Conformational transitions in the Ca2+ + Mg2+-activated ATPase and the binding of Ca2+ ions.

We have studied the fluorescence of the Ca2+ + Mg2+-activated ATPase of sarcoplasmic reticulum labelled with fluorescein isothiocyanate. The change in intensity of fluorescein fluorescence caused by addition of Ca2+ to the labelled ATPase can be interpreted in terms of a two-conformation model for the ATPase, one conformation (E1) having a high affinity for Ca2+, the other (E2) a low affinity. Effects of Ca2+ as a function of pH allow an estimate of the effect of pH on the E1/E2 ratio, consistent with kinetic studies. A model is presented for binding of Ca2+ to the ATPase as a function of pH that is consistent both with the data on the E1/E2 equilibrium and with literature data on Ca2+ binding.     

Two roles of Ca2+ in agonist stimulated Ca2+ oscillations.

We propose a mechanism for agonist-stimulated Ca2+ oscillations that involves two roles for cytosolic Ca2+: (a) inhibition of inositol-1,4,5-trisphosphate (IP3) stimulated Ca2+ release from the endoplasmic reticulum (ER) and (b) stimulation of the production of IP3 through its action on phospholipase C (PLC), via a Gq protein related mechanism. Relying on quantitative experiments by Parker, I., and I. Ivorra (1990. Proc. Natl. Acad. Sci. USA. 87:260-264) on the inhibition of Ca2+ release from the ER using caged-IP3, we develop a kinetic model of inhibition that allows us to simulate closely their experiments. The model assumes that the ER IP3 receptor is a tetramer of independent subunits that can bind both Ca2+ and IP3. Upon incorporation of the action of Ca2+ on PLC that leads to production of IP3, we observe in-phase-oscillations of Ca2+ and IP3 at intermediate values of agonist stimulation. The oscillations occur on a time scale of 10-20 s, which is comparable to the time scale for inhibition in Xenopus oocytes. Analysis of the mechanism shows that Ca(2+)-inhibition of IP3-stimulated Ca2+ release from the ER is an essential step in the mechanism. We also find that the effect of Ca2+ on PLC can lead to an indirect increase of cytosolic Ca2+, superficially resembling "Ca(2+)-induced Ca(2+)-release." The mechanism that we propose appears to be consistent with recent experiments on REF52 cells by Harootunian, A. T., J. P. Y. Kao, S. Paranjape, and R. Y. Tsien. (1991. Science [Wash. DC]. 251:75-78.) and we propose additional experiments to help test its underlying assumptions.     

Ca2+-calmodulin-dependent facilitation and Ca2+ inactivation of Ca2+ release-activated Ca2+ channels

In non-excitable cells, one major route for Ca2+ influx is through store-operated Ca2+ channels in the plasma membrane. These channels are activated by the emptying of intracellular Ca2+ stores, and in some cell types store-operated influx occurs through Ca2+ release-activated Ca2+ (CRAC) channels. Here, we report that intracellular Ca2+ modulates CRAC channel activity through both positive and negative feedback steps in RBL-1 cells. Under conditions in which cytoplasmic Ca2+ concentration can fluctuate freely, we find that store-operated Ca2+ entry is impaired either following overexpression of a dominant negative calmodulin mutant or following whole-cell dialysis with a calmodulin inhibitory peptide. The peptide had no inhibitory effect when intracellular Ca2+ was buffered strongly at low levels. Hence, Ca2+-calmodulin is not required for the activation of CRAC channels per se but is an important regulator under physiological conditions. We also find that the plasma membrane Ca2+ATPase is the dominant Ca2+ efflux pathway in these cells. Although the activity of the Ca2+ pump is regulated by calmodulin, the store-operated Ca2+ entry is more sensitive to inhibition by the calmodulin mutant than by Ca2+ extrusion. Hence, these two plasmalemmal Ca2+ transport systems may differ in their sensitivities to endogenous calmodulin. Following the activation of Ca2+ entry, the rise in intracellular Ca2+ subsequently feeds back to further inhibit Ca2+ influx. This slow inactivation can be activated by a relatively brief Ca2+ influx (30-60 s); it reverses slowly and is not altered by overexpression of the calmodulin mutant. Hence, the same messenger, intracellular Ca2+, can both facilitate and inactivate Ca2+ entry through store-operated CRAC channels and through different mechanisms.     

Divergence of Ca2+ selectivity and equilibrium Ca2+ blockade in a Ca2+ release-activated Ca2+ channel

Prevailing models postulate that high Ca²⁺ selectivity of Ca²⁺ release-activated Ca²⁺ (CRAC) channels arises from tight Ca²⁺ binding to a high affinity site within the pore, thereby blocking monovalent ion flux. Here, we examined the contribution of high affinity Ca²⁺ binding for Ca²⁺ selectivity in recombinant Orai3 channels, which function as highly Ca²⁺-selective channels when gated by the endoplasmic reticulum Ca²⁺ sensor STIM1 or as poorly Ca²⁺-selective channels when activated by the small molecule 2-aminoethoxydiphenyl borate (2-APB). Extracellular Ca²⁺ blocked Na⁺ currents in both gating modes with a similar inhibition constant (K_i; ∼25 µM). Thus, equilibrium binding as set by the K_i of Ca²⁺ blockade cannot explain the differing Ca²⁺ selectivity of the two gating modes. Unlike STIM1-gated channels, Ca²⁺ blockade in 2-APB–gated channels depended on the extracellular Na⁺ concentration and exhibited an anomalously steep voltage dependence, consistent with enhanced Na⁺ pore occupancy. Moreover, the second-order rate constants of Ca²⁺ blockade were eightfold faster in 2-APB–gated channels than in STIM1-gated channels. A four-barrier, three–binding site Eyring model indicated that lowering the entry and exit energy barriers for Ca²⁺ and Na⁺ to simulate the faster rate constants of 2-APB–gated channels qualitatively reproduces their low Ca²⁺ selectivity, suggesting that ion entry and exit rates strongly affect Ca²⁺ selectivity. Noise analysis indicated that the unitary Na⁺ conductance of 2-APB–gated channels is fourfold larger than that of STIM1-gated channels, but both modes of gating show a high open probability (P_o; ∼0.7). The increase in current noise during channel activation was consistent with stepwise recruitment of closed channels to a high P_o state in both cases, suggesting that the underlying gating mechanisms are operationally similar in the two gating modes. These results suggest that both high affinity Ca²⁺ binding and kinetic factors contribute to high Ca²⁺ selectivity in CRAC channels.     

Saltatory propagation of Ca2+ waves by Ca2+ sparks.

Punctate releases of Ca2+, called Ca2+ sparks, originate at the regular array of t-tubules in cardiac myocytes and skeletal muscle. During Ca2+ overload sparks serve as sites for the initiation and propagation of Ca2+ waves in myocytes. Computer simulations of spark-mediated waves are performed with model release sites that reproduce the adaptive Ca2+ release observed for the ryanodine receptor. The speed of these waves is proportional to the diffusion constant of Ca2+, D, rather than D, as is true for reaction-diffusion equations in a continuous excitable medium. A simplified "fire-diffuse-fire" model that mimics the properties of Ca2+-induced Ca2+ release (CICR) from isolated sites is used to explain this saltatory mode of wave propagation. Saltatory and continuous wave propagation can be differentiated by the temperature and Ca2+ buffer dependence of wave speed.     

On the roles of Ca2+ diffusion, Ca2+ buffers, and the endoplasmic reticulum in IP3-induced Ca2+ waves.

We have investigated the effects of Ca2+ diffusion, mobile and stationary Ca2+ buffers in the cytosol, and Ca2+ handling by the endoplasmic reticulum on inositol 1,4,5-trisphosphate-induced Ca2+ wave propagation. Rapid equilibration of free and bound Ca2+ is used to describe Ca2+ sequestration by buffers in both the cytosol and endoplasmic reticulum (ER) lumen. Cytosolic Ca2+ regulation is based on a kinetic model of the inositol 1,4,5-trisphosphate (IP3) receptor of De Young and Keizer that includes activation and inhibition of the IP3 receptor Ca2+ channel in the ER membrane and SERCA Ca2+ pumps in the ER. Diffusion of Ca2+ in the cytosol and the ER and the breakdown and diffusion of IP3 are also included in our calculations. Although Ca2+ diffusion is severely limited because of buffering, when conditions are chosen just below the threshold for Ca2+ oscillations, a pulse of IP3 or Ca2+ results in a solitary trigger wave that requires diffusion of Ca2+ for its propagation. In the oscillatory regime repetitive wave trains are observed, but for this type of wave neither the wave shape nor the speed is strongly dependent on the diffusion of Ca2+. Local phase differences lead to waves that are predominately kinematic in nature, so that the wave speed (c) is related to the wavelength (lambda) and the period of the oscillations (tau) approximately by the formula c = lambda/tau. The period is determined by features that control the oscillations, including [IP3] and pump activity, which are related to recent experiments. Both solitary waves and wave trains are accompanied by a Ca2+ depletion wave in the ER lumen, similar to that observed in cortical preparations from sea urchin eggs. We explore the effect of endogenous and exogenous Ca2+ buffers on wave speed and wave shape, which can be explained in terms of three distinct effects of buffering, and show that exogenous buffers or Ca2+ dyes can have considerable influence on the amplitude and width of the waves.     

L-type Ca2+ channels in Ca2+ channelopathies

Voltage-gated L-type Ca2+ channels (LTCCs) mediate depolarization-induced Ca2+ entry in electrically excitable cells, including muscle cells, neurons, and endocrine and sensory cells. In this review we summarize the role of LTCCs for human diseases caused by genetic Ca2+ channel defects (channelopathies). LTCC dysfunction can result from structural aberrations within pore-forming alpha1 subunits causing incomplete congenital stationary night blindness, malignant hyperthermia sensitivity or hypokalemic periodic paralysis. However, studies in mice revealed that LTCC dysfunction also contributes to neurological symptoms in Ca2+ channelopathies affecting non-LTCCs, such as Ca(v)2.1 alpha1 in tottering mice. Ca2+ channelopathies provide exciting molecular tools to elucidate the contribution of different LTCC isoforms to human diseases.     

The Ca2+-induced membrane transition in mitochondria. II. Nature of the Ca2+ trigger site.

The permeability of isolated mitochondria which have undergone the Ca²⁺-induced transition can be modulated over a wide range simply by adjusting the concentration of free Ca²⁺ in the medium. The effect varies sigmoidally with respect to Ca²⁺ concentration, with an apparent K_m of 16 μm at pH 7.0. It is concluded that the trigger site (by “trigger site” we mean the site of binding of Ca²⁺ which, when Ca²⁺ is bound, will allow the transition in permeability to occur) is possibly also the site for high-affinity Ca²⁺ uptake. Added ADP, NADH and Mg²⁺ inhibit the Ca²⁺-induced permeability of mitochondria which have undergone the Ca²⁺-induced transition. Mg²⁺ and other ions, including H⁺, act like competitive inhibitors of the Ca²⁺ effect. In the presence of Ca²⁺, both neutral and charged molecules of molecular weight <1000 pass readily through the membrane. This response to Ca²⁺ is interpreted as a gating effect at the internal end of hydrophilic channels which span the inner membrane.     

ADP evokes biphasic Ca2+ influx in fura-2-loaded human platelets. Evidence for Ca2+ entry regulated by the intracellular Ca2+ store.

Stopped-flow fluorimetric studies at 37 degrees C have shown that ADP, at optimal concentrations, can evoke Ca2+ or Mn2+ influx in fura-2-loaded human platelets without measurable delay. In contrast, the release of Ca2+ from intracellular stores is delayed in onset by about 200 ms. By working at a lower temperature, 17 degrees C, we have now shown that the rise in cytosolic calcium concentration ([Ca2+]i) evoked by ADP in the presence of external Ca2+ is biphasic. The use of Mn2+ as a tracer for bivalent-cation entry indicates that both phases of the ADP-evoked response are associated with influx. The fast phase of the ADP-evoked rise in [Ca2+]i, which occurs without measurable delay at both 17 degrees C and 37 degrees C, is consistent with Ca2+ entry mediated by receptor-operated channels in the plasma membrane. The delayed phase, indicated by Mn2+ quench, is coincident with the discharge of the intracellular Ca2+ stores. Forskolin did not inhibit the fast phases of ADP-evoked rise in [Ca2+]i or Mn2+ quench, but completely abolished ADP-evoked discharge of the intracellular stores, the delayed phase of the rise in [Ca2+]i observed in the presence of external Ca2+ and the second phase of Mn2+ quench. The timing of the delayed event appears to be modulated by [Ca2+]i: the delayed phase of Mn2+ quench coincides with discharge of the intracellular stores in the absence of added Ca2+, but with the second phase of the ADP-evoked rise in [Ca2+]i in the presence of extracellular Ca2+. Similarly, blockade of the early phase of Ca2+ entry by SK&F 96365 further delays the second phase. It is suggested that a pathway for Ca2+ entry which is regulated by the intracellular Ca2+ store exists in platelets. This pathway operates alongside, and appears to be modulated by the activity of other routes for Ca2+ entry into the cytosol.     

The Ca2+-induced membrane transition in mitochondria. III. Transitional Ca2+ release.

The efflux of Ca²⁺ from mitochondria respiring at steady state, and much of uncoupler-induced Ca²⁺ efflux, is shown to be a consequence of the Ca²⁺-induced membrane transition (the Ca²⁺-induced transition is the Ca²⁺-dependent sudden increase in the nonspecific permeability of the mitochondrial inner membrane which occurs spontaneously when mitochondria are incubated under a variety of conditions (D. R. Hunter, R. A. Haworth, and J. H. Southard, 1976, J. Biol. Chem.251, 5069–5077)). Ca²⁺ release from mitochondria respiring at steady state is shown to be transitional by four criteria: (1) Ca²⁺ release is inhibited by Mg²⁺, ADP, and bovine serum albumin (BSA), all inhibitors of the transition; (2) release is selective for Ca²⁺ over Sr²⁺, a selectivity also found for the transition; (3) the time course of Ca²⁺ release is identical to the time course of the change in the mitochondrial population from the aggregated to the orthodox configuration; and (4) from kinetics, Ca²⁺ release from individual mitochondria is shown to occur suddenly, following a lag period during which no release occurs. Ca²⁺ release induced by uncoupler is shown to be mostly by a transitional mechanism, as judged by four criteria: (1) release of Ca²⁺ is ruthenium red-insensitive and is an order of magnitude faster than Sr²⁺ release which is ruthenium red-sensitive; (2) release of Ca²⁺ is strongly inhibited by keeping the mitochondrial NAD⁺ reduced; (3) the kinetics of Ca²⁺ release indicates a transitional release mechanism; and (4) uncoupler addition triggers the aggregated to orthodox configurational transition which, at higher levels of Ca²⁺ uptake, occurs in the whole mitochondrial population at a rate equal to the rate of Ca²⁺ release. Na²⁺-induced Ca²⁺ release was not accompanied by a configurational change; we therefore conclude that it is not mediated by the Ca²⁺-induced transition.     

Increased intracellular Ca2+ induces Ca2+ influx in human T lymphocytes.

One current hypothesis for the initiation of Ca2+ entry into nonelectrically excitable cells proposes that Ca2+ entry is linked to the state of filling of intracellular Ca2+ stores. In the human T lymphocyte cell line Jurkat, stimulation of the antigen receptor leads to release of Ca2+ from internal stores and influx of extracellular Ca2+. Similarly, treatment of Jurkat cells with the tumor promoter thapsigargin induced release of Ca2+ from internal stores and also resulted in influx of extracellular Ca2+. Initiation of Ca2+ entry by thapsigargin was blocked by chelation of Ca2+ released from the internal storage pool. The Ca2+ entry pathway also could be initiated by an increase in the intracellular concentration of Ca2+ after photolysis of the Ca(2+)-cage, nitr-5. Thus, three separate treatments that caused an increase in the intracellular concentration of Ca2+ initiated Ca2+ influx in Jurkat cells. In all cases, Ca(2+)-initiated Ca2+ influx was blocked by treatment with any of three phenothiazines or W-7, suggesting that it is mediated by calmodulin. These data suggest that release of Ca2+ from internal stores is not linked capacitatively to Ca2+ entry but that initiation is linked instead by Ca2+ itself, perhaps via calmodulin.     

Ca2+ -induced Ca2+ release through localized Ca2+ uncaging in smooth muscle

Ca(2+)-induced Ca(2+) release (CICR) from the sarcoplasmic reticulum (SR) occurs in smooth muscle as spontaneous SR Ca(2+) release or Ca(2+) sparks and, in some spiking tissues, as Ca(2+) release that is triggered by the activation of sarcolemmal Ca(2+) channels. Both processes display spatial localization in that release occurs at a higher frequency at specific subcellular regions. We have used two-photon flash photolysis (TPFP) of caged Ca(2+) (DMNP-EDTA) in Fluo-4-loaded urinary bladder smooth muscle cells to determine the extent to which spatially localized increases in Ca(2+) activate SR release and to further understand the molecular and biophysical processes underlying CICR. TPFP resulted in localized Ca(2+) release in the form of Ca(2+) sparks and Ca(2+) waves that were distinguishable from increases in Ca(2+) associated with Ca(2+) uncaging, unequivocally demonstrating that Ca(2+) release occurs subsequent to a localized rise in [Ca(2+)](i). TPFP-triggered Ca(2+) release was not constrained to a few discharge regions but could be activated at all areas of the cell, with release usually occurring at or within several microns of the site of photolysis. As expected, the process of CICR was dominated by ryanodine receptor (RYR) activity, as ryanodine abolished individual Ca(2+) sparks and evoked release with different threshold and kinetics in FKBP12.6-null cells. However, TPFP CICR was not completely inhibited by ryanodine; Ca(2+) release with distinct kinetic features occurred with a higher TPFP threshold in the presence of ryanodine. This high threshold release was blocked by xestospongin C, and the pharmacological sensitivity and kinetics were consistent with CICR release at high local [Ca(2+)](i) through inositol trisphosphate (InsP(3)) receptors (InsP(3)Rs). We conclude that CICR activated by localized Ca(2+) release bears essential similarities to those observed by the activation of I(Ca) (i.e., major dependence on the type 2 RYR), that the release is not spatially constrained to a few specific subcellular regions, and that Ca(2+) release through InsP(3)R can occur at high local [Ca(2+)](i).     

Ca(2+)-induced Ca2+ release amplifies the Ca2+ response elicited by inositol trisphosphate in macrophages.

We have studied the rise in intracellular calcium concentration ([Ca2+]i) elicited in macrophages stimulated by platelet-activating factor (PAF) by using fura-2 measurements in individual cells. The [Ca2+]i increase begins with a massive and rapid release of Ca2+ from intracellular stores. We have examined the mechanism of this Ca2+ release, which has been generally assumed to be triggered by inositol trisphosphate (IP3). First, we confirmed that IP3 plays an important role in the initiation of the PAF-induced [Ca2+]i rise. The arguments are 1) an increase in IP3 concentration is observed after PAF stimulation; 2) injection of IP3 mimics the response to PAF; and 3) after introduction of heparin in the cell with a patch-clamp electrode, the PAF response is abolished. Second, we investigated the possibility of an involvement of Ca(2+)-induced Ca2+ release (CICR) in the development of the Ca2+ response. Ionomycin was found to elicit a massive Ca2+ response that was inhibited by ruthenium red or octanol and potentiated by caffeine. The PAF response was also inhibited by ruthenium red or octanol and potentiated by caffeine, suggesting that CICR plays a physiological role in these cells. Because our results indicate that in this preparation IP3 production is not sensitive to [Ca2+]i, CICR appears as a primary mechanism of positive feedback in the Ca2+ response. Taken together, the results suggest that the response to PAF involves an IP3-induced [Ca2+]i rise followed by CICR.     

Ca2+ regulation in the Na+/Ca2+ exchanger involves two markedly different Ca2+ sensors

The plasma membrane Na+/Ca2+ exchanger (NCX) is almost certainly the major Ca2+ extrusion mechanism in cardiac myocytes. Binding of Na+ and Ca2+ ions to its large cytosolic loop regulates ion transport of the exchanger. We determined the solution structures of two Ca2+ binding domains (CBD1 and CBD2) that, together with an alpha-catenin-like domain (CLD), form the regulatory exchanger loop. CBD1 and CBD2 are very similar in the Ca2+ bound state and describe the Calx-beta motif. Strikingly, in the absence of Ca2+, the upper half of CBD1 unfolds while CBD2 maintains its structural integrity. Together with a 7-fold higher affinity for Ca2+, this suggests that CBD1 is the primary Ca2+ sensor. Specific point mutations in either domain largely allow the interchange of their functionality and uncover the mechanism underlying Ca2+ sensing in NCX.     

Ca2+ waves in astrocytes.

The glial cell is the most numerous cell type in the central nervous system and is believed to play an important role in guiding brain development and in supporting adult brain function. One type of glial cell, the astrocyte also may be an integral computational element in the brain since it undergoes neurotransmitter-triggered signalling. Here we review the role of the astrocyte in the central nervous system, emphasizing receptor-mediated Ca2+ physiology. One focus is the recent discovery that the neurotransmitter glutamate induces a variety of intracellular Ca2+ changes in astrocytes. Simple Ca2+ spikes or intracellular Ca2+ oscillations often appear spatially uniform. However, in many instances, the Ca2+ rise has a significant spatial dimension, beginning in one part of the cell it spreads through the rest of the cell in the form of a wave. With high enough agonist concentration an astrocyte syncitium supports intercellular waves which propagate from cell to cell over relatively long distances. We present results of experiments using more specific pharmacological glutamate receptor agonists. In addition to describing the intercellular Ca2+ wave we present evidence for another form of intercellular signalling. Some possible functions of a long-range glial signalling system are also discussed.     

The slowing of Ca2+ signals by Ca2+ indicators in cardiac muscle.

The use of high-affinity fluorescent probes for monitoring intracellular free Ca2+ in cardiac muscle is now widespread. We have investigated the consequences of introducing intracellular buffers with the properties of Fura-2 or Indo-1 on the action potential, Ca2+ transient and contractile activity of the myocardium. Our theoretical results suggest that, at the high intracellular concentrations of these fluorescent probes used on occasion to improve the signal-to-noise ratio of the emitted fluorescence, modulation of action potential profile and attenuation of the amplitudes of the Ca2+ transient and contraction can occur, together with subtle changes in the kinetics of these events.     

Ca2+-controlled conformational states of the Ca2+ transport enzyme of sarcoplasmic reticulum.

The fluorescent reagent, S-mercuric N-dansyl-cysteine, reacts specifically with thiols of the purified Ca2+-ATPase of the sarcoplasmic reticulum, producing an increase of fluorescence of fluorescence intensity at 500 nm (lambda ex = 335 nm). The reaction is stoichiometric, and the increase of the fluorescence intensity is proportional to the number of blocked thiols. Twelve reactive thiols per 10(5) daltons of ATPase peptide fall into roughly three classes. Blocking of the most reactive thiol entails little inhibition of enzyme activity. Blocking of the five thiols reacting next (intermediate class) results in almost complete inhibition of both phosphorylated intermediate formation and ATP hydrolysis. The second order rate constants of the reaction of thiols have been determined by stopped flow studies. The most reactive thiol and the six least reactive thiols can each be treated as a single class with respect to the rate constant; five thiols of intermediate reactivity appear to have different rate constants (k2, k3, ..k6). Of these constants, k1, corresponding to the most reactive thiol, does not change with [Ca2+]. Upon increasing [Ca2+] from 10(-9) to 10(-5) M, k2 increase and k7-12 decreases; the changes roughly parallel the activation of ATPase activity and the Ca2+ binding to the high affinity alpha sites (Ikemoto, N. (1975) J. Biol. Chem. 250, 7219-7224). Upon further increase of [Ca2+] k2 decreases and k7-12 increase, in parallel with the inhibition of ATPase activity and with the Ca2+ binding to the low affinity gamma sites.     

Feedback inhibition of Ca2+ release by Ca2+ is the underlying mechanism of agonist-evoked intracellular Ca2+ oscillations in pancreatic acinar cells.

Oscillations of free intracellular Ca2+ concentration ([Ca2+]i) are known to occur in many cell types during physiological cell signaling. To identify the basis for the oscillations, we measured both [Ca2+]i and extracellular Ca2+ concentration ([Ca2+]o) to follow the fate of Ca2+ during stimulation of [Ca2+]i oscillations in pancreatic acinar cells. [Ca2+]i oscillations were initiated by either t-butyloxycarbonyl-Tyr(SO3)-Nle-Gly-Tyr-Nle-Asp-2-phenylethyl ester (CCK-J), which mobilized Ca2+ from the inositol 1,4,5-trisphosphate (IP3)-insensitive pool, or low concentration of cholecystokinin octapeptide (CCK-OP), which mobilized Ca2+ from the IP3-sensitive internal pool. Little Ca2+ efflux occurred during the oscillations triggered by CCK-J or CCK-OP in spite of a large average increase in [Ca2+]i. When internal store Ca2+ pumps were inhibited with thapsigargin (Tg) during [Ca2+]i oscillations, a rapid Ca2+ efflux occurred similar to that measured in intensely stimulated, nonoscillatory cells. Tg also stimulated 45Ca efflux from internal pools of cells stimulated with CCK-J or a low concentration of CCK-OP. Hence, a large fraction of the Ca2+ released during each spike is reincorporated by the internal store Ca2+ pumps. Surprisingly, when the increase in [Ca2+]i during stimulation of oscillations was prevented by loading the cells with 1,2-bis(2-aminophenoxy) ethane-N,N,N',N'-tetraacetic acid, a persistent activation of Ca2+ release and Ca2+ efflux occurred. This was reflected as a persistent increase in [Ca2+]o in cells suspended at low [Ca2+]o or persistent efflux of 45Ca from internal stores of cells maintained at high [Ca2+]o. Since agonist-stimulated Ca2+ release evidently remains activated when [Ca2+]i is highly buffered, the primary mechanism determining Ca2+ oscillations must include an inhibition of Ca2+ release by [Ca2+]i. Loading the cells with 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid had no apparent effect on the levels or kinetics of IP3 formation in agonist-stimulated cells. This suggests that [Ca2+]i regulated the oscillation by inhibition of Ca2+ release independent of its possible effects on cellular levels of IP3.