Basal and Physiological Ca2+Leak from the Endoplasmic Reticulum of Pancreatic Acinar Cells

We have studied the Ca2+ leak pathways in the endoplasmic reticulum of pancreatic acinar cells by directly measuring Ca2+ in the endoplasmic reticulum ([Ca2+] ER ). Cytosolic Ca2+ ([Ca2+] C ) was clamped to the resting level by a BAPTA-Ca2+ mixture. Administration of cholecystokinin within the physiological concentration range caused a graded decrease of [Ca2+] ER , and the rate of Ca2+ release generated by 10 pm cholecystokinin is at least 3× as fast as the basal Ca2+ leak revealed by inhibition of the endoplasmic reticulum Ca2+-ATPase. Acetylcholine also evokes a dose-dependent decrease of [Ca2+] ER , with an EC50 of 0.98 ± 0.06 μm. Inhibition of receptors for inositol 1,4,5-trisphosphate (IP3) by heparin or flunarizine blocks the effect of acetylcholine but only partly blocks the effect of cholecystokinin. 8-NH2 cyclic ADP-ribose (20 μm) inhibits the action of cholecystokinin, but not of acetylcholine. The basal Ca2+ leak from the endoplasmic reticulum is not blocked by antagonists of the IP3 receptor, the ryanodine receptor, or the receptor for nicotinic acid adenine dinucleotide phosphate. However, treatment with puromycin (0.1–1 mm) to remove nascent polypeptides from ribosomes increases Ca2+leak from the endoplasmic reticulum by a mechanism independent of the endoplasmic reticulum Ca2+ pumps and of the receptors for IP3 or ryanodine.

Many physiological and pharmacological responses rely on the ability of intracellular messengers to generate cytosolic Ca 2ϩ signals by releasing Ca 2ϩ from the endoplasmic reticulum (ER) 1 (1)(2)(3)(4). Thus, in the pancreatic acinar cell the neurotransmitter acetylcholine (ACh) stimulates the release of Ca 2ϩ from the ER via inositol 1,4,5-trisphosphate (IP 3 ), while the hormone cholecystokinin (CCK) evokes Ca 2ϩ release by a com-plex interaction between the messengers IP 3 , cyclic ADPribose, and nicotinic acid adenine dinucleotide phosphate (NAADP) (5). Although it is well established that low concentrations of these agonists generate long lasting trains of different forms of cytosolic Ca 2ϩ oscillation (5,6), there are few data on the kinetics of Ca 2ϩ in the ER lumen ([Ca 2ϩ ] ER ), particularly during the action of physiological concentrations of agonists.
After agonist-induced depletion, the Ca 2ϩ content of the ER is replenished by the sarco/endoplasmic reticulum Ca 2ϩ -ATPase (SERCA) pump, but even in the resting state there is a basal leak of Ca 2ϩ from the lumen of ER, which can be revealed by the inhibition of SERCA with thapsigargin or cyclopiazonic acid (CPA) (7,8). Although the molecular nature of the basal leak is still unclear, in permeabilized hepatocytes the temperature dependence and kinetics of the leak suggest that it occurs through a channel (9). Missiaen et al. (10) found that in A7r5 smooth muscle cells the leak rate was fitted by a two-exponential decay.
In skeletal and cardiac muscle, sarcoplasmic reticulum basal Ca 2ϩ efflux may occur through the ryanodine receptor (11,12), and in non-muscle cells it has been suggested that basal Ca 2ϩ leak from the ER reflects the flow of Ca 2ϩ through the IP 3 receptor induced by the action of resting levels of IP 3 (13,14). However in a study in baby hamster kidney fibroblasts, Hofer et al. (7) clearly demonstrated that the leak was not blocked by either the IP 3 receptor antagonist heparin or the ryanodine receptor antagonist ruthenium red.
In contrast to the hypothesis that the basal Ca 2ϩ leak occurs through second messenger-activated Ca 2ϩ channels in the ER membrane, it has been suggested that the leak could occur through the translocon pore complex in the ER membrane (8). Recent studies have suggested that the empty pore of the translocon complex is permeable to small ions and neutral molecules (15)(16)(17). Experimentally, the permeability of the translocon can be modified by puromycin (16,17), an adenosine derivative that purges the translocon of nascent polypeptides, creating an empty pore (18,19), and we have used this tool to investigate for the first time the permeability of the translocon pore to Ca 2ϩ .
In this study we have investigated the basal and agonistevoked pathways of Ca 2ϩ leak by directly measuring the depletion of [Ca 2ϩ ] ER in isolated pancreatic acinar cells. Moreover, to avoid the profound regulatory effects of cytosolic Ca 2ϩ ([Ca 2ϩ ] C ) on Ca 2ϩ release channels (20), [Ca 2ϩ ] in the solution bathing the ER was "clamped" at a quasi-resting level (ϳ90 nM) by dialyzing the cell with a mixture of calcium and the Ca 2ϩ chelator BAPTA. This provides a particularly sensitive method for studying very slow Ca 2ϩ fluxes.
Using this experimental approach we have set out to study the following: (i) the depletion of [Ca 2ϩ ] ER by physiological doses of CCK and ACh, (ii) the relative magnitudes of the basal Ca 2ϩ leak and the Ca 2ϩ leak stimulated by physiological doses of agonists, (iii) the contribution of IP 3 and ryanodine receptors to agonist-evoked and basal Ca 2ϩ leak, and (iv) the effect on [Ca 2ϩ ] ER on the dissociation of nascent polypeptide chains from the ribosome with puromycin.

EXPERIMENTAL PROCEDURES
Cell Preparation and Solutions-Mouse pancreatic acinar cells were isolated by digestion with purified collagenase (200 units ml Ϫ1 , Worthington Biomedical Corp., Lakewood NJ) as described previously (8). Freshly isolated cells were incubated with 5 M fura-2FF/AM (Teflabs, Austin TX) or 5 M mag-fura-2/AM (Molecular Probes Europe BV, Leiden, The Netherlands) and pluronic F-127 (0.025%) for 30 -45 min at 37°C. Cells were attached to poly(L-lysine)-coated cover slips, installed in a flow chamber and placed on the stage of a Nikon Diaphot inverted fluorescence microscope (Nikon Ltd., Kingston, UK). Imaging experiments were performed at room temperature (20 -22°C). Extracellular solutions contained (in mM): NaCl, 140; KCl, 4.7; MgCl 2 , 1.13; CaCl 2 , 1; glucose, 10; and HEPES-NaOH, 10 (pH 7.2) and were perfused rapidly under gravity using electronic valves (Lee Products Ltd., Gerrard's Cross, Bucks, UK). Cells were alternatively illuminated by 340 and 380 nm light from a monochromator using a ϫ40, 1.3 NA objective lens, and emission light with a wavelength longer than 400 nm was collected. Images from a CCD camera (Photonic Sciences, Beaconsfield, UK) were digitized, averaged, and analyzed using a Quanticell 700 m imaging system from Visitech International (Sunderland, UK). In order to remove fura-2FF or mag-fura-2 from the cytosol and control the composition of the intracellular medium, imaging experiments were performed either with patch-clamped cells after dialysis of the cytosol (21) or with streptolysin O-treated cells after permeabilization of the plasma membrane. Experiments with streptolysin O-permeabilized cells allowed the composition of the solution bathing the ER to be changed rapidly (7,22).
Permeabilization of Acinar Cells-Cells attached to cover slips were perfused with intracellular medium containing 0.5 units/ml reduced streptolysin O (Corgenix UK, Peterborough, UK) for 30 s at room temperature (7,22). Experiments were carried out in medium with an ionic composition identical to the patch-pipette intracellular solution.
Measurement of [Ca 2ϩ ] ER -Averaged measurements of the ratio of fluorescence at 340 nm to fluorescence at 380 nm were made at 5-s intervals. In experiments with low, physiological doses of CCK, analysis was performed only in the basolateral, non-nuclear area of the acinar cell, where the density of ER is highest (23,24); this area also has a lower density of other organelles and is less susceptible to movement artifacts. [Ca 2ϩ ] ER was estimated from ratio measurements by an in situ cell calibration (25) using R min values obtained with 10 M ionomycin (Iono) and 10 mM EGTA, and R max values with 10 M ionomycin and 20 mM CaCl 2 . The K d for fura-2FF-free acid was found to be 31.5 M by a cell-free calibration using Ca 2ϩ /EGTA buffers (Molecular Probes Europe BV).
Chemicals and Statistics-All chemicals were obtained from Sigma-Aldrich Co. Ltd. unless otherwise stated. All quantitative data refer to fluorescence ratios of fura-2FF-loaded cells and are presented as means Ϯ S.E. of the mean. Statistical comparisons were by paired or unpaired two-tailed Student's t test, and analysis of sigmoid curves was performed with Microcal Origin using a Boltzmann fit.

RESULTS
Following permeabilization of the plasma membrane or whole-cell dialysis of dye-loaded pancreatic acinar cells, the ratio of fluorescence at 340 to 380 nm rose as cytosolic dye was washed out (8). In 51 fura-FF-loaded patch-clamped cells, the mean ratio in the basolateral non-nuclear area after dialysis was 0.63 Ϯ 0.03, corresponding to a mean apparent [Ca 2ϩ ] ER of 80 M.
Physiological Concentrations of Cholecystokinin Deplete [Ca 2ϩ ] ER -CCK (CCK-8, sulfated form) generated stepwise decreases in ratio, but at picomolar concentrations of the agonist the ratio reached the new steady-state level relatively slowly. At 1 pM CCK, the plasma concentration seen in the fasting state in the mouse 2 and in humans (26), there was a small but detectable decrease in ratio (Fig. 1). In 7 cells the mean decrease was 0.014 Ϯ 0.003 (Fig. 2). However at 10 pM CCK, a concentration seen in the plasma after a meal, there was a more substantial mean decrease in ratio of 0.043 Ϯ 0.006 (13 cells). The CCK-induced decreases in ratio were uniform throughout the basolateral non-nuclear area of the cell, where the resting ratio was highest. 10 nM CCK evoked a large average decrease in the ratio of 0.142 Ϯ 0.005 (Fig. 2).
Depletion of [Ca 2ϩ ] ER by Different Doses of Acetylcholine-ACh (10 nM to 100 M) caused a sharp decrease in ratio ( Fig. 3) with low concentrations (10 -100 nM) often being associated with a small "rebound" from the initial decrease (Fig. 3B). The decrease in the ratio evoked by ACh was most pronounced in the basal area of the cell and was uniform in this region. In 2 J. F. Rehfeld, personal communication. other regions of the cell including the lateral and apical regions, ratio decreases had similar time courses but slightly smaller amplitudes (Fig. 3B). The EC 50 for the ACh-evoked decrease in ratio was 0.98 Ϯ 0.06 M (Fig. 3C). In 10 cells 10 M ACh gave a mean decrease in ratio of 0.135 Ϯ 0.024; this was similar in size to the decrease evoked by 10 nM CCK (Fig. 2). In the presence of a supramaximal concentration of ACh, 10 nM CCK was unable to cause further depletion of [Ca 2ϩ ] ER (n ϭ 5, data not shown).
The Basal Leak Is Considerably Smaller than the Leak Evoked by 10 pM CCK-After a short lag, the application of the SERCA pump inhibitor CPA (10 M) evoked a steady decrease in the ratio (Fig. 4). This is consistent with the unmasking of the basal leak of Ca 2ϩ from intracellular stores following inhibition of Ca 2ϩ uptake. Thapsigargin (2 M) evoked a quantitatively similar depletion of [Ca 2ϩ ] ER (data not shown). The mean rate of decrease evoked by CPA in 14 cells was 0.018 Ϯ 0.006 min Ϫ1 , much smaller than the rate of decrease of the ratio due to 10 M ACh (0.15 Ϯ 0.02 min Ϫ1 , n ϭ 8). When 10 pM CCK was applied to cells in which the basal leak had already been revealed by CPA treatment the mean rate of Ca 2ϩ loss was accelerated (Fig. 4). The decreases in the ratio evoked by CPA and CCK were greatest in the basolateral area; slightly smaller decreases with similar kinetics were seen throughout the cell (Fig. 4). In 7 cells, 10 pM CCK significantly increased (p Ͻ 0.05, paired Student's t test) the mean rate of decline evoked by CPA (by 265 Ϯ 82%), whereas in 6 cells 1 pM CCK increased the CPA-evoked rate of decline by only 10 Ϯ 20%.
IP 3 Receptor Antagonists Block the Action of ACh but Only Partially Block the Action of CCK and Do Not Block the Basal Leak-In 17 cells dialyzed with the competitive IP 3 receptor antagonist heparin (M r 6000; 500 g/ml) the mean resting ratio (0.59 Ϯ 0.03) was not significantly different from cells dialyzed with control pipette solution. Heparin substantially blocked the response to ACh (Fig. 5A), and in 10 cells dialyzed with heparin the response to 10 M ACh was decreased by 80% compared with control cells (Figs. 2 and 5A). Heparin did not completely block the effect of CCK (Fig. 5A). In heparin-dialyzed cells the mean response to 10 nM CCK was not different from the mean response in control cells (Figs. 2 and 5A); however the effect of 10 pM CCK was more than halved by heparin (Fig. 2). In heparin-dialyzed cells CPA was still able to reveal the basal leak (0.012 Ϯ 0.005 min Ϫ1 , n ϭ 4) and deplete [Ca 2ϩ ] ER (Fig.  5B). We also studied the effects of another receptor antagonist flunarizine (10 M) on agonist-evoked and basal Ca 2ϩ release. In our study this compound, which is membrane-permeant and has been previously reported to block directly the Ca 2ϩ channel of the IP 3 receptor (27,28), inhibited the effect of ACh; however CCK and CPA were still able to deplete [Ca 2ϩ ] ER in the presence of flunarizine (n ϭ 4, not shown).

Ryanodine Receptor Antagonists Substantially Inhibit the Action of CCK but Not the Effect of ACh or the Basal Leak-
Dialysis with 8-amino-cyclic ADP-ribose (20 M, Molecular Probes Europe BV), a competitive inhibitor of cyclic ADP-ribose action (29,30), had no effect on mean resting [Ca 2ϩ ] ER (mean ratio ϭ 0.67 Ϯ 0.08 in the presence of 8-amino-cyclic ADPribose, n ϭ 7) or on the action of ACh, but blocked the depletion of [Ca 2ϩ ] ER by 10 pM CCK (Fig. 6A). In 5 cells dialyzed with this inhibitor the mean ratio decrease evoked by 10 pM CCK was reduced by 87 Ϯ 11% (p Ͻ 0.02 compared with control cells). In the presence of 8-amino-cyclic ADP-ribose, CPA was still able to deplete the Ca 2ϩ store at a rate similar to in control conditions (0.02 Ϯ 0.009 min Ϫ1 , n ϭ 4). Inhibition of CPA responses by 8-amino-cyclic ADP-ribose was not seen even when the IP 3 receptor blocker flunarizine was present as well (n ϭ 4). In order to investigate further whether the basal leak could occur through the ryanodine receptor we studied the effect on the basal leak of ruthenium red, a blocker of the ryanodine receptor (20,31). These experiments were performed on permeabilized cells, allowing relatively fast and simple changes of the solutions in contact with the ER. In 16 permeabilized cells CPA (10 M) evoked a mean ratio decrease of 0.021 Ϯ 0.004 min Ϫ1 , but ruthenium red (10 M) had no effect on this basal leak (Fig. 6B). We also studied the effect of nifedipine and verapamil (both 100 M) on the basal Ca 2ϩ leak in permeabilized cells. It has been reported that in sea urchin eggs these Ca 2ϩ channel blockers completely inhibit the release of stored Ca 2ϩ evoked by NAADP (32), but we found that in the pancreatic acinar cell they had no effect on the basal Ca 2ϩ leak evoked by CPA (n ϭ 8 cells for both compounds, data not shown).
Puromycin Evokes a Decrease in [Ca 2ϩ ] ER -In patchclamped and dialyzed cells the protein synthesis inhibitor puromycin (100 -500 M) produced an initial decrease in ratio, which slowed down after a short period (Fig. 7A). Mean fura-2FF ratio decreases with 100 and 500 M puromycin were 0.08 Ϯ 0.02 and 0.167 Ϯ 0.04, respectively (10 cells). Ratio measurements in acinar cells loaded with the low-affinity Ca 2ϩ indicator mag-fura-2 and dialyzed with BAPTA/Ca 2ϩ intracellular solution confirmed the ability of puromycin to deplete [Ca 2ϩ ] ER (n ϭ 4; data not shown). We also studied the effects of puromycin in fura-2FF-loaded acinar cells permeabilized by streptolysin O, because this allowed faster changes of bathing solution during the experiment. Responses to puromycin were seen in permeabilized cells (Fig. 7B), and the mean ratio decrease evoked by 500 M puromycin was 0.11 Ϯ 0.02 (n ϭ 16), and by 1 mM puromycin 0.14 Ϯ 0.03 (n ϭ 7). At 20 M, puro- mycin had no effect (n ϭ 4). To exclude the possibility that the effects of puromycin were mediated via agonist-activated Ca 2ϩ release channels, we applied puromycin to permeabilized cells during blockade of IP 3 and ryanodine receptors. In the presence of the IP 3 receptor inhibitor heparin (500 g/ml), puromycin (500 M) decreased the mean ratio of 10 permeabilized cells by 0.091 Ϯ 0.015 (Fig. 8A), and in the presence of the ryanodine receptor antagonist ruthenium red (10 M) puromycin decreased the mean ratio of 8 permeabilized cells by 0.074 Ϯ 0.019 (Fig. 8B). Neither of these values was significantly different from the effect of puromycin alone. To investigate the possibility that the depletion of [Ca 2ϩ ] ER by puromycin could be because of a decreased activity of the SERCA pump rather than an increase in Ca 2ϩ permeability of the ER membrane, we studied the effect of puromycin on [Ca 2ϩ ] ER when SERCA was inhibited by CPA. Puromycin (500 M) accelerated the Ca 2ϩ leak evoked by CPA (Fig 8C). In 36 cells the mean Ca 2ϩ leak was increased from 0.013 Ϯ 0.002 to 0.018 Ϯ 0.002 (p Ͻ 0.01, paired Student's t test).

DISCUSSION
In this study we were able to dissociate the effects of physiological doses of agonists on the initial or intrinsic Ca 2ϩ efflux from the ER from any effects on Ca 2ϩ signaling due to secondary Ca 2ϩ -induced Ca 2ϩ release. This was achieved by direct measurements of [Ca 2ϩ ] ER while [Ca 2ϩ ] C was clamped at a quasi-resting level with a BAPTA/Ca 2ϩ mixture. In these highly sensitive experimental conditions, where feedback of Ca 2ϩ on its own efflux was prevented, we were able to resolve the depletion of [Ca 2ϩ ] ER produced by the hormone CCK at 1 pM (the fasting plasma level), and at 10 pM, a level achieved in the plasma after a meal. 1 and 10 pM CCK produced ratio changes corresponding to ϳ10 and 32% of the decrease induced by a supramaximal dose of CCK, respectively. Our experiments indicate that SERCA pumps can balance the leak induced by all doses of CCK in the physiological range and prevent substantial depletion of the store. This is important because substantial depletion of ER can inhibit protein synthesis, facilitate protein degradation and affect protein folding (33)(34)(35)(36)(37).
A recent study by Pinton et al. (38) in HeLa cells has indicated that overexpression of the anti-apoptotic protein bcl-2 can decrease the Ca 2ϩ content of the ER, suggesting that this protein could mediate (or regulate) Ca 2ϩ leak from the ER. Moreover, this study implies a new "trophic" action of low physiological doses of Ca 2ϩ -releasing hormones, i.e. protection of the ER from Ca 2ϩ overload and consequently from apoptotic destruction of the cell.
Previous measurements of cytosolic Ca 2ϩ have indicated that in intact cells physiological concentrations of CCK generate, after a delay of 1-2 min, a mixture of fast, local spikes and slow global Ca 2ϩ oscillations (5,30,39). Our study describes an increase of the permeability of the ER that is purely due to action of the hormone (without amplification by Ca 2ϩ -induced Ca 2ϩ release). The mechanism by which moderate increases in Ca 2ϩ efflux from the ER are converted into different forms of Ca 2ϩ oscillation by other cellular mechanisms is an interesting subject for further theoretical and experimental studies. For ACh the physiological concentration range in the vicinity of the pancreatic acinus is unknown, therefore we have characterized the effects of a broad range of ACh concentration on [Ca 2ϩ ] ER . The ability of low doses of ACh (10 and 100 nM) to generate a pattern of short lasting local Ca 2ϩ spikes similar to that produced by IP 3 infusion into patch-clamped cells (6) has been well characterized (5,39). Our measurements of [Ca 2ϩ ] ER suggest that 10 and 100 nM ACh cause between 13 and 25% of the maximal agonist-dependent ratio changes, respectively. The depletion of [Ca 2ϩ ] ER induced by these doses of ACh, and by physiological doses of CCK, is clearly quantitatively similar. Therefore the difference in the patterns of cytosolic Ca 2ϩ re- sponses of the two agonists cannot simply be explained by different levels of Ca 2ϩ efflux from the ER.
In this study we found that the ability of ACh to induce Ca 2ϩ leak from the ER was blocked by heparin, a competitive IP 3 receptor antagonist (40), and by flunarizine, which does not affect IP 3 binding to its receptor but blocks the Ca 2ϩ release channel activated by IP 3 (27,28). This strongly suggests that the primary mechanism of ACh-induced Ca 2ϩ leak is mediated by IP 3 receptors.
In contrast, the role of IP 3 receptors in acinar cell Ca 2ϩ signaling by CCK is more complex. Using a biochemical radioreceptor assay Matozaki et al. (41) showed that supraphysiological (Ͼ50 pM) concentrations of CCK generate measurable amounts of IP 3 , and studies using permeabilized cells (42) and patch-clamped cells (5,43) have reported that inhibitors of the IP 3 receptor completely block CCK-evoked [Ca 2ϩ ] C signals. However in another study Thorn et al. (39) described that when acinar cells were dialyzed with low concentrations (Ͻ250 g/ ml) of heparin, physiological doses of CCK (5-20 pM) still produced long lasting oscillations in [Ca 2ϩ ] C (although short, IP 3type [Ca 2ϩ ] C spikes were blocked). In the present study in which "intrinsic," messenger-evoked Ca 2ϩ efflux from the ER and secondary Ca 2ϩ -induced Ca 2ϩ release were dissociated, we found that inhibition of the IP 3 receptor caused partial inhibition of the effect of 10 pM CCK, as did inhibition of the ryanodine receptor with 8-NH 2 cyclic ADP-ribose. These data support the hypothesis that although ACh-induced efflux is mediated by IP 3 , physiological concentrations of CCK stimulate efflux dependent on multiple messengers, including IP 3 and cyclic ADP-ribose. Because inhibition of IP 3 receptors has been found to block [Ca 2ϩ ] C oscillations evoked by the putative Ca 2ϩ releasing messengers cyclic ADP-ribose and NAADP (5), this suggests that CCK-evoked [Ca 2ϩ ] C oscillations rely on the recruitment of IP 3 receptors to amplify the [Ca 2ϩ ] C signal. Our data show that the CCK-induced Ca 2ϩ efflux is partially dependent on IP 3 receptors. We found that heparin did not block the depletion of [Ca 2ϩ ] ER by 10 nM CCK, suggesting that although nanomolar concentrations of CCK do generate IP 3 (41), when IP 3 action is blocked by heparin other CCK-stimulated messengers could be able to evoke a substantial release of Ca 2ϩ from the ER. Our observations that in the presence of supramaximal doses of ACh, the addition of supramaximal doses of CCK is unable to deplete [Ca 2ϩ ] ER further, suggesting that although primary CCK-induced Ca 2ϩ efflux does involve additional Ca 2ϩ -releasing messengers, it does not involve a separate Ca 2ϩ store. Qualitatively similar time courses of [Ca 2ϩ ] ER were seen within different regions of a single cell during depletion of [Ca 2ϩ ] ER with secretagogues and with CPA (Figs. 3 and  4). This supports the concept that the endoplasmic reticulum of the pancreatic acinar cell acts as a single agonist-releasable Ca 2ϩ store (8,22,24).
Hofer et al. (7) reported that neither heparin nor the ryanodine receptor antagonist ruthenium red blocked the basal leak revealed by SERCA inhibition. Our study supports this finding. We showed that cells dialyzed with heparin had a similar resting [Ca 2ϩ ] ER to control cells, suggesting no substantial differences from control cells in the pump/leak relationship. Neither heparin, nor the membrane-permeant Ca 2ϩ channel blocker flunarizine, which does not affect IP 3 binding to its receptor but has been reported to inhibit Ca 2ϩ release by IP 3 (27,28), blocked the basal leak evoked by CPA. Furthermore, neither of the two ryanodine receptor antagonists used (ruthenium red and 8-NH 2 cyclic ADP-ribose) nor nifedipine or verapamil, which in sea urchin eggs completely block the Ca 2ϩ release evoked by NAADP (32), had any effects on the basal Ca 2ϩ leak. We therefore found no evidence to suggest that in the pancreatic acinar cell the basal Ca 2ϩ leak occurs through any of the Ca 2ϩ release channels so far identified in the ER membrane.
In contrast it has been hypothesized that the basal Ca 2ϩ leak from the ER into the cytosol may occur through the aqueous pore in the translocon of the ER membrane during the protein synthetic cycle (8). During the normal cycle of protein synthesis the permeability of the translocon is tightly controlled, possibly because of the binding of the ribosome to the cytosolic surface of the translocon pore (44). The permeability of the translocon is also regulated at the luminal side of the pore by the prominent ER chaperone BiP (45), this protein being released from the translocon shortly after the completion of ribosome-nascent chain targeting. However in the empty state, when the translocon pore is ribosome-bound but unoccupied by polypeptide, the ribosome-translocon complex seems to allow the passage of small molecules (17).
In the present study we have examined the effect of puromycin on the Ca 2ϩ permeability of the ER membrane, and found a substantial puromycin-induced Ca 2ϩ efflux. Puromycin is an antibiotic that selectively terminates ribosomal translation by releasing the nascent polypeptide from the protein channel of the ribosome (18,19). Simon et al. (15,16) used an electrophysiological approach to show that in pancreatic rough microsomes puromycin activates an ion-permeable pore. Interestingly, spontaneous openings of large conductance ion channels in the ER membrane, possibly representing subconductance states of the translocon channel, were reported even in the absence of puromycin (16). These spontaneous openings could be responsible for the basal Ca 2ϩ leak. If such a translocon-mediated Ca 2ϩ leak exists, then it should be particularly prominent in the pancreatic acinar cell with its extremely well developed rough endoplasmic ER and very high protein-synthesizing activity. Importantly, a very recent study by Potter and Nicchitta (46) has demonstrated that ribosomes maintain stable associations with translocons after the termination of protein synthesis. According to our results, in the protein-free state such endogenous ribosome-translocon complexes could serve as mediators of the basal Ca 2ϩ leak from the ER.
In our study we have shown, for the first time, that removal of the polypeptide chain from the ribosome by puromycin causes a depletion of [Ca 2ϩ ] ER by a mechanism independent of IP 3 receptors and ryanodine receptors and independent of inhibition of the SERCA pump. Our study therefore provides experimental support for the hypothesis that the basal Ca 2ϩ leak from the rough ER occurs through translocon pores in the ER membrane.