Differential effects of protein kinase C activation on calcium storage and capacitative calcium entry in NIH 3T3 cells.

In NIH 3T3 cells, treatment with phorbol 12-myristate 13-acetate (PMA) reduced the release of Ca2+ by thapsigargin, but did not activate Ca2+ entry; Ca2+ influx was triggered after the residual pool was emptied by thapsigargin, and this Ca2+ influx was similar to that induced by thapsigargin in control cells. The effect of PMA was due to decreased Ca2+ storage because 1) Ca2+ release by ionomycin was similarly affected by PMA, and in both control and PMA-treated cells, ionomycin did not release Ca2+ following thapsigargin treatment; 2) PMA reduced 45Ca2+ accumulation; and 3) studies with Ca2+ indicator compartmentalized into the endoplasmic reticulum indicated that stored Ca2+ was reduced by PMA. Although PMA did not itself activate Ca2+ entry, PMA potentiated Ca2+ entry with low concentrations of cyclopiazonic acid. With a somewhat higher concentration of cyclopiazonic acid, PMA had no effect on calcium entry. Thus, protein kinase C has two apparent actions on calcium signaling in NIH 3T3 cells: 1) reduced intracellular Ca2+ storage capacity and 2) augmented calcium entry with submaximal intracellular Ca2+ pool depletion. These actions indicate a complex and potentially important role for the protein kinase C system in calcium homeostasis in this cell type.

In NIH 3T3 cells, treatment with phorbol 12-myristate 13-acetate (PMA) reduced the release of Ca 2؉ by thapsigargin, but did not activate Ca 2؉ entry; Ca 2؉ influx was triggered after the residual pool was emptied by thapsigargin, and this Ca 2؉ influx was similar to that induced by thapsigargin in control cells. The effect of PMA was due to decreased Ca 2؉ storage because 1) Ca 2؉ release by ionomycin was similarly affected by PMA, and in both control and PMA-treated cells, ionomycin did not release Ca 2؉ following thapsigargin treatment; 2) PMA reduced 45 Ca 2؉ accumulation; and 3) studies with Ca 2؉ indicator compartmentalized into the endoplasmic reticulum indicated that stored Ca 2؉ was reduced by PMA. Although PMA did not itself activate Ca 2؉ entry, PMA potentiated Ca 2؉ entry with low concentrations of cyclopiazonic acid. With a somewhat higher concentration of cyclopiazonic acid, PMA had no effect on calcium entry. Thus, protein kinase C has two apparent actions on calcium signaling in NIH 3T3 cells: 1) reduced intracellular Ca 2؉ storage capacity and 2) augmented calcium entry with submaximal intracellular Ca 2؉ pool depletion. These actions indicate a complex and potentially important role for the protein kinase C system in calcium homeostasis in this cell type.
Hormone agonists coupled to receptor activation of phospholipase C increase intracellular calcium through two distinct but related pathways. Initially, inositol 1,4,5-trisphosphate (IP 3 ), 1 formed from phospholipase C-mediated breakdown of phosphatidylinositol 4,5-bisphosphate, binds to channel receptors located in the endoplasmic reticulum (ER), leading to channel opening and release of stored calcium to the cytoplasm. Subsequently, depletion of intracellular Ca 2ϩ stores activates a pathway for Ca 2ϩ entry across the plasma membrane, which has been termed "capacitative Ca 2ϩ influx" (1,2). Several reports utilizing reagents that inhibit the Ca 2ϩ -ATPase responsible for Ca 2ϩ storage within the ER (for example, thapsigargin and cyclopiazonic acid) have suggested that Ca 2ϩ store depletion provides a full and sufficient signal for activation of capacitative Ca 2ϩ entry (reviewed in Refs. 3 and 4). Studies on human neutrophils in which Ca 2ϩ stores were fully or partially de-pleted by IP 3 , chelating agents, or a Ca 2ϩ ionophore suggested a correlation between the percentage of Ca 2ϩ pool depletion and the magnitude of capacitative Ca 2ϩ entry (5).
In this report, we describe a novel action of the PKC activator phorbol 12-myristate 13-acetate (PMA). Treatment of NIH 3T3 cells at 37°C with PMA caused a marked reduction in the calcium storage capacity of the ER. However, capacitative calcium entry was not activated. Rather, with submaximal depletion of intracellular stores by cyclopiazonic acid, PMA facilitated calcium entry. This indicates that PKC may be an important and complex regulator of cellular calcium storage and homeostasis.

MATERIALS AND METHODS
NIH 3T3 Cell Culture-NIH 3T3 cells were maintained at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, 5 mM glutamine, 50 units/ml penicillin, and 50 units/ml streptomycin. After 3 days in culture, subconfluent cells were passed at a dilution of 1:10. For measurements of [Ca 2ϩ ] i , the cells were plated to subconfluence on glass coverslips 2 days before use. The cells were then incubated with fura-2/AM or fura-F/AM as described below.
Fura-2 and Fura-F Loading-The coverslips with attached cells were mounted in a Teflon chamber (Bionique Testing Laboratories, Inc., Saranac Lake, NY) and incubated in DMEM containing 1 M fura-2/AM (Molecular Probes, Inc, Eugene, OR) for 15 min at 37°C and 5% CO 2 . The cells were then washed and bathed in a HEPES-buffered physiological saline solution (HPSS; 116 mM NaCl, 5.4 mM KCl, 0.8 mM MgSO 4 , 1.8 mM CaCl 2 , 20 mM HEPES, and 10 mM glucose) at room temperature for at least 20 min before Ca 2ϩ measurements were made. In experiments assessing the intraluminal ER [Ca 2ϩ ], cells were incubated with 10 M fura-F (K d ϭ 25 M (13)), a gift from Dr. Elizabeth Murphy (NIEHS, National Institutes of Health, Research Triangle Park, NC) in DMEM for 60 min at 37°C and 5% CO 2 in order to compartmentalize the dye maximally within the ER (as determined in preliminary experiments). Cells were subsequently washed and bathed in HPSS similarly to the fura-2-loaded cells.
Fluorescence Measurements-The fluorescence of the fura-2-or fura-F-loaded cells was monitored with a photomultiplier-based system, mounted on a Nikon Diaphot microscope equipped with a Nikon 40ϫ (1.3 NA) Neofluor objective. The fluorescence light source was provided by a Deltascan D101 (Photon Technology International, Monmouth Junction, NY) equipped with a light path chopper and dual excitation monochromators. The light path chopper enabled rapid interchange between two excitation wavelengths (340 and 380 nm), and a photo-* 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  multiplier tube monitored the emission fluorescence at 510 nm, selected by a barrier filter (Omega Optical Inc., Brattleboro, VT). All experiments were carried out at 25°C in a field of cells (four to eight cells/ field). Calibration and calculation of [Ca 2ϩ ] i were carried out as described previously (15). 45 Ca 2ϩ Studies-Cultures were grown to subconfluence in 6-well plates and treated with Me 2 SO or 1.6 M PMA for 60 min at 37°C in DMEM and 5% CO 2 . Cells were washed twice in HPSS and kept at 37°C in HPSS containing 1 mM Ca 2ϩ . Uptake was initiated by addition of 1 Ci/ml 45 Ca 2ϩ and was terminated by quickly aspirating the medium and washing with HPSS containing 1 mM Ca 2ϩ , followed by addition of 1 ml of 1 mM EDTA, pH 8.0, and 0.1% Triton X-100. Samples were harvested, and their 45 Ca 2ϩ content was measured by liquid scintillation spectroscopy.
Immunodetection of Calreticulin and IP 3 Receptor-Subconfluent cultures grown in 175-cm 2 flasks were treated with Me 2 SO or 1.6 M PMA for 60 min at 37°C in DMEM and 5% CO 2 . Cells were subsequently rinsed with TES buffer (10 mM TES, pH 7.4, 20 mM sucrose, 100 mM NaF, 15 mM EDTA, and 2 mM EGTA), scraped in 1 ml of ice-cold TES buffer (supplemented with 2 g/ml aprotinin and 0.5 g/ml leupeptin)/flask, and disrupted by sonication (20 pulses). The homogenates were subsequently diluted (1:2) with solubilizing Laemmli buffer (16). Aliquots from control or PMA-treated cells were subjected to 7% polyacrylamide slab gel electrophoresis (acrylamide/bisacrylamide ratio ϭ 37.5:1), with electrophoresis performed at 25°C at a constant voltage of 100 V. Electrophoretic mobility, as a function of molecular weight, was obtained by parallel electrophoresis of molecular weight standards. Proteins were subsequently transblotted to nitrocellulose paper according to Towbin et al. (17) at 150 mA for 10 -14 h. Efficacy of the transfer was monitored by Coomassie Blue staining of the gels. For immunodetection of calreticulin and the IP 3 receptor, a procedure based on chemiluminescence (ECL, Amersham Corp.) was used. A rabbit polyclonal anti-calreticulin antibody (Affinity Bioreagents, Inc.) and an affinitypurified goat anti-rat cerebellum IP 3 receptor antibody (a gift from Dr. Solomon Snyder) were used to detect calreticulin and the IP 3 receptor, respectively, in control and PMA-treated cells.
Fluorescence Microscopy of Rhodamine 6G-labeled Cells-For rhoda-mine 6G labeling, cells grown to subconfluence on coverslips were treated with Me 2 SO or 1.6 M PMA for 60 min at 37°C in DMEM and 5% CO 2 . They were subsequently washed with HPSS, mounted on Teflon chambers, and bathed in HPSS containing 1.8 mM Ca 2ϩ . 1 M rhodamine 6G was added to the medium, and 10 min later, cells were visualized. (Dye-dependent labeling of intracellular organelles was very rapid; within 5-10 min, a defined cytoplasmic tubular network was observed.) Statistical Analysis-All individually shown experiments (Ca 2ϩ tracings, immunoblots, or fluorescent stainings) are representative of at least three separate experiments; dose-response relationships and time courses depict the average of a minimum of three experiments subjected to analysis of variance. Results were considered statistically significant with p Ͻ 0.05.

RESULTS
NIH 3T3 fibroblasts were exposed to either 1.6 M PMA or vehicle (0.1% (v/v) Me 2 SO) for 1 h at 37°C. The cells were loaded with fura-2, and [Ca 2ϩ ] i was measured before and after addition of 2 M thapsigargin ( Fig. 1). In PMA-treated cells kept in nominally Ca 2ϩ -free HPSS buffer, the thapsigargininduced Ca 2ϩ peak was diminished in comparison with control cells; however, although this might be taken as indicative of depletion of Ca 2ϩ stores by PMA, addition of 1.8 mM Ca 2ϩ prior to thapsigargin did not evoke Ca 2ϩ influx in PMA-treated or control cells (Fig. 1A). Substantial Ca 2ϩ entry was seen only after the residual stored Ca 2ϩ was depleted with thapsigargin. Neither the magnitude nor the rate of the [Ca 2ϩ ] i rise associated with influx was significantly different from that in control cells ( Fig. 1A; data not shown). The action of PMA to produce an apparent reduction in releasable Ca 2ϩ did not result from redistribution of Ca 2ϩ to other sites because, in PMA-treated cells, addition of 10 M ionomycin (in the presence of 200 M EGTA) after thapsigargin did not induce any further Ca 2ϩ mobilization (Fig. 1B). Furthermore, PMA-dependent reduction in stored Ca 2ϩ was also observed when 10 M ionomycin (in the presence of 200 M EGTA) was used, instead of thapsigargin, to assess stored Ca 2ϩ (data not shown; see Fig. 10B). Thus, in NIH 3T3 cells, PMA significantly reduces the amount of sequestered Ca 2ϩ without any accompanying activation of Ca 2ϩ entry. Fig. 2A shows that the action of the phorbol ester occurred in a time-and temperature-dependent manner. Depletion of the thapsigargin-sensitive calcium stores was observed following PMA treatment at 37°C, but not with treatment at 25°C. At 37°C, Ca 2ϩ depletion by 1.6 M PMA appeared to be quite rapid and was maximal by 60 min. The effect of PMA on thapsigargin-released Ca 2ϩ was also dependent on PMA concentration (Fig. 2B); in cells incubated with the phorbol ester for 60 min at 37°C, half-maximal inhibition of the thapsigargin release of stored Ca 2ϩ required ϳ16 nM PMA (thapsigargininduced Ca 2ϩ release was 114 Ϯ 5 and 83 Ϯ 7 nM in control and 16 nM PMA-treated cells, respectively).
In the experiments shown in Fig. 2A, prior to and following addition of thapsigargin, 1.8 mM Ca 2ϩ was restored to the medium to assess calcium entry. Again, despite the time-dependent effects of PMA on Ca 2ϩ store capacity, Ca 2ϩ entry was never activated until thapsigargin was added; as discussed above for Fig. 1A, following the complete depletion of stores with thapsigargin, the extent of Ca 2ϩ entry was statistically indistinguishable in control and PMA-treated cells ( To confirm that the reduction in Ca 2ϩ storage caused by PMA resulted from activation of PKC, cells were treated under the same conditions used for PMA with oleoylacetylglycerol ( Fig. 3A). Oleoylacetylglycerol mimicked the effect of PMA in a concentration-dependent manner and was not additive with PMA, suggesting that both compounds acted via activation of PKC to promote Ca 2ϩ pool depletion. A similar result was achieved with the less potent phorbol ester phorbol dibutyrate (Fig. 3B). The effects of PMA were at least partially reversed by the specific protein kinase C inhibitor chelerythrine chloride That PMA induces a net loss of sequestered Ca 2ϩ was confirmed by measurement of 45 Ca 2ϩ uptake. Fig. 4 shows that steady-state 45 Ca 2ϩ accumulation in attached NIH 3T3 cells was significantly diminished by incubation with PMA. However, the initial rate of uptake and the rate of efflux after removal of 45 Ca 2ϩ from the medium appeared to be unaffected by phorbol ester treatment.
We next carried out experiments to assess differences in the levels of stored calcium by using a Ca 2ϩ -sensitive fluorescent indicator compartmentalized in the endoplasmic reticulum. Fura-F was used in these experiments because its K d for Ca 2ϩ is 25 M (13), in the range expected in the lumen of the endoplasmic reticulum. The prolonged incubation time with fura-F/AM at 37°C encourages compartmentalization of the dye into intracellular organelles (18). The signal from cytoplasmic noncompartmentalized dye was reduced or eliminated by addition of 20 mM MnCl 2 to the cells; this caused an increase in the fluorescence ratio as the relative contribution of dye sequestered in Mn 2ϩ -inaccessible compartments with elevated [Ca 2ϩ ] was increased (Fig. 5). This elevated fluorescence ratio results in large part from Ca 2ϩ in the endoplasmic reticulum because addition of thapsigargin reduces the ratio as Ca 2ϩ is released from endoplasmic reticulum stores. As shown in Fig.  5, in cells pretreated with either PMA or a 1 M concentration of the reversible Ca 2ϩ -ATPase inhibitor cyclopiazonic acid, the increase in ratio on addition of Mn 2ϩ and the decrease in ratio on addition of thapsigargin were substantially diminished as compared with control cells. These data suggest (but do not prove) that the average endoplasmic reticulum luminal [Ca 2ϩ ] is lower in cells treated with either PMA or cyclopiazonic acid. In each case, this could result from a reduction in the volume of the Ca 2ϩ -storing compartment relative to other sites of dye sequestration, from partial loss of Ca 2ϩ from all of the endoplasmic reticulum, or from a complete loss of Ca 2ϩ from a fraction of the total endoplasmic reticulum. 2 In NIH 3T3 cells, virtually all of the sequestered Ca 2ϩ appears to be stored in a thapsigargin-sensitive compartment because once Ca 2ϩ is released by thapsigargin, no further release of Ca 2ϩ by ionomycin is observed. However, because it appears that this Ca 2ϩ pool can be substantially reduced without affecting the entry of Ca 2ϩ , we next sought to determine whether the entire thapsigargin-sensitive Ca 2ϩ pool is homogeneously sensitive to IP 3 . As shown in Fig. 6, cells were microinjected with a 1 mM concentration (pipette concentration) of the nonhydrolyzable IP 3 analog (2,4,5)IP 3 in the absence of external Ca 2ϩ and treated subsequently with 2 M thapsigargin; (2,4,5)IP 3 appeared to cause the release of total stored Ca 2ϩ because application of 2 M thapsigargin was without any additional effect. Next, IP 3 was generated from activation of phospholipase C-␥ by a maximal concentration (100 ng/ml) of platelet-derived growth factor (PDGF) in control and PMA-treated cells (Fig. 7). In control cells, addition of PDGF in the absence of external Ca 2ϩ increased [Ca 2ϩ ] i , and following the return of [Ca 2ϩ ] i to base-line levels, additional Ca 2ϩ was 2 An obvious question is, does PMA treatment cause a release of Ca 2ϩ to the cytoplasm that is detectable in fura-2-loaded cells? Experiments of this nature were attempted, but with equivocal results; while a gradual rise in [Ca 2ϩ ] i was seen following addition of PMA to cells (at 37°C), this appeared to result, at least in part, from an unexpected PMA-induced redistribution of fura-2 into a high Ca 2ϩ cellular compartment, presumably the ER. (In all other experiments, cells were loaded with fura-2 after pretreatment, as described under "Materials and Methods"; this apparently does not result in dye redistribution.) The significance of this phenomenon is currently under investigation.

FIG. 3. The PMA-induced depletion of the thapsigargin-sensitive Ca 2؉ pool is reproduced by other activators of PKC. y axis values denote the peak [Ca 2ϩ
] i level in response to 2 M thapsigargin in Ca 2ϩ -deficient buffer following the different treatments. A, the PKC activator oleoylacetylglycerol (OAG) mimicked the PMA action in a concentration-dependent fashion and was not additive with PMA. B, similarly, activation of PKC by phorbol dibutyrate (PDBu) reproduced the PMA effect in a concentrationdependent way. No depletion occurred with the inactive 4␣-PMA (data not shown). Oleoylacetylglycerol and phorbol esters were applied for 1 h at 37°C. mobilized by subsequent application of 2 M thapsigargin. This indicates that the quantity of IP 3 produced in PDGF-stimulated cells is insufficient to empty the entire thapsigargin-and IP 3 -sensitive Ca 2ϩ store. In cells pretreated with PMA, PDGFinduced Ca 2ϩ release was significantly diminished, while the subsequent thapsigargin response was almost completely blunted (Fig. 7). (Note that there was always a substantial and highly variable delay preceding the rise in [Ca 2ϩ ] i following PDGF treatment, but this delay was not affected by treatment with PMA.) Together, the results from Figs. 6 and 7 suggest that PMA treatment does not selectively alter Ca 2ϩ storage in agonist-or IP 3 -insensitive pools and that all of the thapsigargin-sensitive stores are also sensitive to IP 3 .
Because the reduction in stored Ca 2ϩ by PMA treatment required up to 1 h for maximal effect, we considered the possibility that the total ER content of the cells may have diminished. Therefore, total homogenates from control or PMAtreated cells were subjected to Western blots for two ER markers, the IP 3 receptor and calreticulin, an ER resident protein. As shown in Fig. 8, the PMA treatment did not affect the density of either marker, suggesting that the Ca 2ϩ pool depletion elicited by PMA was not a consequence of loss of ER membrane.
Activation of PKC by phorbol esters is known to induce morphological changes due to cytoskeleton rearrangement in different cell types (19 -21). To determine if the activation of PKC by PMA promoted any significant alterations in cell morphology, control and PMA-treated cells were visualized by differential interference contrast microscopy (Fig. 9, A and B, respectively). The flattened, planar aspect of control cells, characteristic of naive fibroblasts, was changed to a rounded shape after a 60-min incubation with 1.6 M PMA, agreeing with previous observations that activation of PKC alters NIH 3T3 cell structure (22). To determine if the morphology of the ER was affected by the PMA treatment, control and PMA-treated cells were stained with the fluorescent dye rhodamine 6G to label the ER and were examined by confocal fluorescence microscopy (Fig. 9, C and D, respectively). In control cells, the fluorescent staining was spread, and a tubular network of organelles resembling the normal structure of stacks of ER was observed; on the other hand, in PMA-treated cells, the spread fluorescent labeling was lost and became punctate.
One possible explanation for the inability of PMA to activate calcium influx despite its ability to cause an apparent depletion of Ca 2ϩ stores would be that the extent of depletion by PMA was beneath some critical threshold amount of depletion required to activate influx. Alternatively, because PMA is known to inhibit capacitative calcium influx in some systems (10, 11, 23), such an effect might mask a small activation of influx due to Ca 2ϩ depletion. To deal with these issues, experiments were performed in which Ca 2ϩ pools were depleted to different extents by the use of submaximal concentrations of the reversible Ca 2ϩ -ATPase inhibitor cyclopiazonic acid in the absence of external Ca 2ϩ . Capacitative Ca 2ϩ entry was assessed by subsequent addition of 1.8 mM Ca 2ϩ . The cells were returned to Ca 2ϩ -free medium, and the remaining sequestered Ca 2ϩ was determined by addition of 10 M ionomycin ϩ 200 M EGTA. Fig. 10A illustrates the protocol; the upper and lower tracings are representative experiments in which 10 and 0.1 M cyclopiazonic acid were used, respectively. As shown in Fig. 10B, each of three concentrations of cyclopiazonic acid (0.5, 1.0, and 2.5 M) induced a graded (but nonlinear), submaximal (as compared with 2 M thapsigargin) depletion of intracellular stores, as indicated by a graded diminution in residual ionomycin-releasable Ca 2ϩ ; at all three concentrations of cyclopiazonic acid, this depletion was greater in the PMA-treated cells. Fig. 10C shows the data for Ca 2ϩ entry from these same experiments. Again, the three concentrations of cyclopiazonic acid induced a graded increase in Ca 2ϩ entry, and PMA did not inhibit Ca 2ϩ entry under any of the conditions examined. Rather, PMA significantly augmented Ca 2ϩ entry at the two lowest cyclopiazonic acid concentrations (0.5 and 1.0 M). This effect of PMA was not obtained with the inactive steroisomer 4␣-PMA (with 0.5 M cyclopiazonic acid; control entry, 30 Ϯ 12 nM, n ϭ 4; entry in cells treated with 1.5 M 4␣-PMA, 27 Ϯ 9 nM, n ϭ 4; p Ͼ 0.05). PMA treatment did not augment Ca 2ϩ entry due to 2.5 M cyclopiazonic acid, despite the fact that, at this concentration of cyclopiazonic acid, both Ca 2ϩ store depletion and Ca 2ϩ entry were submaximal, and Ca 2ϩ store depletion was significantly greater in the PMA-treated cells. Note that PMA treatment induced an apparent greater depletion of Ca 2ϩ stores than did 0.5 or 1.0 M cyclopiazonic acid, despite the fact that these concentrations of cyclopiazonic acid induced significant influx of Ca 2ϩ , while PMA did not.
These results indicate two effects of PMA on calcium homeostasis in NIH 3T3 cells. Minimal depletion of intracellular stores by cyclopiazonic acid activated significant Ca 2ϩ entry, while an apparently greater depletion by PMA alone did not. This indicates that the failure of PMA to activate entry is not simply due to an inability to release Ca 2ϩ to some threshold level. Also, PMA did not inhibit cyclopiazonic acid-induced Ca 2ϩ entry at any of the concentrations examined. Rather, in a restricted range of cyclopiazonic acid concentrations, entry was actually augmented by PMA. Therefore, we conclude that the effects of PMA on Ca 2ϩ storage are decidedly unlike those of the Ca 2ϩ -ATPase inhibitors. In addition, the ability of PMA to augment Ca 2ϩ entry in a specific range of pool depletion appears to indicate an action of protein kinase C on calcium entry that is independent of the action of PMA on intracellular Ca 2ϩ storage. DISCUSSION This study demonstrates that PMA treatment of NIH 3T3 cells results in a substantial loss of Ca 2ϩ from thapsigarginand IP 3 -sensitive Ca 2ϩ pools, but this apparent depletion of stored Ca 2ϩ does not lead to activation of capacitative Ca 2ϩ entry (Fig. 1). The effect of PMA on sequestered Ca 2ϩ is concentration-, time-, and temperature-dependent (Fig. 2) and appears to be mediated by PKC (Fig. 3); however, it may result from the activation of a temperature-dependent pathway initiated or regulated by PKC rather than from a direct action of PKC since activation of PKC by PMA has been observed at 25°C in several systems.
We have also found that PMA treatment reduces stored Ca 2ϩ in an epidermal cell line (A431) as well as in freshly isolated lacrimal cells (data not shown), suggesting that this phenomenon may be a general one. This suggestion is further supported by two recent reports on glioma C6 cells (24) and platelets (25), where it was shown that PMA treatment decreased thapsigargin-dependent Ca 2ϩ release, in agreement with our results. However, those data differ from our data in that the phorbol ester treatment also inhibited thapsigargin-activated Ca 2ϩ influx, no experiments were carried out to determine if Ca 2ϩ storage was affected, and the authors concluded that the effects on release and entry were in some way related (i.e. by increased Ca 2ϩ extrusion, for example). While our findings may not be directly applicable to these other systems, they may indicate that the effects of PMA on release of Ca 2ϩ and on entry are in fact not related. Thus, with the results of our work included, there now exist cell types in which PMA induces an inhibition of entry, with no effect on the size of the stores, an inhibition of entry as well as a diminution in stored Ca 2ϩ , and now a diminution in store Ca 2ϩ with no effect on entry.
The mechanism underlying the action of PKC on Ca 2ϩ storage in NIH 3T3 cells is not known. Possible Ca 2ϩ transport systems that could be modulated by the activated enzyme are the ER Ca 2ϩ -ATPase, the ER leak channel, the ER IP 3 receptor, or the plasma membrane Ca 2ϩ -ATPase. Although their activities have not been directly measured in this study, the initial unidirectional rates of 45 Ca 2ϩ uptake in intact cells (Fig.  4), a primary function of Ca 2ϩ -ATPase-dependent Ca 2ϩ sequestration, are not different in control versus PMA-treated fibroblasts; these data suggest that the activity of this Ca 2ϩ transporter is not regulated by PKC and thus cannot mediate the PMA action. The 45 Ca 2ϩ efflux rates (Fig. 4), a result of the combined activities of the leak channel and the plasma membrane Ca 2ϩ -ATPase, are the same in control and PMA-treated cells, suggesting that these Ca 2ϩ transport systems are not affected by PKC activation either. One possibility consistent with these observations is that a subfraction of the ER that is specifically PMA-sensitive is so completely emptied by PMA treatment that it does not contribute to the measured kinetics. Modulation of IP 3 receptor function may be a possibility to account for the ER Ca 2ϩ loss, but probably not through changes in IP 3 levels since phospholipase C activity is inhibited by PKC activation in these cells. Since the IP 3 receptor has been suggested to interact with ankyrin, a cytoskeletal protein (26,27), it is possible that PKC activation increases the permeability of this Ca 2ϩ channel through a cytoskeletal modulation resulting in Ca 2ϩ loss.
Our data suggest that activation of PKC in NIH 3T3 cells induces profound changes in the architecture of the ER, perhaps a fragmentation of that organelle (Fig. 9) without net loss of its membrane (Fig. 8). These results are supported by a FIG. 10. Relationship between store depletion by cyclopiazonic acid and Ca 2؉ entry in NIH 3T3 cells. Ca 2ϩ pools in NIH 3T3 cells were depleted to varying extents by the use of different concentrations of the Ca 2ϩ -ATPase inhibitor cyclopiazonic acid (CPA) or thapsigargin (TG) in nominally Ca 2ϩ -free medium. Capacitative Ca 2ϩ entry was assessed by extracellular addition of 1.8 mM Ca 2ϩ in the presence of cyclopiazonic acid; two washes in nominally free Ca 2ϩ mediumcontaining cyclopiazonic acid followed, and enough time was allowed for [Ca 2ϩ ] i to return to base-line levels. 10 M ionomycin (iono.) ϩ 200 M EGTA ϩ cyclopiazonic acid were then added to assess the residual ionomycin-releasable Ca 2ϩ pool. A shows representative tracings illustrating the protocol used; the upper and lower tracings are cells treated with 10 or 0.1 M cyclopiazonic acid, respectively. B depicts the relationship between cyclopiazonic acid concentration (or 2 M thapsigargin) and residual ionomycin-releasable calcium. C depicts the relationship between cyclopiazonic acid concentration (or 2 M thapsigargin) and capacitative Ca 2ϩ influx. For B and C, open bars are controls, and closed bars are cells treated with PMA (from three to six independent experiments). *, significantly different from control. recent report on the effects of phorbol ester on cell morphology and localization of overexpressed PKC isoforms in NIH 3T3 fibroblasts (22); after a 15-min treatment with 100 nM PMA, PKC-␣ localized in the cell periphery and accumulated in cell margins, but a portion of the activated PKC-␣ concentrated in punctate regions in the cytoplasm, near the nucleus, and the punctate labeling was confirmed by double staining to be in the ER. Interestingly, colocalization of PKC-␣ and ER proteins was seen even in the absence of PKC activation (22). Since PKC-␣ is expressed in wild-type NIH 3T3 cells (22,28), we can speculate that activation of this PKC isoform and its association with components of the ER are somehow causally related to the loss of sequestered Ca 2ϩ .
These results raise an important question: Are these modifications in ER structure a result or a cause of the Ca 2ϩ pool depletion promoted by PMA? An earlier study in starfish oocytes reported that sperm-induced fertilization or microinjection of IP 3 , maneuvers that deplete the ER of Ca 2ϩ , transiently fragmented the ER (29). Furthermore, it was suggested that, since the period when the ER was fragmented temporally correlated with the time when [Ca 2ϩ ] i was high, the fragmentation may have been caused by loss of Ca 2ϩ from the ER (29). Our findings suggesting that activation of PKC depletes Ca 2ϩ and causes ER fragmentation are in good agreement with that report, but the temporal relationship between these events remains to be established in NIH 3T3 cells.
Recent studies have suggested that activation of PKC may affect capacitative Ca 2ϩ entry by either stimulating or inhibiting its activity depending on the cell type (rat thyroid cells (8), human neutrophils (23), Xenopus oocytes (11), pancreatic cells (12), and RBL-2H3 cells (10)). Our data indicate that, in NIH 3T3 cells, PKC does not inhibit but rather augments capacitative calcium entry. Curiously, this augmentation only occurs with a minimal degree of calcium store depletion, as obtained with 0.5 and 1.0 M cyclopiazonic acid. While the depletion of Ca 2ϩ stores by PMA might be expected to contribute to capacitative Ca 2ϩ entry, this does not appear to be the mechanism by which PMA augments entry in this situation. PMA induces an apparent Ca 2ϩ store depletion that is greater than that due to 0.5 M cyclopiazonic acid, but does not activate entry, and with 2.5 M cyclopiazonic acid, there is a substantial increase in intracellular store depletion in the PMA-treated cells, but no additional Ca 2ϩ entry. Thus, the ability of PMA to deplete Ca 2ϩ stores and the ability of PMA to potentiate Ca 2ϩ entry are not well correlated, and it appears likely that the effect on Ca 2ϩ entry results from an additional site of action of PMA somewhere in the pathway signaling capacitative Ca 2ϩ entry.
In summary, this study provides a number of novel and potentially important insights regarding the regulation of Ca 2ϩ homeostasis and Ca 2ϩ signaling by protein kinase C. The data clearly establish that, in a variety of cell types, activation of the protein kinase C pathway has profound effects on cellular Ca 2ϩ storage. Yet, these changes in Ca 2ϩ storage do not impact in a major way on the ability of ER calcium stores to regulate capacitative calcium entry. Further research is necessary to understand the molecular and subcellular actions of protein kinase C in regulating Ca 2ϩ storage in NIH 3T3 and other cell types.