

Cytosolic free [Ca²⁺] was raised by polarizing the membrane to −120 mV to increase Ca²⁺ leak from the external solution. As discussed above in connection with Fig. 3, the effects on InsP₃-evoked release were to greatly reduce the latency before Ca²⁺ efflux, and also to decrease the magnitude of the InsP₃-evoked efflux. In Fig. 3 the upper record shows the Ca²⁺ efflux of 5.6 μM · s⁻¹ evoked by 1.2 μM InsP₃ at 0 mV, the middle trace an InsP₃-evoked efflux of 1.7 μM · s⁻¹ 2 min later in the same cell, after [Ca²⁺]_i had been elevated to ∼2 μM by Ca²⁺ from the external solution at V = −120 mV, and the lower record shows a further control producing a flux of 6.2 μM · s⁻¹ at 0 mV 2 min later. The main effect of prior elevation of [Ca²⁺]_i to ∼2 μM was to suppress the InsP₃-evoked efflux to 33% (±6% SD, n = 6) of the control flux in the same cell. This confirms more rigorously data reported previously (Ogden et al., 1990) in which elevation of [Ca²⁺]_i by a bile acid was shown to inhibit InsP₃-evoked release as judged by the rate of change of the Ca²⁺-dependent K⁺ conductance in guinea pig hepatocytes. Other effects noted in these experiments were a reduction of the delay in InsP₃ activation of efflux (discussed above) and also a larger second response amplitude, possibly as a result of additional Ca²⁺ loading of the stores by the prior elevation of [Ca²⁺]_i.

Rate of Termination of Ca²⁺ Flux Correlates with d[Ca²⁺]_i/dt

Inspection of many [Ca²⁺]_i records from hepatocytes, and comparison with data from other tissues such as Purkinje neurons, showed that the duration of the period of InsP₃-evoked Ca²⁺ flux was brief when the flux was high, suggesting a mechanistic link between duration and Ca²⁺ flux. Evidence was presented above that InsP₃-evoked flux was inhibited at the peak [Ca²⁺]_i, indicating that the rate of the process terminating Ca²⁺ release determines the duration of the Ca²⁺ flux. The relation between duration and magnitude of flux was examined by making use of cell to cell variation of the flux, d[Ca²⁺]_i/dt, shown in Fig. 5 D. The rate of termination was measured as the reciprocal of the 10–90% risetime of the [Ca²⁺]_i and is plotted in Fig. 8 A against the flux measured as d[Ca²⁺]_i/dt (moles · s⁻¹ · liters⁻¹ cytosol) for each response. A good linear correlation was found between the rate of termination and the magnitude of flux (r = 0.91). Further, the large cell– cell variation when Ca²⁺ flux or peak [Ca²⁺]_i was plotted against InsP₃ concentration (see Fig. 5) is not apparent in the dependence of duration on flux shown in Fig. 8.

Figure 8.

Relation between rate of termination of Ca-flux and d[Ca²⁺]_i/dt in hepatocytes and Purkinje neurons. The duration of the InsP₃-evoked Ca²⁺ flux estimated as the reciprocal of the 10– 90% risetime (as indicated in Fig. 1) plotted against d[Ca²⁺]_i/dt. (A) Rate of termination (duration⁻¹) of the rise of [Ca²⁺]_i plotted against d[Ca²⁺]_i /dt on linear scales data from 48 hepatocytes. (B) Data from 14 isolated Purkinje neuronal soma, linear scales. Squares represent control cells, triangles data where [Ca²⁺]_i was elevated via voltage-dependent Ca²⁺ channels before InsP₃ release, inhibiting d[Ca²⁺]_i/dt to 10–20% of control. (C) Data from hepatocytes and Purkinje neurons plotted on log scales.

Cerebellar Purkinje neurons have large InsP₃-evoked Ca²⁺ fluxes, measured by d[Ca²⁺]_i/dt, and brief duration of [Ca²⁺]_i rise. This difference between tissues was used to see if a similar relation holds more generally over a wide range of Ca²⁺ flux. A plot of data obtained with the same protocol from Purkinje neurons, shown in Fig. 8 B, also has a good correlation between rate of termination and d[Ca²⁺]_i/dt. The data from hepatocytes and Purkinje neurons are shown plotted together on log scales in Fig. 8 C. If termination of Ca²⁺ flux is produced by Ca²⁺ binding to and inactivating channels then the rate of termination should be proportional to local Ca²⁺ concentration and hence the flux produced locally by open channels. The Ca²⁺ concentration in the cytosol close to InsP₃-gated channels due to Ca²⁺ flux through open channels will reach a high level quickly after activation, in <10 ms at distances up to 0.5 μm (see, e.g., Stern, 1992). In this model the rate of termination should be proportional to free [Ca²⁺] and hence to d[Ca²⁺]_i/dt. This mechanism and an alternative, store depletion, are discussed below.

Prolongation of [Ca²⁺]_i Increase by High InsP₃ Concentration

Evidence was obtained of prolongation of [Ca²⁺]_i as the InsP₃ concentration released by photolysis was increased and occurred even in conditions where the initial flux was the same. This effect was seen by integrating the [Ca²⁺]_i records in twin pulse experiments and is illustrated by the experiment in Fig 9. Initially (upper trace), a single pulse of 0.4 μM InsP₃ produced Ca²⁺ flux of d[Ca²⁺]_i/dt = 21 μM · s⁻¹, 10–90% risetime of high d[Ca²⁺]_i/dt of 240 ms, and area 34 μM · s. In the middle record the InsP₃ concentration was doubled by applying 2 pulses of 0.4 μM separated by 400 ms, producing no additional flux with the second pulse. The effect was to increase the area under the [Ca²⁺]_i trace to 52 μM · s without substantially increasing the flux (25 μM · s⁻¹), the duration of high d[Ca²⁺]_i/dt (10– 90% rise 280 ms), or peak [Ca²⁺]_i, thus indicating that the effect was to prolong the declining phase of [Ca²⁺]_i by prolonging release or inhibiting sequestration of free Ca²⁺. Subsequent photolytic release of 1 μM InsP₃ in the same cell produced a Ca²⁺ flux of d[Ca²⁺]_i/dt = 38 μM · s⁻¹, 10–90% risetime of 180 ms and area 67 μM · s, showing that responses to 0.4 μM and 2× 0.4 μM were submaximal. In five experiments of this kind the areas resulting from a single pulse releasing 1.2 μM InsP₃ were compared with the areas for 2 pulses of 1.2 μM InsP₃ in quick succession. The twin pulse areas were increased to 190 ± 42% (mean ± SEM, n = 5) relative to single pulse with no increase of peak flux. This can be compared with the area evoked by a single pulse releasing 2.4 μM InsP₃ which produced 201 ± 30% (mean ± SEM, n = 5) of the area evoked by a single pulse of 1.2 μM InsP₃ in the same cell. Comparison of the 10–90% risetimes in twin pulse experiments with those in controls shows that risetimes are prolonged by 116 ± 7% (mean ± SEM, n = 5) and that the 190% prolongation occurs predominantly after the peak [Ca²⁺]_i. Experiments in Ca²⁺-free medium showed no effect of Ca²⁺ removal on duration and that the prolonged [Ca²⁺]_i increase at high InsP₃ concentration is not a result of enhanced Ca²⁺ entry from the extracellular fluid. These data suggest that the InsP₃-evoked flux is prolonged at high InsP₃ concentration even when the peak flux is the same as at low concentration. The simplest explanation would be repeated activation of receptors recovering from Ca²⁺ inactivation, an effect that would be prolonged by the higher concentrations. Alternatively, a weaker inhibitory effect of elevated free Ca²⁺ at higher InsP₃ concentrations (Combettes et al., 1994) or a reduced rate of Ca²⁺ pumping into stores or out of the cell may account for the prolonged responses in double flash experiments.

Figure 9.

Dependence of the duration of [Ca²⁺]_i increase on InsP₃ concentration in hepatocytes. Photolytic release of InsP₃ indicated by the arrows in a single hepatocyte. Areas under the [Ca²⁺]_i increase compared between records following release of InsP₃. Records separated by 2 min. 500 μM furaptra, 0 mV. (Top record) Single pulse of 0.4 μM InsP₃ produced flux 21 μM · s⁻¹, peak [Ca²⁺]_i 6 μM and area 34 μM · s. (Middle record) Twin pulses of 0.4 μM InsP₃ separated by 500 ms. Flux 25 μM · s⁻¹, peak [Ca²⁺]_i 7 μM and area 52 μM · s. (Lower record) Single pulse of 1 μM InsP₃. Flux 38 μM · s⁻¹, peak [Ca²⁺]_i 9 μM, area 67 μM · s.

discussion

The hormonal activation of guinea pig hepatocytes, for example by norepinephrine, results in release of Ca²⁺ from intracellular stores as a series of spikes of high [Ca²⁺]_i relative to resting [Ca²⁺]_i, with fast rise and decline and a period of ∼20 s (Field and Jenkinson, 1987; rat hepatocytes, Woods et al., 1987). Spikes of [Ca²⁺]_i of similar waveform and period can be generated by perfusing InsP₃ directly into the cytosol from a patch pipette in whole cell patch clamp (Capiod et al., 1987; Ogden et al., 1990). This observation suggests that the mechanisms important in forming the spikes involve properties inherent in the InsP₃ receptor and its interaction with InsP₃ and free Ca²⁺. The aim of the experiments described above was to analyze in situ in single cells the kinetics of activation and inactivation of Ca²⁺ release, following InsP₃ release as a well defined pulse, to see how receptor properties shape the Ca²⁺ spike. The time course of [Ca²⁺]_i change after InsP₃ release comprises a latency, a period of rapid Ca²⁺ efflux into the cytosol which terminates abruptly at the peak where d[Ca²⁺]_i/dt is zero, and a slow recovery as Ca²⁺ is removed from the cytosol. InsP₃-evoked Ca²⁺ flux into the cytosol was measured as the rate of change of free cytosolic [Ca²⁺], d[Ca²⁺]_i/dt, and the rate of termination of the Ca²⁺ flux as the reciprocal of the 10– 90% risetime to the peak [Ca²⁺]_i.

InsP₃ Concentrations

InsP₃ reproducibly produced Ca²⁺ release in hepatocytes in whole cell patch clamp at concentrations of 0.1–0.2 μM or greater, approximately the same as that reported in permeabilized guinea pig hepatocytes (Burgess et al., 1984; Combettes et al., 1989) and similar to those in whole cell clamp endothelial cells (Carter and Ogden, 1992) and astrocytes, but much less than the concentrations needed in cerebellar Purkinje neurons (Khodakhah and Ogden, 1993). As judged by shortening of delays and the saturation of d[Ca²⁺]_i/dt, maximal InsP₃ concentrations were between 2 and 10 μM.

At low InsP₃ concentration (0.4 μM), a delay of mean 290 ms was found between release and the start of Ca²⁺ flux. In a previous study with photorelease of metabolically stable 5-thio-InsP₃ in hepatocytes, it was found that delays of many seconds duration occurred at low concentrations of 5-thio InsP₃ (Wootton et al., 1995), suggesting that in the absence of phosphatase activity lower concentrations of InsP₃ might be effective and produce longer delays. Further, this observation suggests that the accumulation of InsP₃ is important during hormone action, requiring InsP₃ generation locally at a rate that exceeds breakdown in order to produce a response and resulting in gradients of InsP₃ concentration (see Marty et al., 1989, for discussion).

It was found that the [Ca²⁺]_i increase was prolonged as InsP₃ concentration was increased and was due to an action additional to the increase in peak Ca²⁺ flux. This effect was demonstrated in closely spaced twin pulse experiments when, although the InsP₃ concentration was doubled, the peak flux activated was not increased and the duration of elevated [Ca²⁺]_i increased by a factor of 1.9. Comparison of 10–90% risetimes between twin pulse and control responses shows that the prolongation occurs after the peak [Ca²⁺]_i when the efflux from stores has fallen to a low level. If InsP₃ remaining after the peak [Ca²⁺]_i produces a low level of Ca²⁺ release, possibly due to the recycling of inactivated InsP₃ channels on a seconds timescale, then this result is expected simply because the higher concentration takes longer to metabolize and the low level activation of receptors persists for longer. The possibility that a second pool of activatable receptors is recruited to produce the prolongation is not supported by the observation that the second pulse of InsP₃ produces no immediate increase in Ca²⁺ flux into the cytosol, requiring simultaneous activation of a new pool of receptors as the initial pool is inhibited to sustain this explanation. The importance of breakdown of InsP₃ in the total duration of action is also apparent from data with 5-thio-InsP₃, which induces a small, steady, persistent component of Ca²⁺ release in hepatocytes (Wootton et al., 1995) and Purkinje neurons (Khodakhah and Ogden, 1995).

Origin of the Delay in Activation of Ca²⁺ Efflux at Low InsP₃ Concentrations

The delay between InsP₃ release in the cytosol and the rise of fluorescence of fluo-3 from baseline levels at resting [Ca²⁺]_i decreased from mean 290 ms at 0.4 μM InsP₃ to 30 ms or less at high concentration, close to the kinetic limit imposed by photolytic release of InsP₃. Similar delays were reported in fast perfusion experiments with permeabilized guinea pig hepatocytes (Champeil et al., 1989). Delays have also been reported in InsP₃ activation of Ca²⁺ release in smooth muscle (Walker et al., 1987; Somlyo et al., 1992), RBL cells (Meyer et al., 1990), oocytes (Parker and Ivorra, 1990), and Purkinje neurons (Khodakhah and Ogden, 1993, 1995). Evidence that Ca²⁺ ions act cooperatively with InsP₃, thus generating the delay at low InsP₃ concentration, was obtained in experiments in which the [Ca²⁺]_i was elevated before InsP₃ release, resulting in delays reduced or abolished when compared with controls in the same cell. In previous work (Ogden et al., 1990), elevation of [Ca²⁺]_i by a bile acid was shown to reduce the delay in the activation of the Ca²⁺-dependent K⁺ conductance. Reduction of the latency has also has been reported by Iino and Endo (1992) in smooth muscle and by Parker et al. (1996) in Xenopus oocytes. The whole cell averaged [Ca²⁺]_i required to produce this effect here was ∼2 μM, but the local concentration could be different because influx is via the plasma membrane. In the same recordings, it was found that [Ca²⁺]_i was high enough to have produced inhibition of the InsP₃-evoked Ca²⁺ efflux.

A minimal delay at high InsP₃ concentration has been reported in 2 studies, by Parker et al. (1996) in Xenopus oocytes with a protocol similar to that used here, who found a minimal delay of about 30 ms at resting [Ca²⁺]_i, and by Marchant and Taylor (1996) who found a minimal delay of about 30 ms in a perfused microsomal preparation of hepatocytes. Champeil et al. (1989) found a minimum resolvable delay of 20 ms in perfused permeabilized hepatocytes, which is, however, attributable to the resolution of rapid mixing studies. In the experiments reported here, the mean latency at 5–10 μM InsP₃ was 30 ms with a minimum of 25 ms. The photolysis of caged InsP₃ has a half-time of 3 ms, so it might be expected that a 5–10-fold supramaximal concentration would act with no detectable delay. Responses at 25–100 μM InsP₃ showed delays that could be distinguished from the optical artifact, supporting the idea that there is a minimal delay. The effect of raising [Ca²⁺]_i in reducing the delay to zero supports the interpretation of Parker et al. (1996) that Ca²⁺ priming of the receptor occurs before activation by InsP₃.

It is not clear from the present experiments, as in many other studies, whether there is cooperativity produced by multiple InsP₃ binding at constant [Ca²⁺]_i, as reported at low InsP₃ concentrations in RBL cells (Meyer et al., 1988). The concentrations of InsP₃ are higher here than those required in RBL and may be in the range where the “low concentration” Hill coefficient is substantially underestimated.

The delay in activation of the Ca²⁺-dependent K⁺ conductance is 100 ms longer than the delay of fluo-3 fluorescence at all but very high InsP₃ concentrations. This can be explained by the activation of the Ca²⁺- dependent K⁺ conductance at free [Ca²⁺] greater than resting levels and by the strongly cooperative activation by Ca²⁺ (Ogden et al., 1990; Capiod and Ogden, 1989a ). There is no evidence that the additional delay in activation of the conductance results from slow diffusion of Ca²⁺ to the membrane.

InsP₃-evoked Ca²⁺ Flux

After the latency, free [Ca²⁺] rises quite linearly with a rate, d[Ca²⁺]_i/dt, that depends on the flux of Ca²⁺ into unit cytosolic volume (moles · s⁻¹ · liter⁻¹) and on the proportion of Ca²⁺ bound to endogenous and exogenous buffers. If the unitary flux through InsP₃-gated channels is similar from one to another, then d[Ca²⁺]_i/dt measures in each cell the density of channels open in unit volume. This quantity d[Ca²⁺]_i/dt was found to increase with InsP₃ concentration up to 2.4 μM in each cell without showing evidence of cooperative InsP₃ binding and was maximal by about 7 μM. In guinea pig hepatocytes the peak [Ca²⁺]_i also increased with InsP₃ concentration in this range in each cell but with lower slope. The peak [Ca²⁺]_i at high InsP₃ concentration were 5–9.5 μM and rates of change d[Ca²⁺]_i/dt up to 52 μM · s⁻¹. These maximum values are small compared with cerebellar Purkinje neurons (Khodakhah and Ogden, 1995) and aortic endothelial cells (Carter and Ogden, manuscript in preparation) examined with the same methods, but are generally high compared with [Ca²⁺]_i reported for hormonal stimulation of hepatocytes in studies with fluorescent Ca²⁺ indicators (for review, Rooney and Thomas, 1991). In Xenopus oocytes (Parker et al., 1996), an increase in the flux, measured as rate of change of calcium green fluorescence, was found as InsP₃ concentration was increased, although the InsP₃ and Ca²⁺ concentrations were not measured.

A striking finding in these experiments was a large cell to cell variation in d[Ca²⁺]_i/dt measured at each InsP₃ concentration. This may be due to a large variation in density of InsP₃ receptors from cell to cell, a variation in the driving potential for Ca²⁺ efflux due to different degrees of store loading, or different extents of Ca²⁺ binding to cytosolic buffers. A large variation of peak [Ca²⁺]_i from cell to cell was also found here and in cerebellar Purkinje neurons (Khodakhah and Ogden, 1995).

Termination of InsP₃-evoked Ca²⁺ Release: Comparison with Cerebellar Purkinje Neurons

The net flux into the cytosol at the peak [Ca²⁺]_i is zero, indicating that the Ca²⁺ flux due to InsP₃ has declined to a low level at this time. It was noted that the duration of the period of high, fairly constant d[Ca²⁺]_i/dt during the rise of [Ca²⁺]_i was shortened when flux, d[Ca²⁺]_i/dt, was high. The rate of the process underlying the termination of flux, whatever the mechanism, can be represented by the reciprocal of the 10–90% rise time of [Ca²⁺]_i. A good linear correlation was found between this rate and d[Ca²⁺]_i/dt in liver cells, shown in Fig. 8 A, suggesting a simple one-step mechanism. The d[Ca²⁺]_i/dt represents the flux of free Ca²⁺ into unit cytosolic volume and will depend on the density of InsP₃ receptors. To test this idea, data from Purkinje neurons, which have a very high receptor density and high Ca²⁺ flux into the cytosol (Khodakhah and Ogden, 1995), were analyzed and presented in the same way in Fig. 8 B. A linear correlation with similar slope was found but over a much higher range of flux, suggesting that the same mechanism operates to terminate the Ca²⁺ flux in Purkinje neurons. The data from hepatocytes and Purkinje neurons are plotted together in Fig. 8 C on log–log scales, illustrating a similar correlation over a wide range, four log units, in both tissues. The large cell–cell variability seen when the flux or free [Ca²⁺] were plotted against InsP₃ concentration, was not present in the correlation between duration and flux, supporting a mechanistic link between rate of termination and the flux. Empirically, this correlation indicates that high Ca²⁺ flux into the cytosol via InsP₃ channels, produced, for instance, by a high channel density as in Purkinje neurons, results in a more rapid termination of Ca²⁺ release, producing a fast rising but brief pulse of free [Ca²⁺]. Local differences in InsP₃ receptor density might be expected to produce localized fast rising, brief free [Ca²⁺] changes, as well as producing localized high InsP₃ sensitivity as proposed by Hirose and Iino (1994).

Mechanism of Termination

The rate of termination correlates with the rate of appearance of free [Ca²⁺] in the cytosol of hepatocytes and Purkinje neurons over a wide range of flux. Purkinje neurons have a very large endogenous Ca²⁺ buffering, estimated as 2,000 Ca²⁺ ions bound for each free, due to high density of Ca²⁺-binding proteins (Fierro and Llano, 1996), compared with estimates of 50–100:1 in other cells (Neher, 1995), supporting the idea that free Ca²⁺ is important rather than total Ca²⁺ flux. The inhibition produced by cytosolic free [Ca²⁺] of the InsP₃ receptor has been shown in many tissues (Payne et al., 1988; Iino, 1990; Ogden et al., 1990; Ivorra and Parker, 1990; Bezprozvanny et al., 1991; Finch et al., 1991; Combettes et al., 1993). Strong inhibition of InsP₃-evoked Ca²⁺ release by Ca²⁺ influx through voltage-gated channels has been shown in Purkinje neurons (Khodakhah and Ogden, 1995) and through nicotinic channels in Xenopus oocytes (Parker et al., 1996). Similar experiments were less straightforward in hepatocytes. Previously (Ogden et al., 1990), a bile acid had been used to release Ca²⁺ from InsP₃-independent stores and shown to reduce the rate of rise of the Ca²⁺-dependent K⁺ conductance after InsP₃ release. In the present study, the [Ca²⁺]_i was raised more reproducibly by polarizing the membrane to −120 mV to increase the leak of Ca²⁺, and [Ca²⁺]_i was measured here with a fluorescent indicator. InsP₃-evoked flux was inhibited to ∼33% by prior elevation of [Ca²⁺]_i to ∼2 μM. Controlling the [Ca²⁺]_i close to resting levels with a high concentration of free fura-2 during InsP₃-evoked release prolonged the period of flux and prevented inactivation of the response normally seen to a second pulse of InsP₃ (5-thio-InsP₃ in this case). These experiments support the idea that free Ca²⁺ adjacent to the release sites inactivates InsP₃ receptor/channels. The free Ca²⁺ close to open channels is proportional to the flux through the channels and is established very quickly, in ∼10 ms at 0.5 μm calculated on the basis of Ca²⁺ diffusion from a discrete source (see Stern, 1992; Roberts, 1994). The linear correlation between rate of termination and free d[Ca²⁺]_i/dt, measuring flux into unit cytosolic volume, can therefore be explained by accumulation of free [Ca²⁺] close to the release sites and a first order inactivation by Ca²⁺ binding to channels. Open channels have a high and invariant [Ca²⁺] within the channel because of high constant flux density in the pore, and fast inactivation by pore [Ca²⁺] has been reported to determine the channel open lifetime of 2–3 ms (Bezprozvanny and Ehrlich, 1994). The whole cell flux d[Ca²⁺]_i/dt measured here depends on the open probability of InsP₃ channels and operates on a slower timescale of 10–1,000 ms. Peak open probability of InsP₃-gated channels is low (<0.15; Bezprozvanny et al., 1991) so a large proportion of channels are closed even with high InsP₃ concentrations. Because open channels have a high, constant rate of inhibition due to high pore [Ca²⁺], the dependence of termination rate on whole cell flux indicates an inhibition of closed channels by Ca²⁺ accumulating in the cytosol within a few hundred nm adjacent to open channels, an effect that would increase at high density of open channels (observed here as high d[Ca²⁺]_i/dt). This is supported by the observations discussed in connection with Fig. 6, that a second pulse of InsP₃ is ineffective even when applied during the period of high flux when open probability is high but submaximal, indicating that the process terminating flux acts on the closed channels. The first order dependence on [Ca²⁺]_i implied in the linear correlation between rate of termination and flux in both hepatocytes and Purkinje neurons differs from the conclusions of Oancea and Meyer (1996) from Ca²⁺ injection experiments, who found a relation suggesting 3–4 Ca²⁺ ions bind to produce inhibition of InsP₃-evoked Ca²⁺ transients in RBL cells.

Alternatively, the correlation between duration and size of flux could also be explained on the basis of store depletion, the higher the flux the sooner store Ca²⁺ runs out. The observation that the peak [Ca²⁺]_i is increased by increasing the InsP₃ concentration released in each cell, showing that, at low InsP₃, stores are not depleted when the flux is terminated. Furthermore, the prolonged responses in high fura-2 concentration also argue against store depletion acting alone. However, the observation, e.g., in Fig. 3, that loading the store by increasing Ca²⁺ influx into the cytosol increases the peak [Ca²⁺]_i evoked by InsP₃ indicates that store loading can influence Ca²⁺ flux in hepatocytes. The strong inhibitory effect of raised [Ca²⁺]_i suggests that Ca²⁺ feedback is the primary mechanism, but store depletion or InsP₃-evoked desensitization of receptors (Hajnozcky and Thomas, 1994; Ilyin and Parker, 1994) may contribute to termination of the Ca²⁺ flux.

The conclusions from these experiments are (a) that the properties of the InsP₃ receptor can account for the time course of [Ca²⁺]_i spikes in guinea pig hepatocytes, the onset and rise determined by the activation and inactivation of InsP₃ receptor, and the minimum duration of the interspike period by the recovery of InsP₃ sensitivity described previously (Ogden et al., 1990), and (b) that the differences in kinetics of InsP₃-evoked Ca²⁺-release between hepatocytes and Purkinje neurons can be explained by the differences in InsP₃ receptor density. The longer latencies at low concentration, slower rise of free [Ca²⁺], and longer duration of the InsP₃-evoked flux into the cytosol are mostly explicable by the low density of InsP₃ receptors in liver, producing small flux and correspondingly long periods of Ca²⁺ release due to slow Ca²⁺-inactivation of release channels. These properties are consistent with prolonged, pulsatile Ca²⁺ signalling in liver cells which underlies hormonal regulation of processes such as glycogenolysis. The different receptor subtypes, mainly type 2 InsP₃ receptors with some type 1 in hepatocytes, and type 1 alone in Purkinje neurons (DeSmedt et al., 1994) would appear secondary to receptor density in determining the kinetics of Ca²⁺ release but may be important in determining other factors such as InsP₃ sensitivity or susceptibility to phosphorylation. More generally, the linear relation between Ca²⁺ flux into unit cytosolic volume and rate of termination of Ca²⁺ release, shown here for hepatocytes and Purkinje neurons over a wide range of Ca²⁺ flux, indicates that regulation of the InsP₃ receptor by cytosolic free [Ca²⁺] produces fast rising but brief pulses of [Ca²⁺]_i at high InsP₃ receptor densities.