(Received for publication, September 17, 1996, and in revised form, November 8, 1996)
From the Department of Cellular and Molecular Physiology, Tufts University School of Medicine, Boston, Massachusetts 02111
The regulation of the inositol 1,4,5-trisphosphate (IP3) receptor in liver was analyzed using a novel superfusion method. Hepatic microsomes were loaded with 45Ca2+, and superfused at high flow rates to provide precise control over IP3 and Ca2+ concentrations ([Ca2+]) and to isolate 45Ca2+ release from reuptake. 45Ca2+ release was dependent on both [Ca2+] and IP3. The initial rate of 45Ca2+ release was a biphasic function of [Ca2+], increasing as [Ca2+] approached 3 µM but decreasing at higher concentrations, suggesting that the hepatic IP3 receptor is regulated by [Ca2+] at two sites, a high affinity potentiation site and a low affinity inhibitory site. The relationship between initial rates and IP3 concentration was steep (Hill coefficient of 3.4), suggesting that activation of the calcium channel requires binding of at least 3 IP3 molecules. IP3 concentrations above 10 µM produced rapid decay of release rates, suggesting receptor inactivation. Superfusion with 10 µM IP3 under conditions that minimize calcium release ([Ca2+] < 1 nM) inhibited 45Ca2+ release in response to subsequent stimulation (400 nM Ca2+). These data suggest sequential positive and negative regulation of the hepatic IP3 receptor by cytosolic calcium and by IP3, which may underlie hepatocellular propagation of regenerative, oscillatory calcium signals.
In many cells including hepatocytes, binding of certain hormones to cell surface receptors results in the activation of phospholipase C, thereby generating the intracellular second messenger inositol 1,4,5-trisphosphate (IP3).1 The intracellular receptor for IP3 is a calcium channel that permits efflux of Ca2+ sequestered in the endoplasmic reticulum out into the cytoplasm. Calcium diffuses through the cytoplasm, and under certain conditions, kinematic waves of elevated Ca2+ concentration can be observed (1). The amplitude of these waves is roughly constant, but the frequency of the waves correlates with extracellular receptor agonist concentration. Thus, calcium waves couple the activation of cell surface receptors to activation of calcium-dependent intracellular processes in a frequency-encoded manner (2). The importance and complexity of changes in Ca2+ concentration in transmitting information implies that calcium signaling is tightly regulated.
The IP3 receptor occupies a central position among the multitude of molecules involved in intracellular calcium metabolism (1, 2). It possesses regulatory properties that in principle can account for the spatio-temporal complexity of calcium signals (1, 2). In many tissues, cytoplasmic Ca2+ concentrations modulate IP3-dependent calcium release in a biphasic manner (3-7). At submicromolar concentrations, Ca2+ activates the release process, whereas, at higher concentrations, Ca2+ is inhibitory. This regulation by calcium has been proposed to be a means of controlling calcium oscillations (1, 2). The role of IP3 in regulating the activity of its receptor is less clear. Results that support (8) as well as undermine (5, 9, 10) cooperative receptor activation have been reported. Recently, an inhibitory role for IP3 has been proposed (11).
The aim of this study was to characterize the regulation of the hepatic IP3 receptor by both Ca2+ and IP3. A superfusion assay, which was developed to study IP3-dependent calcium release from brain microsomes (5), was adapted for use with liver microsomes. This assay achieves subsecond time resolution, allows for precise control of extravesicular Ca2+ and IP3 concentrations, and isolates the release of 45Ca2+ from its reuptake. We show that both extravesicular Ca2+ and IP3 can activate and inhibit the calcium channel function of the IP3 receptor.
D-myo-Inositol 1,4,5-trisphosphate (hexapotassium salt) was purchased from LC Laboratories (Woburn, MA). Magnesium ATP, MOPS, dithiothreitol (DTT), EDTA, EGTA, HEPES, Percoll, phenylmethylsulfonyl fluoride, antipain, leupeptin, pepstatin, and other chemicals were obtained from Sigma. Hanks' buffer saline solution was from Life Technologies, Inc. Fura 2 was from Molecular Probes (Eugene, OR). 45CaCl2 was purchased from DuPont NEN.
Preparation of MicrosomesA low-sedimentation low-density
microsomal fraction (12) was prepared from sodium
pentobarbital-anaesthetized male Sprague-Dawley rats (200-300 g),
perfused in situ with 100 ml of Hanks' buffer saline
solution containing 0.5 mM EGTA, 1 mM DTT, and
protease inhibitors (phenylmethylsulfonyl fluoride, 100 µg/ml,
leupetin, 2 µg/ml, pepstatin, 1 µg/ml, and antipain, 2 µg/ml, pH
7.3). The tissue was minced in 40 ml of ice-cold sucrose buffer (250 mM sucrose, 5 mM HEPES, 5 mM KOH, 1 mM EDTA, 1 mM DTT, and protease inhibitors, pH
adjusted with KOH to 7.4), homogenized with a glass tissue grinder
(BellCo) (10 loose strokes, 10 tight strokes), and centrifuged for 10 min at 500 × g. The supernatant was centrifuged for 10 min at 2,000 × g and the resulting supernatant was
further centrifuged for 40 min at 36,000 × g to
sediment the microsomes (low sedimentation). The pellet was resuspended
in sucrose buffer containing 35% (v/v) Percoll and fractionated by
centrifugation at 36,000 × g for 30 min. The superior
low density fraction was washed twice in 10 volumes of MK buffer (120 mM KCl, 20 mM MOPS, pH 7.2), with 1 mM DTT and protease inhibitors. The final pellet was
resuspended at 2-5 mg/ml in MK buffer with protease inhibitors and
0.7 M glycerol, divided into 50-µl portions, and stored
at 70 °C.
Microsomes were thawed on ice and loaded with 45Ca2+ for 10 min at room temperature in 2.5 volumes of loading buffer (MK buffer containing 5 mM MgATP, 50 µM EGTA, 45CaCl2 (free 45Ca2+ 10 µM), 1 mM DTT, and protease inhibitors, pH 7.2). Loaded microsomes were then superfused as described previously (13). Briefly, the microsomes were washed in 16 volumes of basal buffer (100 mM KCl, 20 mM MOPS, 1 mM EGTA, 5 mM MgCl2, free Ca2+ less than 1 nM), applied to a filter support, and mounted in a superfusion chamber accessed by three solenoid-driven valves that were operated by computer. Each valve controlled the delivery of a separate pressurized solution to the chamber. Microsomes were superfused at 40 p.s.i. (flow rate ~ 2 ml/s) for 1.05 s with basal buffer to establish a baseline rate of release. Release was evoked at t = 0, by switching to a second stimulation buffer (100 mM KCl, 20 mM MOPS, 4 mM EDTA, 5 mM MgCl2, sufficient CaCl2 to obtain the desired free calcium concentration, and IP3 at the concentration specified for each experiment). The superfusate containing released 45Ca2+ was collected with a rotating fraction collector, which divided the effluent stream into 70-ms fractions. Radioactivity in each fraction and that remaining on the filter was measured by liquid scintillation counting (Beckman LS 6800) after addition of 2 ml of Biosafe II (Research Products International, Mount Prospect, IL) to each fraction. Experiments were performed in groups of 6, 3 with the desired concentrations of IP3 and 3 with no IP3 to control for calcium-dependent release, in a randomized sequence to minimize any time-dependent changes in the preparation.
Calculations45Ca2+ release rates were calculated by dividing the amount of radioactivity in each fraction by the total amount remaining on the filter, expressed as % of total. In all experiments, the baseline calcium release rates were adjusted to the mean of 45Ca2+ release during the 3 fractions preceding the switch to the stimulating buffer. This adjustment was generally on the order of 5-10% of the baseline rate of release. Net IP3-dependent calcium release was calculated by subtracting the release rates observed in the absence of IP3 from the rates in the presence of IP3 for each concentration of Ca2+ (5). Cumulative IP3-dependent calcium release was obtained by stepwise addition of IP3-dependent calcium release rates. The peak rate of release was defined as the largest IP3-dependent 45Ca2+ release rate in a single fraction, expressed in percent per s. The initial rate of release was defined as the rate of 45Ca2+ release (expressed in percent per s) derived from the cumulative release between t = 0 and the fraction with maximum rate of release.
Calcium BuffersFree Ca2+ concentrations in the loading and stimulation buffers were calculated according to Fabiato and Fabiato (14), assuming a nominal Ca2+ concentration of 3 µM in distilled water. These calculations agreed with measurements of free Ca2+ obtained with Fura 2 (for free Ca2+ concentrations less than 1 µM) or with an ion-selective electrode (for free Ca2+ concentrations above 1 µM).
Superfusion of
45Ca2+ loaded hepatic microsomes with a
stimulation buffer containing 400 nM Ca2+ and
no IP3 resulted in a small but sustained increase in the rate of 45Ca2+ efflux (Fig.
1A). When the stimulus buffer contained 400 nM Ca2+ and 10 µM
IP3, the release event reached a maximum within two fractions (~140 ms), and subsequently decayed toward a plateau rate
that was significantly higher than that observed with 400 nM Ca2+ alone.
IP3-dependent release was blocked by heparin
(100 µg/ml) (Fig. 1B), and was not due to
calcium-dependent 45Ca2+ release
(such as that mediated by the ryanodine receptor), since superfusion
with 100 µM Ca2+ in the absence of
IP3 produced the same amount of calcium release as did
superfusion with 400 nM Ca2+ alone (not shown).
Nonetheless, the small amount of calcium-dependent 45Ca2+ release was subtracted from the total
release rates to arrive at the net
IP3-dependent release rates (see
"Experimental Procedures"). The net
IP3-dependent release event (Fig.
1C) could be well described as the sum of two exponential
decay phases, a fast phase that decayed with a time constant
1 = 130 ms, and a slower component that decayed slowly
(
2 = 0.73 s). The cumulative
45Ca2+ release event could be reconstructed by
sequentially summing the release rates (Fig. 1D).
Characterization of the Effects of Ca2+ on the IP3 Receptor
The effects of extravesicular
Ca2+ concentration on IP3 receptor activity
were examined using a standard IP3 concentration of 10 µM, and measuring the net
IP3-dependent 45Ca2+
release rates. Increasing the Ca2+ concentrations from
nanomolar levels up to 3 µM resulted in marked changes in
the kinetics of 45Ca2+ release; the maximal
rate of release was greater and occurred more rapidly, and the decay of
the fast component of release was accelerated (Fig.
2A). However, with increases in
Ca2+ concentrations above 3 µM, the maximal
rate of release and the cumulative release of
45Ca2+ decreased. Analysis of the relationship
between the maximal rate of release or the initial rate of release and
the extravesicular Ca2+ concentration was biphasic (Fig.
2B) as has been shown for the IP3-dependent 45Ca2+
release in rat brain (3, 5). Significantly, when we analyzed the
relationship between cumulative release and extravesicular Ca2+ concentration, we observed stimulation of release with
Ca2+ concentrations less than 1 µM, but the
concentration-response relationship was shifted to lower
Ca2+ concentrations. This observation emphasized the need
to make measurements on a subsecond time scale in order to study the
aspects of regulation of the transient properties of the
IP3 receptor.
Characterization of the Effects of IP3 on the IP3 Receptor
To investigate the potency of
IP3 in this system, the IP3 concentration of
the stimulation buffer was varied while holding the free
Ca2+ concentration constant (400 nM). As the
IP3 concentration in the stimulation buffer was increased
from 0.3 to 50 µM, the maximal rate of release increased
and occurred faster (Fig. 3A). The fast component of calcium release was not detectable at IP3
concentrations of 300 nM or less, but could be observed as
the IP3 concentration was increased into the micromolar
range. The amplitude of the slow component of release was similar at
all IP3 concentrations, but the peak rate was reached more
rapidly as the concentration of the agonist was increased. Decay of the
fast component increased with the IP3 concentration, but
the decay was not solely dependent on depletion of intravesicular
calcium. Because of the increased rate of decay at IP3
concentrations above 10 µM, the intravesicular 45Ca2+ content at the end of the fast phase was
greater than for lower concentrations of IP3 where decay of
release rates was slower.
Inspection of the cumulative IP3-dependent 45Ca2+ release clearly showed sigmoidal kinetics, most demonstrably at lower IP3 concentrations (Fig. 3B). Concentration-response relationships were plotted for initial release rates and cumulative release after 2.55 s of stimulation (Fig. 3C). The initial rate data were fit by non-linear least squares regression, yielding the following parameters; the Hill coefficient (nH) was 3.4, the Vmax value was 4.4% per s, and the apparent Km value for IP3 was 4.5 µM. Significantly, analysis of the relationship between IP3 concentration and the cumulative release measured after 2.55 s of stimulation was best described by a normal hyperbolic fit, i.e. the Hill coefficient of these data is 1.2, Vmax value was 1.7% per s, and the apparent Km value for IP3 was 0.9 µM. This analysis demonstrates that the sigmoidicity of the concentration-response analysis is best observed when measuring the initial release events.
IP3-dependent Inhibition of the IP3 ReceptorTo test the possibility that
IP3 itself inhibited IP3-dependent
calcium release, we took advantage of the fact that the IP3 receptor requires both calcium and IP3 for its activity.
This permitted us to pretreat the microsomes with IP3 and
no added Ca2+ ([Ca2+] < 1 nM)
for various intervals, conditions that produced only small increases in
the rate of 45Ca2+ release. Thus, the
IP3 receptor can be occupied by ligand without depleting
intravesicular 45Ca2+, which could confound
attempts to measure receptor inactivation. The vesicles were superfused
with basal buffer as usual, and at t = 1.50 s, the
superfusion buffer was switched to a prestimulus solution that
contained either 1 nM Ca2+ (
,
) or 1 nM Ca2+, 10 µM IP3
(
) (Fig. 4A). At t = 0, the solution was again switched to a test buffer containing 400 nM Ca2+, 10 µM IP3
(
,
), or 400 nM Ca2+ only (
).
Prestimulation with 10 µM IP3, 1 nM Ca2+ produced a small amount of
45Ca2+ release (Fig. 4, B and
C), but not enough to significantly reduce the amount of
45Ca2+ contained in the vesicles (less than
0.2% of total). However, prestimulation with IP3
significantly inhibited the response to the test buffer. The initial
rate of 45Ca2+ release was diminished by 42%
by prestimulation with 10 µM IP3. When the
peak rate of release observed after prestimulation was scaled to the
peak rate of release in the control experiment (×), it was apparent
that there was no qualitative change in the rate of decay or the
relative amplitudes of the two components, consistent with
agonist-dependent inactivation of the IP3
receptor, leading to a diminished response to the test buffer
due to a decrease in the number of receptors available to conduct
Ca2+ efflux.
The aim of this study was to examine the regulation of the IP3 receptor by two second messengers implicated in its function, IP3 and Ca2+. We used a subsecond kinetic approach to measure 45Ca2+ release from microsomes prepared from rat liver. IP3-dependent 45Ca2+ release from hepatic microsomes was rapid and transient when strong stimulation was used. The decay of this release event was fit by two exponentials, representing an initial fast phase superimposed on a slower phase of Ca2+ release, as noted previously (9, 15). The Ca2+ release specifically represents IP3-dependent 45Ca2+ release since it was abolished by heparin, which interferes with the binding of IP3 to its receptor (16). The activity was regulated by both Ca2+ and IP3. Furthermore, pre-exposure of the receptor to IP3 (at [Ca2+] < 1 nM) produced a decrease in the activity of the receptor in response to subsequent test solutions. Our results are consistent with a model for the IP3 receptor where either Ca2+ or IP3 are sufficient to promote receptor inactivation, but both are necessary to promote receptor activation.
The superfusion method used in this study possesses unique advantages.
Classical filtration studies rely on the calculation of the
IP3-dependent calcium release by measuring the
45Ca2+ remaining in the stores of different
experimental samples before and after partial discharge by
IP3. The superfusion method directly and continuously
measures release rates. The high flow rate (2 ml/s) relative to the
small volume of the superfusion chamber (~30 µl) allows rapid
solution changes ( = 40 ms) and precise control of the
extravesicular concentrations of IP3 and calcium. Continuous superfusion of the membranes alleviates possible agonist depletion or degradation. The resulting rapid chelation of released 45Ca2+ by EDTA isolates the efflux of
Ca2+ from its reuptake and minimizes accumulation of
Ca2+ near the channel pore which might influence its
activity. This contrasts with studies of permeabilized hepatocytes, in
which changes in fluorescence of the calcium-sensitive dye Fura 2 were measured. This experimental protocol is limited, due to membrane permeabilization (17), the inability to isolate Ca2+
release from Ca2+ reuptake, and poor control over changes
in IP3 and Ca2+ concentrations in the vicinity
of the IP3 receptor.
Many cells have multiple mechanisms in place to produce calcium release activity. In addition to the IP3 receptor, hepatocytes contains a ryanodine receptor (18), which is structurally and functional related to the IP3 receptor. The ryanodine receptor is thought to be responsible for Ca2+-induced Ca2+ release from the stores, and can be distinguished from the IP3 receptor by functional criteria. In addition to the being a ligand-gated receptor, the IP3 receptor is relatively insensitive to Mg2+ (5, 19). Any ryanodine receptor activity in this preparation would be blocked by the high Mg2+ concentration (5 mM) used in the superfusion buffers. Superfusion with Ca2+ at high concentrations induced only a minimal Ca2+ release, an observation consistent with low levels of ryanodine receptor activity. In any case, IP3-dependent Ca2+ release was always corrected, using the control experiment performed at the same free Ca2+ concentration in the absence of IP3.
The initial rate of calcium release was a steep function of IP3 concentration. The apparent Hill coefficient of 3.4 suggests that binding of at least 3 IP3 molecules to the channel is required for activation of the ion channel activity, as suggested previously (8). Our results differ from those reported by Finch et al. (5) who used the same assay with synaptosomes-derived microsomal vesicles. Their experiments were performed at 10 µM Ca2+, which could explain a loss in apparent cooperativity. Combettes (20), using permeabilized hepatocytes, reported significantly reduced apparent cooperativity at a Ca2+ concentration of 10 µM compared to 100 nM. In contrast, the relationship between IP3 concentration and cumulative Ca2+ release (after 2.55 s of stimulation) had a Hill coefficient of ~1.2. The apparent potency of IP3, when based on cumulative release (0.9 µM), was somewhat higher than that based on the initial rate of release (4.5 µM). Using a stopped-flow method with permeabilized hepatocytes, Champeil et al. (9) determined comparable apparent affinities (0.2 and 1 µM respectively). At 10 µM free Ca2+ concentration, i.e. at saturating concentrations of Ca2+ required for the coagonist effect, Finch et al. (5) found that the initial release rate and the cumulative release had the same dependence on IP3 concentration, suggesting that IP3-dependent Ca2+ release was transient even at low IP3 concentrations. Combettes (19) reported that, at low concentrations of both Ca2+ and IP3, the release event is sustained, which is consistent with our data. This was subsequently confirmed by Finch (21); at 10 and 300 nM free Ca2+ concentrations, well below the range of extravesicular Ca2+ concentrations that inhibit release, the apparent potency of IP3 was higher when cumulative measurements of Ca2+ release over 2 s were used rather than the maximum rate of 45Ca2+ release.
IP3-dependent inactivation of IP3 receptor has been reported previously (11). In these experiments, the Ca2+ channel of the IP3 receptor function was assessed by measuring the quench of Fura 2 fluorescence by Mn2+ entry into the stores. This approach assumes that Mn2+ moves inward through the IP3 receptor from the cytoplasm of permeabilized hepatocytes into empty Ca2+ stores, and mimics the flux of Ca2+ through the Ca2+ channel. Our design allows preincubation with IP3 without significant movement of Ca2+ and measures directly Ca2+ efflux from microsomes with a better time resolution. Preincubation with IP3 inhibited release in a time dependent manner. It appears that inhibition of the Ca2+ channel by IP3 is faster than previously thought and might be rapid enough to explain the transient nature of the IP3-dependent Ca2+ release observed at constant extravesicular Ca2+.
A biphasic effect of cytosolic Ca2+ on IP3-dependent Ca2+ release has been described in Xenopus oocytes (3), smooth muscle (4), brain (5, 6), and hepatocytes (7). This property is important to our understanding of the mechanisms of Ca2+ waves (2, 22, 23), but has been recently suggested to be an effect produced by the metal-free Ca2+ chelators used to buffer calcium (19, 24). EDTA has similar affinities for Ca2+ and Mg2+, and in the presence of excess of Mg2+, nearly all of the Ca2+-free EDTA is complexed with Mg2+. Under these conditions where metal-free EDTA concentrations are minimal, Combettes (19) found that the cumulative amount and initial rate of IP3-dependent Ca2+ release had only weak dependence on Ca2+ concentrations between 30 nM and 2 µM. Measuring IP3-dependent Ca2+ release from permeabilized hepatocytes after 5 s, the same investigators reported that increasing the free Ca2+ concentration from 1 nM to 1 µM in this EDTA buffer did not stimulate IP3-dependent Ca2+ releases at high IP3 doses (20). We exposed hepatic microsomes to IP3 (10 µM) and the desired free Ca2+ concentration using EDTA as a calcium buffer. We found that when measured on a subsecond time scale, cytosolic Ca2+ at submicromolar concentration activated the initial rate of IP3-dependent 45Ca2+ release, consistent with the role of Ca2+ as a coagonist of the IP3 receptor. Our results do not exclude the possibility that metal-free chelators are inhibitors of the IP3 receptor, but they confirm that at submaximal IP3 concentration, Ca2+ activates IP3-dependent calcium release during the first 100-200 ms of stimulation.
The transient nature of the IP3-dependent calcium release observed in these experiments is potentially an important property of the IP3 receptor. Repetitive oscillatory calcium signals that diffuse as kinematic waves can only occur if mechanisms exist to terminate calcium release. Many cellular mechanisms exist which could explain the transient nature of the IP3 induced calcium release from hepatic microsomes. For example, sequential increases and decreases in IP3 concentration, local transients of Ca2+ concentration near the mouth of the pore, or depletion of intravesicular Ca2+ all could lead to transient release patterns. These possibilities seem unlikely in our system, since agonist concentrations were held constant during the experiment, and the vesicles were superfused at high flow rates with a strong Ca2+ buffer to clamp the extravesicular Ca2+ concentration. As for depletion, the vesicles were always loaded to the same extent, and therefore to the same initial intravesicular Ca2+ concentration. At higher IP3 concentrations, the fast phase of release decayed more rapidly, resulting in higher luminal Ca2+ content at the end of the release event (Fig. 3b), arguing that the decay is not solely due to depletion of intravesicular Ca2+.
The simplest model to account for our observations incorporates
IP3-dependent as well as
calcium-dependent inactivation of the receptor. By mass
action, increases in IP3 concentrations lead to an increase
in the fraction of receptors in the fully liganded conformation (Fig.
5). By analogy with cell surface ionotropic receptors
such as the nicotinic acetylycholine receptor (25) and
N-methyl-D-aspartate receptors (26), this
conformation is in equilibrium with two states, the open state that
conducts calcium, and an inactive conformation that does not. As the
calcium concentration is increased into the micromolar range, the
equilibrium between fully-liganded receptor and the active ion channel
is shifted to favor the activated state, allowing for calcium efflux.
At nanomolar calcium concentrations, the fully liganded receptor enters
the inactive conformation without going through the active state. Thus,
at low calcium concentrations, application of IP3 leads to
a significant fraction of receptors that are occupied by
IP3, but the open state of the receptor is bypassed,
resulting in accumulation of the receptor in the inactivated state.
Numerical simulations (27) using this model faithfully reproduce a
number of important features of the empirical results, including the kinetics of the release event, the biphasic effect of Ca2+
on the response, and the effect of pre-exposure to IP3 at
subnanomolar [Ca2+]. Repetitive propagation of kinematic
Ca2+ waves within hepatocytes can only occur with limited
autocatalytic mechanisms capable of terminating and reinitiating
Ca2+ signals. Our data, supported by this quantitative
model, show that sequential positive and negative feedback regulation
of the channel by both cytosolic Ca2+ and IP3
may provide such mechanisms.
![]() |
![]() |
![]() |
In conclusion, we have conducted experiments designed to obtain quantitative data regarding the regulation of the IP3 receptor by the two second messengers, Ca2+ and IP3. We have provided substantive evidence that both IP3 and Ca2+ alter the calcium channel activity of the receptor in a similar fashion, such that either ligand promotes receptor inactivation, but both coagonists are required to maximize calcium release activity. The coincidence properties of the IP3 receptor are likely to have profound implications in the maintenance of intracellular calcium waves, and in regulation of cellular activity.