Cell, Vol. 63, 1025-1032,

November

30, 1990, Copyright

0 1990 by Cell Press

Receptor-Activated Cytoplasmic Ca*+ Spiking Mediated by lnositol Trisphosphate Is Due to Ca*+-Induced Ca*+ Release Makoto Wakui: Yuri V. Osipchuk,t and Ole H. Petersen MRC Secretory Control Research Group The Physiological Laboratory University of Liverpool Liverpool, L69 36X England

Summary Receptor-mediated inositol 1,4,5-trisphoephate (Ins(1,4,5)P3) generation evokes fluctuations in the cytoplasmic Ca2+ concentration ([Ca2+Ji). Intracellular Ca2+ infusion into single mouse pancreatic acinar cells mimicks the effect of external acetylcholine (ACh) or internal lns(1,4,5)P3 application by evoking repetitive Ce2+ release monitored by Ca2+-activated Cl- current. Intracellular infusion of the lne(1,4,5)P3 receptor antagonist heparin fails to inhibit Ca2+ splking caused by Ca2+ infusion, but blocks ACh- and Ins(l,4,5)P3-evoked Ca2+ oscillations. Caffeine (1 mM), a potentiator of Ca2+-induced Ca2+ release, evokes Ca2+ spiking during subthreshold intracellular Ca2+ infusion. These meults indicate that ACh-evoked Ca2+ oscillations an due to pulses of Ca2+ release through a caffeine-sensitive channel triggered by a small steady Ins(l,4,5)P3-evoked Ca2+ flow. Introduction Many hormones and neurotransmitters evoke an increase in the cytoplasmic Ca*+ concentration ([Ca2+]i) of their target cells. Such Ca*+ signals are important for the control of many different cellular functions, including metabolism, secretion, contraction, and growth (Berridge and Irvine, 1964, 1969; Berridge, 1967). Cytoplasmic Ca*+ signals can be initiated by the opening of voltage-sensitive Ca*+ channels when receptor activation evokes depolarization of the cell membrane (Jan and Jan, 1969) or by generation of inositol 1,4,5-trisphosphate (lns(1,4,5)Ps), a messenger that releases Ca*+ from endoplasmic reticulum (ER) Ca2+ stores (Streb et al., 1963; Berridge and Irvine, 1964, 1969) and also controls voltage-insensitive Ca*+ channels in the plasma membrane, alone (Gardner, 1969) or together with inositol (1,3,4,5)-tetrakisphosphate (lns(1,3,45)P.,) (Petersen, 1989). Draining Ca*+ from the ER can cause proteins such as endoplasmin (GRP 94) to be rapidly secreted (Booth and Koch, 1989) and the release of Ca*+ from the ER may therefore not only provide a cytosolic signal, but also a signal in the ER lumen. lns(1,4,5)Ps production can be induced by a vari’ Present address: Department of Physiology, Tohoku School of Medicine, Sendai 960, Japan. + Present address: A. A. Sogomoletz Institute of Physiology, of Sciences, Kiev-24, 252601, USSR.

University, Academy

ety of agonists interacting with their specific cell surface receptors that stimulate phosphoinositidase C (phospholipase C) (PIC) via functionally distinct guanine nucleotide binding (G) proteins (Schnefel et al., 1988; Ashkenazi et al., 1989). PIC hydrolyzes phosphatidylinositol4Bbiphos phate (PIP*) thus generating lns(1,4,5)Ps as well as 1,2-diacylglycerol (Berridge, 1987). Activation of receptors linked to lns(1,4,5)P3 formation generally evokes oscillating cytoplasmic Ca*+ signals (spikes) when submaximal agonist concentrations are used (Berridge and Irvine, 1989; Goldbeter et al., 1990). The frequency of such cytoplasmic Ca*+ spikes is often dependent on agonist concentration and may provide the basis for particularly precise cellular control mechanisms (Berridge and Irvine, 1989). Ca*+ oscillations in the ER luminal compartment may be of importance in controlling vesiculation and consequent packaging of secretory and resident proteins (Sambrook, 1990). It has been suggested that oscillating or spiking Ca*+ signals can be explained by fluctuating Ins(l,4,5)P3 levels (Woods et al., 1987; Meyer and Stryer, 1988), but intracellular infusion of Ins(l,4,5)P3 (Wakui et al., 1989; Swann et al., 1989) or the nonhydrolyzable analog Ins(l,4,5)PS3 (Wakui et al., 1989) can generate repetitive Ca*+ spikes. Ferris et al. (1990) have suggested that a constant lns(1,4,5)P3 concentration can generate cytosolic Caa oscillations due to fluctuating ATP levels, based on data showing that low ATP concentrations increase Ins(l,4,5)P+voked Ca2+ flux through reconstituted lns(1,4,5)P3 receptors, whereas high ATP concentrations inhibit this effect. A two-pool quantitative model for signal-induced Ca*+ oscillations has recently been proposed that also predicts the occurrence of periodic cytoplasmic Ca2+ spikes in the absence of lns(1,4,5)P3 oscillations (Goldbeter et al., 1990). In this model a steady lns(1,4,5)P3 level causes a steady Ca*+ flow into the cytosol that evokes emptying of separate intracellular Ca*+ pools via Ca*+-induced Ca2+ release. The emptying of these Ins(l,4,5)P3-insensitive pools gives rise to a Ca*+ spike, and the time taken to refill the pools determines the interspike interval (Goldbeter et al., 1990). According to this model, cytoplasmic Ca*+ spikes should be evoked by intracellular Ca*+ infusion. In experiments on single internally perfused pancreatic acinar cells such a phenomenon has been directly demonstrated (Osipchuk et al., 1990). The finding that a small, steady intracellular Ca2+ infusion evokes repetitive cytoplasmic Ca*+ spikes thus mimicking the effect of intracellular lns(1,4,5)P3 infusion (Osipchuk et al., 1990) may however, also be interpreted differently. PIC can be activated by Ca*+ (Downes and Michel, 1981; Taylor et al., 1986; Meyer and Stryer, 1988) and intracellular Ca*+ infusion could therefore result in lns(1,4,5)P, generation. Two different channels that mediate Ca*+ release from intracellular nonmitochondrial stores into the cytosol are known. The Ins(l,4,5)P3-activated Ca*+ channel has been isolated and sequenced (Furuichi et al., 1989; Maeda et al., 1990) reconstituted and functionally investigated in

Cell 1026

lipid bilayers (Ehrlich and Watras, 1988; Ferris et al., 1989), and localized to the ER (Ross et al., 1989). The Ca2+-induced Ca2+ release channel has been isolated from muscle sarcoplasmic reticulum (SR) (Lai et al., 1988), sequenced (Takeshima et al., 1989), and studied functionally by reconstitution in lipid bilayers (Lai et al., 1988) or after expression in Chinese hamster ovary cells (Penner et al., 1989). The Ins(l,4,5)P3-activated Ca2+ channel can be blocked specifically by heparin (Ehrlich and Watras, 1988; Ferris et al., 1989), a substance that has no effect on the SR Ca2+ channel (Erhlich and Watras, 1988), which can be activated by submicromolar to micromolar Ca2+ concentrations, specifically from the cytoplasmic side (Meissner et al., 1988; Lai et al., 1988). Caffeine, a well known potentiator of Ca2+-induced Ca2+ release (Endo, 1977), can open SR CP+ release channels at normal cytoplasmic Ca2+ concentrations (Penner et al., 1989), but has no effect on Ins(l,4,5)P3-activated Ca2+ channels (Ehrlich and Watras, 1988). We now show that intracellular heparin infusion blocks cytoplasmic Ca2+ spikes evoked by external acetylcholine (ACh) or internal lns(1,4,5)P3 application but fails to inhibit Ca2+ spiking induced by intracellular Ca2+ infusion. Caffeine potentiates the Ca2+ induction of Ca2+ oscillations and can also, under certain circumstances, itself evoke such oscillations. These results show that continued opening of Ins(l,4,5)P3-activated Ca2+ channels is needed to sustain repetitive cytoplasmic Ca2+ spiking evoked by agonist-receptor interaction, but they also demonstrate that Ca2+-induced Ca2+ spikes can occur independently of lns(1,4,5)P3. It would appear that Ins(1,4,5)P3 evokes a small, steady outflow of Ca2+ into the cytosol from a heparin-sensitive channel generating repetitive Ca2+ spikes from caffeine-sensitive channels. Results Acetylcholine-Evoked Ca*+ Spiking Is Abolished by Intracellular Heparin Infusion We monitored changes in [Ca2’]i by patch-clamp wholecell recording of Ca2+-dependent Cl- current in single collagenase-isolated mouse pancreatic acinar cells (Wakui et al., 1989). The validity of this approach has been confirmed recently in a study where [Ca2+]i in single pancreatic acinar cells was assessed by simultaneous microfluorimetry (fura-2) and patch-clamp recording of Ca2+dependent Cl- current (Osipchuk et al., 1990). Figure 1 shows ACh-evoked repetitive increases in Ca2+-dependent Cl- current, which we shall refer to as cytoplasmic Ca2+ spikes, confirming previous results on the same cells (Wakui et al., 1989; Osipchuk et al., 1990). In the experiments presented here, heparin is introduced into the tip of the patch-clamp pipette via the application of pressure from a fine tube inserted into the pipette. There is direct continuity between the pipette solution and the intracellular compartment, and heparin infused into the pipette solution can therefore diffuse into the cell (Figure 1). In the experiment illustrated in Figure 1, a reduction in the Ca2+ spike frequency is already observed within

about 1 min of applying pressure to the heparin-containing infusion tube. Thereafter, the interspike interval gradually increases and after about 5 min no further Ca2+ spikes are observed. In every one of the four experiments of this type carried out, intracellular heparin infusion abolished ACh-evoked Ca2+ spikes. ACh-evoked Ca2+ spikes in these cells do not stop spontaneously and there is also no tendency for a slow-down in frequency (see Figure 1 in Wakui et al., 1989). Ca2+ Spikes Evoked by Internal lns(1,4,5)P3 Application Are Blocked by Intracellular Heparin Infusion It has previously been shown that intracellular application of lns(1,4,5)P3 evokes repetitive Ca2+ spikes in pancreatic acinar cells and that the individual spikes are generally of a shorter duration than those evoked by ACh stimulation (Wakui et al., 1989; Osipchuk et al., 1990). Figure 2 shows repetitive Ca2+ spikes evoked by pressure application of 10 VM lns(1,4,5)P3 via an intrapipette infusion tube. This concentration (10 PM) of lns(1,4,5)P3 is not evoking a maximal effect since it has previously been shown that 100 PM lns(1,4,5)Pa can induce a sustained Ca2+ signal (Wakui et al., 1989). When external ACh (50 nM) is added on top of the internal lns(1,4,5)P3 stimulus, a quasisustained Ca2+ signal is obtained (Figure 2a). Discontinuation of ACh stimulation is followed by a short silent phase before the repetitive Ca2+ spiking evoked by lns(1,4,5)P3 is resumed (Figure 2a). Twelve experiments of the type shown in Figure 2a were carried out giving similar results, although there was considerable cell-to-cell variation in spike frequency (2-10 per min). Figure 2b shows the result from an experiment in which intrapipette infusion of heparin abolishes lns(1,4,5)P3 (10 FM)-evoked Ca2+ spikes. Subsequent ACh (50 nM) applications also fail to evoke any responses (Figure 2b). Four experiments of this type were carried out, all giving similar results. In another series of experiments, heparin (200 pglml) was present in the pipette solution from the start of the experiment, and intrapipette infusion of lns(1,4,5)P3 (10 PM) was initiated 2-3 min after the establishment of the whole-cell recording configuration. In these cases, lns(1,4,5)P3 did not evoke any effect (n = 5). Intracellular Ca2+ Infusion Can Evoke Cytoplasmic Ca2+ Spikes and These Are Not Blocked by Internal Heparin Application Figure 3a shows that pressure application of Ca2+ solution into the tip of the patch-clamp pipette evokes a graded response consisting of discrete pulses of Ca2+-activated Cl- current. The maximal spike frequency in these types of experiments was 7-10 per min. At the highest pressure, sufficient Ca2+ enters the cell to ensure a steady elevation of [Ca2+]i. The trace shown in Figure 3a is typical of seven out of ten experiments. In the remaining three experiments, Ca2+ infusion failed to evoke any effect, probably because of insufficient flow through the fine intrapipette infusion tube. When ACh (50 nM) is applied during Ca2+-induced Ca2+ spiking, a quasisustained os-

Heparin -

” Pressure injection

’

Figure

50nY

2OOp0/ml

ACh

1. ACh-Evoked

hepartn r

Pulses

(Spikes)

of Cap+-Dependent

Cl- Current

Are Blocked

ACh

by lntrapipette

1

Perfusion

with a Heparin-Containing

Solution

The cartoon illustrates the recording configuration. pans-membrane currents are recorded under voltage-clamp conditions at -30 and 0 mV. At 0 mV there are only very small ACh-evoked current changes as the Cl- equilibrium potential is close to 0. At -30 mV there is a large electrical gradient favoring Cl- efflux seen as inward current (downward deflections). In this particular experiment there is a slight change in the apparent ACh null potential, most likely because of a slow and modest increase in the intracellular Cl- concentration as the final equilibration between the pipette and intracellular solutions occurs.

cillating Ca*+ signal is generated that continues until ACh is removed (Figure 3b) (n = 4). When the pipette solution contained heparin (200 pglml) from the start of the experiment, internal Ca*+ infusion still evoked Ca*+ spikes in all the three cells where this protocol was tried (Figures 3c, and 3d), although the ACh effect had been blocked. Caffeine Stimulates CaZ+ Spike Generation at a Low Concentration (1 mm) but Inhibits It at a High Concentration (20 mM) In the experiment illustrated in Figure 4a, a slow Ca*+ infusion sustained by low pressure fails by itself to evoke Ca*+ release, but additional application of caffeine (1 mM) from the outside causes discrete and repetitive Ca*+ spikes (n = 5). It has been shown previously that caffeine can induce Ca*+ oscillations when applied in the presence of a subthreshold dose of ACh (Osipchuk et al., 1990) but in the absence of stimulation, caffeine (l-20 mfvt) never had an effect (n = 5). However, when the cells were internally perfused with a solution not containing the normally used low concentration (0.25 mM) of the Ca*+ chelator EGTA, caffeine (1 mM) could repeatedly and reversibly evoke Ca*+ spikes. In the experiment shown in Figure 4b, a high concentration of caffeine (20 mM) is ini-

200Wml

I

heperh

lowhwP3

b , , , , , , , , ,, , , , , , Figure

2. lntrapipette

Infusion

of lns(1,4,5)Ps

(a) illustrates an experiment in which of heparin blocks Ins(l,4,5)P,-evoked

(InsPs)

tially applied, but has no effect until just after it has been removed, when a few Ca*+ spikes appear. Thereafter, 1 mM caffeine reversibly evokes repetitive Ca*+ spikes. Later in the experiment, 20 mM caffeine immediately stops the repetitive Ca*+ pulses, but after its removal a few Ca2+ spikes reappear. In some cells the initial priming with 20 mM caffeine was needed in order to obtain subsequently Ca*+ spike responses to stimulation with 1 mM caffeine. In other cells, 1 mM caffeine itself, without any previous priming with high caffeine doses, evoked effects of the type shown in Figure 4b. In a series of ten experiments without EGTA in the pipette solution there were only three cells that did not respond to caffeine alone. The effect of caffeine is unlikely to have anything to do with a possible increase in the level of cyclic AMP caused by inhibition of phosphodiesterase since 20 mfvl theophylline failed to mimick the action of caffeine (n = 5). The inhibitory effect of the high caffeine concentration was further investigated in cells that were stimulated to generate Ca*+ spikes by internal lns(1,4,5) PSa application. We have shown previously that this nonhydrolysable inositol trisphosphate analog mimic& the action of Ins(1,4,5)P3 and that 30 PM Ins (1,4,5)PSa quantitatively give an effect very similar to that evoked by 10 t.&l Ins (1,4,5)P3 (Wakui et al., 1999). (The Ins (1,4,5)PS3 we used was a mix-

Evokes

5OnM ACh

, Repetitive

Inward

Gas+-Dependent

7

Cl- Current

the additional external ACh application evokes a quasisustained Ca2+ spikes. Heparin also blocks the ACh response.

signal.

Pulses (b) shows

that intrapipette

infusion

Cell 1028

roqu 20

40

60

hmHgl

100

I

150 Lv\

c

I

ry-f-f7-

a 80

Figure 3. lntrapipette Ca*+ Repetitive Ca*+ Spikes

ca2*

20 7r/mV

‘---Aomv

1 OOMM Ca* +

2w;nlz I’ ’

!’

1

I r

--&

I

‘1

I

ioopd

Ca2+, 2OmmM

I

1 I

7

111 1’1’

‘1 m-&

caffeine 20

I I 1

2’3‘3pA

(mM1 0mV

-/

-30mV

,

1

I

,

(a) shows graded response to varying Ca*+ infusion rates (varied by changing the pressure applied to the infusion tube). With increasing pressure, larger and more frequent Gas+ spikes are seen. Finally, at the highest pressure chosen, a sustained Ca2+ signal is seen following a few spikes. (b) shows that external ACh applied during the internal Ca2+ infusion evokes a quasisustained response. (c) shows repetitive discrete pulses evoked by Ca*+ infusion into a cell that was internally perfused with heparin-containing solution from the beginning of the experiment. Later (data not shown) this cell failed to respond to ACh (50 nM). (d) illustrates an experiment in which the pipette was filled with heparin from the beginning and ACh failed to elicit a response, but subsequent Ca*+ infusion into the cell evoked several Ca*+ spikes that in this case were followed by a sustained response.

Figure 4. Caffeine Ca2+ Spikes

-__ 1 mM caffaha

Evokes

Ca2+ lonophores Can Induce Repetitive CaZ+ Spikes The effects of internal Ca2+ infusion can be mimicked to some extent by external application of the Ca2+ ionophore ionomycin. Figure 8a is typical of four out of eight experiments. In two of the remaining four experiments, where 10 nM ionomycin had no effect (Figure 8b), increasing the external Ca2+ concentration from l-20 mM evoked a short train of Ca2+ spikes. After an interval, a sustained rise in the Ca2+-dependent Cl- current was then observed. A higher ionomycin concentration (100 nM) evoked a sustained rise in [Ca2+]i after a few Ca2+ spikes (Figure 8c) (n = 7). With 5 mM EGTA in the pipette solution, ionomycin (100 nM) had no effect (n = 4). In all cases of ionomycinevoked Ca2+ oscillations (10 nM), the pulses of Ca2+dependent Cl- current stopped after a few minutes (Figure 8a), and in two such cases where ACh (50 nM) was then applied, no response occurred. The Ca2+ ionophore A23187 (0.1-l vM) induced similar effects to those obtained with lo-100 nM ionomycin (n = 7).

ture of the El and the L isomers, but it is only the D isomer that is active. The 30 PM concentration of D,L-lns(1,4,5)P& therefore corresponds to 15 WM D-Ins (1,4,5)PS3 (Wakui et al., 1989)) Figure 5 shows that 20 mM caffeine rapidly blocks CP+ spiking. This acute inhibitory effect is quickly reversible (n = 8). In fact, the frequency of Ca2+ spiking always transiently increased immediately after caffeine removal. This probably reflects the stimulatory action of low caffeine concentrations during the period when caffeine is washed out. When internal stimulation with Ins(l,4,5)PS3 is initiated during a period of exposure to 20 mM caffeine, no effect is observed until caffeine has been removed from the bath (n = 5) (Figure 5b). Figure 5c shows that external application of the cell-permeable dibutyryl cyclic AMP (10 mM) causes a small and reversible increase in the baseline current, but fails to prevent the acute inhibitory effect of caffeine as well as the transient acceleration of spiking immediately after caffeine removal (n = 8).

,

2OOpA

\‘I--

Infusion

1

20

Applied

Externally

Evokes

In (a), a low pressure applied to the Ca*+containing infusion tube causes an insufficient Ca*+ stimulus to elicit a response by itself, but caffeine applied during the Ca2+ infusion evokes Gas+ spikes. This effect is reversible and can be repeated. (b) shows the result of an experiment in which there is no intrapipette infusion tube and where the pipette solution exceptionally did not contain the Ca2+ chelator EGlA. Initially, 20 mM caffeine has no effect until after its removal, when a few Ca*+ spikes occurred. At a concentration of 1 mM, caffeine, thereafter reversible, evokes repetitive Ca*+ spikes. At the end of this experiment, 20 mM caffeine is seen to inhibit acutely and reversibly the Caa+ spike generation.

7$9piaplasmic

Ca2+ Oscillations

Figure 5. lns(1,4,5)PS3 (Ins PS,)-Evoked Cytoplasmic Gas+ Spikes Are Inhibited Acutely and Reversibly by 2g,mM Caffeine Dibutyryl cyclic AMP (dbcAMP) (10 mt) ap plied externally fails to influence the action of caffeine, but does evoke a very slight and fully reversible increase in the baseline current.

r-

30IlM Ins PS, I

10 mY dbcAMP 20 ml csffelne

I

1

ZOOpA

30s

Figure 6. The Ca 2+ lonophore Evokes Cytoplasmic Ca2+ Spikes

a

1

low

(a) ionomycin (10 nM) applied externally evokes Ca2+-dependent Cl- current pulses for a few minutes. (b) in some experiments there is no response to ionomycin until the external Ca2+ concentration is increased from l-20 mM. (c) high concentration of ionomycin (100 nM) evokes a sustained response, after a few initial spikes, and this sustained response is reduced when external Ca2+ is removed.

bnomych 2onw

cac12

b lOOnhI

lonomycin

bnomych 2+ -2mv

C

&

2OOpA

mib\ -3OmV

Discussion The results presented here provide direct evidence for the hypothesis proposed by Berridge (Berridge, 1988; Berridge and Irvine, 1989; Goldbeter et al., 1990) that receptor-activated cytoplasmic Ca*+ oscillations in electrically nonexcitable cells are due to a Ca2+-induced Ca2+ release mechanism. The schematic diagram shown in Figure 7 illustrates the main conclusions that can be drawn from our experiments. The finding that the repetitive Ca2+ spikes induced by infusion of Ca*+ in the patch pipette are not blocked by the Ins(1,4,5)P3 receptor antagonist heparin shows that Ca*+-induced Ca2+ release does not depend on lns(1,4,5)P3 or the Ins(i,4,5)P3-activated Ca2+ channel. Since both the ACh- and Ins(l,4,5)P3-evoked Ca2+ spiking can be abolished by heparin, it is clear that sustained receptor-activated Ca2+ oscillations depend on continued binding of lns(1,4,5)P3 to its receptors. Since furthermore the Ins(l,4,5)P3-evoked Ca2+ spiking can be mimicked by intracellular Ca*+ infusion, the conclusion must be that a constant lns(1,4,5)P3 level causes a steady Ca*+ release and that it is this steady outflow of Ca2+ into

63 ,* Caffeine

Figure 7. Simplified cerning the Actions

Diagram Illustrating of ACh, lns(1,4,5)P3,

the Main Conclusions and Ca2+

Con-

Abbreviations: mR = muscarinic receptor, G = GTP-binding protein, PIC = phosphoinositidase C (phospholipase C), IP, = lns(1,4,5)P3. Since the ACh-evoked Ca2+ signals studied in this work are derived from intracellular stores, this diagram does not depict entry and exit pathways for Ca2+ across the plasma membrane. The different Ca2+ accumulation mechanisms in the two Ca* stores were discovered by Thevenod et al. (1999).

Cell 1030

the cytosol that is responsible for the repetitive Ca*+ spikes. It has been demonstrated previously that heparin can completely inhibit lns(1,4,5)P3 binding to its purified receptor as well as abolish Ca*+ flux through this protein reconstituted in lipid vesicles (Ferris et al., 1989). Our conclusion, derived from experimental evidence, that receptor-induced Ca*+ oscillations are mediated by a constant flow of Ca*+ through Ins(l,4,5)Ps-activated Ca2+ channels is in complete agreement with the quantitative minimal model developed by Goldbeter et al. (1990). In general, the source of this constant Ins(l,4,5)P3-generated Ca2+ flow into the c,tosol could be the Ins(l,4,5)P3-sensitive ER Ca*+ pool, as illustrated in Figure 7, or it could be the extracellular Ca*+ entering the cytosol via Ins(1 ,4,5)P3or lns(1,4,5)P3/lns(l,3,4,5)P4-sensitive Ca*+ pathways in the plasma membrane (Kuno and Gardner, 1987; Morris et al., 1987; Penner et al., 1988; Irvine, 1990; Berridge, 1990). In histamine-stimulated endothelial cells there is a constant opening of the plasma membrane Ca*+ uptake pathway when [Ca*+]i is spiking (Jacob, 1990), showing that external Ca*+ can be the source of a steady flow into the cytosol, triggering discrete pulses of Ca*+-induced Ca*+ release. In the experiments presented here the Ins(l,4,5)Pa-evoked steady flow of Ca*+ into the cytosol is almost exclusively derived from internal stores since it has been shown that in the internally perfused single mouse pancreatic acinar cells, Ca2+ oscillations evoked by ACh, lns(1,4,5)P3, or Ins(l,4,5)PSs are completely unaffected by removal and readmission of external Ca*+ during the first 7-8 min after the start of stimulation (Wakui et al., 1989). Since the Ca*+-induced Ca*+ release is not blocked by heparin it must occur through channels other than those activated by Ins(1 ,4,5)P3. Ca*+-activated Ca*+ release channels from SR are known not to be blocked by heparin (Ehrlich and Watras, 1988) and since caffeine, an activator of the Ca*+-induced Ca*+ release channel (Penner et al., 1989) can mimick the effect of intracellular Ca2+ infusion, in our experiments under conditions of minimal intracellular Ca*+ buffering (there was no EGTA in the internal perfusion solution), it is likely that the Ca*+ release giving rise to the cytoplasmic spikes occurs through channels of a type somewhat similar to those characterized in muscle SR (Lai et al., 1988). Schmid et al. (1990) have recently discovered caffeineactivated Ca*+ channels in ER vesicles from pancreatic acinar cells and it seems likely that the caffeine- and Ca2+-evoked repetitive Ca*+ spikes observed in the present work are due to the opening of such channels. Caffeine (1 mM) alone never evokes any effect under our standard intracellular Ca2+ buffering condition (0.25 mM EGTA) (Osipchuk et al., 1990), but this concentration of the drug clearly induces Ca2+ oscillations when applied during subthreshold intracellular Ca*+ infusion (Figure 4a). This is further evidence indicating that the caffeine-sensitive channel is the one activated by Ca2+. Figure 7 suggests that the caffeine-sensitive Ca2+-induced Ca*+ release channel is only present in Ins(l,4,5)Ps-insensitive Ca*+ pools. In pancreatic acinar cells there is clear evidence for the existence of at least two separate intracellular non-

mitochondrial Ca*+ pools, one sensitive and the other insensitive to lns(1,4,5)P3 (Thevenod et al., 1989; Schulz et al., 1989). Why does the Ca2+-induced Ca*+ release occur as discrete pulses of Ca2+ outflow into the cytosol? In Goldbeter et al.‘s (1990) quantitative model, this is due to emptying of the Ins(l,4,5)Ps-insensitive pool caused by the rapid Ca*+ outflux. The pool therefore needs to be recharged before a new spike can occur, and in this period [Ca*+], is low because of the intensive pumping of Ca*+ into the pool. Only when the pool is full and the Ca*+ pumping stops does [Ca2+]i rise sufficiently to give renewed Ca*+-induced Ca*+ release. We have no direct evidence relating to this point, but our data are consistent with such a model. Another or additional mechanism may be negative feedback by Ca*+ on the Ca*+ release (Parker and Ivorra, 1990; Wakui and Petersen, 1990). For the muscle Ca*+activated Ca2+ release channel it has been shown that very high Ca*+ levels (millimolar) can inhibit opening of the channel (Meissner et al., 1986). Caffeine is generally thought to sensitize the Ca*+ release channel to the action of Ca*+ (Endo, 1977). Our results with the low caffeine concentration (1 mM) fit in with this concept. The acute inhibitory effects of a high caffeine concentration (20 mM) could be interpreted as being due to increased sensitivity to a negative feedback action of Ca2+. It seems unlikely that the acute inhibitory actions of caffeine (20 mM) are due to emptying of the pools, since Ca*+ spiking resumes immediately after removal of the drug (Figure 5). In rat chromaffin cells, there are often spontaneous cytoplasmic Ca*+ spikes, and caffeine (2-10 mM) induces an increase in the fluctuation frequency or evokes oscillations in cells that are initially silent. Agonists like bradykinin and histamine can also increase or induce Ca*+ spiking, but whereas neomycin (a drug that inhibits lns(1,4,5)Ps formation) blocks the agonist-evoked actions, it has no effect on those evoked by caffeine (Malgaroli et al., 1990). The results from the chromaffin cells indicating that caffeinesensitive Ca*+ release can cause Ca*+ spiking (Malgaroli et al., 1990) are in agreement with our data on the pancreatic acinar cells that, unlike the chromaffin cells, are electrically nonexcitable (Petersen and Findlay, 1987). The results presented here show that intracellular Ca*+ infusion evokes repetitive cytosolic Ca*+ spikes mediated by Ca*+ release through caffeine-sensitive channels. The Ca*+-induced repetitive Ca*+ spikes are very similar to those evoked by internal lns(1,4,5)P3 application, but can still be observed when the Ins(l,4,5)P+ensitive Ca*+ channels are blocked. Continuous Ins- (1,4,5)P3 activation of heparin-sensitive channels is necessary for sustained receptor-induced Ca2+ spiking. Experimental

Procedures

Cell Isolation Fragments of mouse pancreas were digested by pure collagenase (Worthington; 200 U/ml, 20-30 min, 37%) in the presence of 1 mM Ca’+ to obtain single cells as previously described (Wakui et al., 1999; Osipchuk et al., 1990).

Cytoplasmic 1031

Ca2+ Oscillations

Media for Experiments The standard extracellular solution contained (in mM): NaCl 140, KCI 4.7, CaC12 1.0, MgCI, 1.13, HEPES 10, and glucose 10 (pH was 7.2). In the pipette solutions, KCI was present at a concentration of 140 mM. The EGTA concentration was normally 0.25 mM, but in some experiments there was no EGTA, whereas in a few others 5 mM EGTA was present. No Ca*+ was added to the pipette solulion, but 1.13 mM MgClp and 5 mM Na2ATP were present. The HEPES concentration was 10 mM and the pH was always 7.2. The Ca2+ solution that was perfused into the patch-clamp pipette tip by pressure application to a thin infusion tube (20-300 mm Hg) was a normal pipette solution except for the absence of EGTA and the presence of 1 mM CaCI,. The free Caz+ concentration was calculated to be about 100 )IM. The Ins(l,4,5)P3-containing fluid was a standard pipette solution with 10 uM Ins(l,4,5)Ps (Sigma). In some experiments the phosphorothioate analog lns(1,4,5)PSs was used (kindly provided by Dr. B. V. L. Potter, Leicester University). Heparin was the low molecular weight (3000), sodium salt from Sigma. lonomycin and A23187 were from Calbiochem and dissolved in dimethyl sulfoxide (DMSO). The final DMSO concentration was normally less than 0.1% and always less than 1% and had no effect on its own (n = 4). All experiments were carried out at room temperature. Whole-Cell Cl- Current Recording The tight-seal, whole-cell current recording configuration of the patchclamp technique was used for the measurement of the transmembrane current from single cells (Hamill et al., 1981; Jauch et al., 1986), but it incorporated the additional feature of internal perfusion of the pipette tip (Jauch et al., 1986; Osipchuk et al., 1990) via a thin poly thene tube (portex). Acinar cells were voltage-clamped at a holding potential of -30 mV, and depolarizing voltage jumps of 100 ms duration to a membrane potential of 0 mV were applied repeatedly throughout the experiments. In all experiments shown, the Cl- equilibrium potential (ECl-) is about 0. Because of compression of pen-recording traces, all records seem to show currents at -30 and 0 mV simultaneously. Dotted horizontal lines indicate0 current level, and downward deflections represent inward current. Acknowledgments This work was supported by an MRC Programme Grant (0. H. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC Section 1734 solely to indicate this fact. Received

May 14, 1990; revised

September

4, 1990.

References Ashkenazi, A., Peralta, E. G., Winslow, J. W., Ramachandran, J., and Capon, D. J. (1989). Functionally distinct G proteins selectively couple different receptors to PI hydrolysis in the same cell. Cell 56, 487-493. Berridge, teracting

M. J. (1987). lnositol trisphosphate and diacylglycerol: two insecond messengers. Annu. Rev. Biochem. 56, 159-193.

Berridge, potential

M. J. (1988). lnositol trisphosphate-induced oscillations in Xenopus oocytes. J. Physiol.

Berridge, 9583-9586.

M. J. (1990).

Calcium

oscillations.

membrane 403, 589-599.

J. Biol.

Chem.

265,

Berridge, M. J., and Irvine, R. F. (1984). lnositol trisphosphate, a novel second messenger in cellular signal transduction. Nature 372, 315-321. Berridge, signalling.

M. J., and Irvine, R. F. (1989). Nature 347, 197-205.

lnositol

phosphates

Booth, C., and Koch, G. L. E. (1989). Perturbation of cellular induces secretion of luminal ER proteins. Cell 59, 729-737.

and cell calcium

Endo, M. (1977). Calcium Physiol. Rev. 57, n-108.

release

from

the sarcoplasmic

reticulum.

Ferris, C. D., Huganir, Ft. L.,Supattapone, S., and Snyder, S. H. (1989). Purified inositol 1,4,5-trisphosphate receptor mediates calcium flux In reconstituted lipid vesicles. Nature 342, 87-89. Ferris, C. D., Huganir, R. L., and Snyder, S. H. (1990). Calcium flux mediated by purified inositol 1.4,5-trisphosphate receptor in reconstituted lipid vesicles is allosterically regulated by adenine nucleotides. Proc. Natl. Acad. Sci. USA 87; 2147-2151. Furuichi, T., Yoshikawa, S., Miyawaki, A., Wada, K., Maeda, N., and Mikoshiba, K. (1989). Primary structure and functional expression of the inositol 1,4,5-trisphosphate-binding protein PdoO. Nature 342, 32-38. Gardner, 15-20.

P (1989).

Calcium

and

T lymphocyte

activation.

Cell 59,

Goldbeter, A., DuPont, G.. and Berridge. M. J. (1990). Minimal model for signal-induced Ca*+ oscillations and for their frequency encoding through protein phosphorylation. Proc. Natl. Acad. Sci. USA 87, 1461-1465. Hamill. 0. P., Marty, A., Neher, E.. Sakmann, B., and Sigworth, F. J. (1981). Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pfliigers Arch. 397, 85-100. Irvine, R. F. (1990). ‘QuantaI’ try by inositol phosphates-a 5-9.

Ca2+ release and the control of Ca2* enpossible mechanism. FEBS Lett. 263,

Jacob, R. (1990). Agonist-stimulated divalent cation entry into single cultured human umbilical vein endothelial cells. J. Physiol. 427, 55-77. Jan, L. Y., and Jan, Y. N. (1989). 56, 13-25.

Voltage-sensitive

ton channels.

Jauch, P., Petersen, 0. H., and Lauger, P (1986). Electrogenic of the sodium-alaninecotransporter in pancreaticacinarcells: seal whole-cell recordings. J. Membr. Biol. 94, 99-115.

Cell

properties I. Tight-

Kuno, M., and Gardner, P (1987). Ion channels activated by inosltol 1,4,5-trisphosphate in plasma membrane of human T-lymphocytes. Nature 326, 301-304. La& F. A., Erickson, H. P.. Rousseau, E., Liu, A.-Y., and Meissner. G. (1988). Purification and reconstitution of the calcium release channel from skeletal muscle. Nature 331, 315-319. Maeda, N., Niinobe, M., and Mikoshiba. K. (1990). A cerebellar purkinje marker Pm0 protein is an inositol 1,4,5-trisphosphate (Ins P3) receptor protein. Purification and characterization of Ins P3 receptor complex. EMBO J. 9, 61-67. Malgaroli, A., fluctuations trisphosphate dine-sensitive

Fesce, R., and Meldolesi, J. (1990). Spontaneous [Ca2+], in rat chromaffin cells do not require inositol-1,4,5elevations but are generated by a caffeine and ryanointracellular Ca2+ store. J. Biol. Chem. 265, 3005-3008.

Meissner, G., Darling, E., and Eveleth. J. (1986). Kinetics of rapid Ca*+ release by sarcoplasmic reticulum. Effects of Ca2+,Mg2+ and adenine nucleotides. Biochemistry 25, 236-244. Meyer, T., and Stryer, L. (1988). Molecular model for receptorstimulated calcium spiking. Proc. Natl. Acad. Sci. USA 85, 5051-5055. Morris, A. P, Gallacher, D. V., Irvine. R. F., and Petersen, Synergism of inositol trisphosphate and tetrakisphosphate ing Ca2+-dependent K+ channels. Nature 330. 653-655.

0. H. (1987). in activat-

Osipchuk, Y. V., Wakui, M., Yule, D. I., Gallacher, D. V., and Petersen, 0. H. (1990). Cyioplasmic Ca*+ oscillations evoked by receptor stimulation, G-protein activation, internal application of inositol trisphosphate or Ca2+: simultaneous microfluorimetry and Ca*‘-dependent Cl- current recording in single pancreatic acinar cells. EMBO J. 9, 697-704. Parker, I., and Ivorra, I. (1990). Inhibition by Ca2+ of inositol trisphosphate-mediated Ca2+ liberation: a possible mechanism for oscillatory release of Ca*+. Proc. Natl. Acad. Sci. USA 87; 260-264.

Downes, C. P., and Michell, R. H. (1981). The polyphosphoinositide phosphodiesterase of erythrocyte membranes, Biochem. J. 198, 133-140.

Penner, R., Matthews, G.. and Neher, E. (1988). influx by second messengers in rat mast cells.

Ehrlich, B. E.. and Watras, J. (1988). vates a channel from smooth muscle 336, 583-586.

Penner, R., Neher, E., Takeshima. H., Nishimura, S., and Numa, S. (1989). Functional expression of the calcium release channel from skeletal muscle ryanodine receptor cDNA. FEBS Lett. 259, 217-221.

lnositol 1,4,5trisphosphate sarcoplasmic reticulum,

actiNature

Regulation of calcium Nature 334, 499-504.

Cell 1032

Petersen, 0. H. (1989). Does inositol tetrakisphosphate the receptor-mediated control of calcium mobilization? 10, 3x-303. Petersen, 0. H., and Findlay, I. (1987). creas. Physiol. Rev. 67, 1054-1116.

Electrophysiology

play a role in Cell Calcium of the pan-

Ross, C. A., Meldolesi, J., Milner, T. A., Satoh, T., Supattapone, S., and Snyder, S. H. (1989). lnositol 1,4$trisphosphate receptor localized to endoplasmic reticulum in cerebellar purkinje neurons, Nature 339, 468-470. Sambrook, J. F. (1990). The involvement secretory proteins from the endoplasmic

of calcium in transport of reticulum. Cell 67, 197-199.

Schmid, A., Kremer-Dehlinger, M.. Schulz, I., and Gogelein, H. (1990). Voltage-dependent InsPs-insensitive calcium channels in membranes of endoplasmic reticulum vesicles. Nature 346, 374-378. Schnefel, S., Bannc. H., Eckhardt, L., Schultz, G., and Schulz, I, (1988). Acetylcholine and cholecystokinin receptors functionally coupled by different G-proteins to phospholipase C in pancreatic acinar cells. FEES Lett. 230, 125-130. Schulz, I., Thevenod, F., and Dehlinger-Kremer, M. (1989). Modulation of intracellular free Ca2+ concentration by IPs-sensitive and IPsinsensitive nonmitochondrial Ca2+ pools. Cell Calcium 10, 325-338. Streb, H., Irvine, R. F., Berridge, M. J., and Schulz, I. (1983). Release of Ca*+ from a non-mitochondrial intracellular store of pancreatic acinar cells by inositol 1,4,5trisphosphate. Nature 306, 87-69. Swann, K., Igusa, Y., and Miyazaki, S. (1989). Evidence for an inhibitory effect of protein kinase C on G-protein-mediated repetitive calcium transients in hamster eggs. EMBO J. 6, 3711-3718. Takeshima, H., Nishimura, S., Matsumoto, T., Ishida, H., Kangawa, K., Minamino, N., Matsuo, H., Ueda, M., Hanaoka, M., Hirose, T., and Numa, S. (1989). Primary structure and expression from complementary DNA of skeletal muscle ryanodine receptor. Nature 339,439-445. Taylor, C. W., Merritt, J. E., Putney, fects of Ca*+ on phosphoinositide Biochem. J. 238, 785-772.

J. W., and Rubin, R. I? (1986). Efbreakdown in exocrine pancreas.

Thevenod, E, Dehlinger-Kremer, M., Kemmer, T. I?, Potter, B. V. L., and Schulz, I. (1989). Characterization trisphosphate-sensitive (IsCaP) and -insensitive mitochondrial Caz+ pools in rat pancreatic acinar cells. 109, 173-186.

Christian, A.-L., of inositol 1,4,5(IisCaP) nonJ. Membr. Biol.

Wakui, M., and Petersen, 0. H. (1990). Cytoplasmic Ca2+ oscillations evoked by acetylcholine or intracellular infusion of inositol trisphosphate or Ca*+ can be inhibited by internal Cap+. FEBS Lett. 263, 208-208. Wakui, M., Potter, B. V. L.. and Petersen, 0. H. (1989). Pulsatile intracellular calcium release does not depend on fluctuations in inositol trisphosphate concentration. Nature 339, 317-320. Woods, N. M., Cuthbertson, K. S. R., and Cobbold, Agonist-induced oscillations in cytoplasmic free calcium in single rat hepatocytes. Cell Calcium 6, 79-100. Note

Added

Correspondence

in Ptuof should

be addressed

to 0. H. I?

P H. (1987). concentration