Mechanism of Inhibition of Sequestration of Protein Kinase C α/βII by Ceramide

Sustained activation of protein kinase C (PKC) isoenzymes α and βII leads to their translocation to a perinuclear region and to the formation of the pericentrion, a PKC-dependent subset of recycling endosomes. In MCF-7 human breast cancer cells, the action of the PKC activator 4β-phorbol-12-myristate-13-acetate (PMA) evokes ceramide formation, which in turn prevents PKCα/βII translocation to the pericentrion. In this study we investigated the mechanisms by which ceramide negatively regulates this translocation of PKCα/βII. Upon PMA treatment, HEK-293 cells displayed dual phosphorylation of PKCα/βII at carboxyl-terminal sites (Thr-638/641 and Ser-657/660), whereas in MCF-7 cells PKCα/βII were phosphorylated at Ser-657/660 but not Thr-638/641. Inhibition of ceramide synthesis by fumonisin B1 overcame the defect in PKC phosphorylation and restored translocation of PKCα/βII to the pericentrion. To determine the involvement of ceramide-activated protein phosphatases in PKC regulation, we employed small interference RNA to silence individual Ser/Thr protein phosphatases. Knockdown of isoforms α or β of the catalytic subunits of protein phosphatase 1 not only increased phosphorylation of PKCα/βII at Thr-638/641 but also restored PKCβII translocation to the pericentrion. Mutagenesis approaches in HEK-293 cells revealed that mutation of either Thr-641 or Ser-660 to Ala in PKCβII abolished sequestration of PKC, implying the indispensable roles of phosphorylation of PKCα/βII at those sites for their translocation to the pericentrion. Reciprocally, a point mutation of Thr-641 to Glu, which mimics phosphorylation, in PKCβII overcame the inhibitory effects of ceramide on PKC translocation in PMA-stimulated MCF-7 cells. Therefore, the results demonstrate a novel role for carboxyl-terminal phosphorylation of PKCα/βII in the translocation of PKC to the pericentrion, and they disclose specific regulation of PKC autophosphorylation by ceramide through the activation of specific isoforms of protein phosphatase 1.

amide from the salvage of free sphingoid bases, formed in turn from the breakdown of complex sphingolipids. Activation of this salvage pathway was implicated in preventing translocation of PKC to the pericentrion in response to PMA. The ability of ceramide to inhibit translocation of PKC to the pericentrion suggested that ceramide interfered with a key step involved in mediating the effects of PMA on PKC. Therefore, it became important to define the mechanism by which ceramide exerted its effects on PKC. Several enzymes have been shown to be activated in vitro by ceramide and have, thus, emerged as candidate transducers of ceramide action. These ceramide mediators include kinase suppressor of Ras (24), cathepsin D (25), PKC (26), and ceramide-activated protein phosphatases, which include protein phosphatase 1 (PP1) (27,28) and protein phosphatase 2A (PP2A) (29 -33). Moreover, multiple studies have shown that ceramide signaling results in dephosphorylation of key proteins such as the retinoblastoma proteins (20), Bcl-2 (34), PKC␣ (35), Akt (36), p38 (23,37), and SR proteins (38), and these substrates in turn mediate specific downstream functions of ceramide.
Moreover, exogenous ceramide has been shown to influence both PLD (39) and PKC (35,40) itself, raising the possibility that ceramide generated from the salvage pathway could target either or both of these to inhibit PKC translocation to the pericentrion. Indeed, several studies have shown that short-chain ceramides inhibit PLD both in vitro (39) and in cells (40,41). Exogenous ceramide has also been shown to induce dephosphorylation of PKC␣ on several phosphorylation sites including Thr-497 or Ser-657 (42). Given these considerations, it became important to determine the mechanisms by which PKCs translocate to the pericentrion and the mechanisms by which ceramide interferes with this process.
In this study we employed HEK-293 and MCF-7 cells and investigated the mechanisms by which ceramide evokes a negative feedback pathway for cPKC translocation to the pericentrion. The results demonstrate a novel role for carboxyl-terminal phosphorylation sites (autophosphorylation sites) of cPKCs (43)(44)(45) in the translocation of PKC to the pericentrion, and they disclose very specific effects of the salvage-generated ceramide on regulation of the phosphorylation of these sites through activation of members of the PP1 family of protein phosphatases.
Confocal Microcopy-Cells growing on glass coverslips were fixed for 10 min at room temperature with 4% paraformaldehyde in phosphate-buffered saline (PBS) and washed with PBS. Confocal laser microscopy was performed using an LSM510 microscope (Carl Zeiss, NY). Cells were counted as translocated to the pericentrion if GFP-PKCs was observed in the perinuclear region, and GFP-PKCs translocation was assessed by blinded quantification of confocal microscopy pictures (3 fields for each counted for an approximate sampling of 50 -100 cells for each determination).
Indirect Immunofluorescence-Transfected cells with GFPtagged PKCs were fixed for 10 min at room temperature with 4% formaldehyde in PBS and washed with PBS. Cells were treated for 10 min with 0.1% Triton X-100, washed with PBS, and blocked for 1 h with PBS containing 2% human serum. The primary antibodies specific for Rab11 or CD59 were diluted in PBS containing 2% human serum and incubated for 90 min at room temperature or overnight at 4°C. Samples were washed with PBS, and AlexaFluor488-or Alex-aFluor555-conjugated anti-IgG antibodies were applied for 1 h at room temperature in PBS containing 2% human serum. Confocal laser microscopy was performed using an LSM510 microscope (Carl Zeiss, NY).
Palmitate Labeling-For chase experiments, MCF-7 cells were labeled (24 -30 h) in 35-mm dishes with 2 Ci/ml [ 3 H]palmitate in RPMI 1640 supplemented with 10% FBS. Cells were then washed 3 times with PBS, and RPMI 1640 supplemented with 10% FBS was added. Cells were pretreated with 0.4% 1-butanol for 10 min followed by stimulation. The culture media were removed, and the cells were washed rapidly three times with ice-cold PBS. Total cellular lipids were extracted by the modified method of Bligh and Dyer (65). Lipids were separated by thin-layer chromatography on a silica gel G plate with ethyl acetate/iso-octane/acetic acid (9:5:2, v/v/v) as the developing system (6). The plates were sprayed with EN 3 HANCE Spray (PerkinElmer Life Sciences) to amplify the tritium signal and then exposed for autoradiography for 24 h. The area corresponding to phosphatidyl-butanol was scraped off, and the radioactivity was measured by liquid scintillation counting.
Western Blotting-Cells were washed three times with PBS supplemented with Halt tm phosphatase inhibitor mixture (Pierce) and then lysed using Laemmli buffer. The protein samples (20 g) were subjected to SDS-polyacrylamide gel electrophoresis. Proteins were electrophoretically transferred to nitrocellulose membranes, blocked with PBS, 0.1% Tween 20 containing 5% nonfat dried milk, washed with PBS-Tween, and incubated with rabbit polyclonal antibody for phospho-PKC␣/ ␤II (1:1000) or rabbit polyclonal antibody for PKC␣ (1:2000) in 0.1% Tween 20 containing 5% nonfat dried milk. The blots were washed with PBS-Tween and incubated with secondary antibody conjugated with horseradish peroxidase in PBS-Tween containing 5% nonfat dried milk. Detection was performed using enhanced chemiluminescence reagent.

Effects of Ceramide on PLD Activation and GFP-PKC␣ Translocation in PMA-stimulated MCF-7
Cells-PLD activation has been shown to be required for the translocation of PKC␣ and/or PKC␤II to the pericentrion in HEK-293 cells or COS-7 cells (5,7). In a previous study we showed that PKC␤II failed to translocate to the pericentrion in MCF-7 (6). This inhibition was mediated by ceramide generated from the salvage pathway. Because ceramide was previously shown to inhibit PLD activation (39 -41), we wondered whether the effects of ceramide on PKC in MCF-7 cells are due to inhibition of PLD. MCF-7 cells labeled with [ 3 H]palmitate were stimulated with PMA for 1 h in the presence of 0.4% 1-butanol, and then phosphatidylbutanol levels were determined as a measure of PLD activity. As shown in Fig. 1A, generation of phosphatidylbutanol was stimulated after PMA treatment. This stimulation was not affected by FB1. Under these conditions, FB1 inhibited ceramide formation and restored the translocation of GFP-PKC␤II (6). FB1 treatment also resulted in GFP-PKC␣ sequestration and colocalization with the Rab11-positive subset of the recycling compartment (Fig. 1B). These results show that 1) ceramide generated from the salvage pathway does not inhibit PLD and 2) the effects of ceramide on PKC translocation to the pericentrion are not due to inhibition of PLD.
PKC␣/␤II Phosphorylation upon PMA Treatment in MCF-7 and HEK-293 Cells-Because activation of PLD by PMA appears to be a direct result of the action of PMA and the interaction of PKC with PLD, the above results suggested that the site of ceramide action is independent of PLD activation and may be at the level of PKC itself. The association of PKC␣/␤II with membranes has been shown to be modulated by specific phosphorylation of residues Thr-638/641 or Ser-657/660 that have been identified as autophosphorylation sites (Thr-641 in PKC␤II corresponds to Thr-638 in PKC␣, whereas Ser-660 in PKC␤II corresponds to Ser-657 in PKC␣) (44,45,47,48). We, therefore, wondered if the translocation of PKC␣/␤II to the pericentrion might be modulated by phosphorylation. The effects of PMA on the phosphorylation status of PKC␣/␤II were evaluated in both MCF-7 cells and HEK-293 cells, since PMA induces PKC translocation to the pericentrion in HEK-293 cells but not in MCF-7 cells (3)(4)(5)(6). Those two cell lines were treated with 100 nM PMA, and then acute phosphorylation of PKC␣/␤II at both Thr-638/641 and Ser-657/660 was assessed using two distinct antibodies targeting those sites; one antibody against Thr-638/ 641 is specific for PKC␣/␤II, whereas the second antibody against Ser-657/660 detects phosphorylation on both conventional and novel PKC. Using the last antibody, we were able to detect phosphorylation of overexpressed GFP-PKC␣ and GFP-PKC␤II after PMA treatment in both cell lines (data not shown). This immunoreactivity does not negate phosphorylation of other PKC isoforms but clearly indicated phosphorylation of PKC␣/␤II at Ser-657/660 in those cell lines.
PMA treatment of HEK-293 and MCF-7 over the indicated time course induced acute phosphorylation of PKC␣/␤II at both Thr-638/641 and Ser-657/660 in HEK-293 cells. By contrast, PKC␣/␤II phosphorylation at Thr-638/641 was not seen in PMA-stimulated MCF-7 cells ( Fig. 2A). Because PKC␤ was not detected in MCF-7 (49,50), the antibody against phosphorylated PKC␣/␤II at Thr-638/641 is suggested mostly to detect phosphorylated PKC␣ at Thr-638 in this cell line. During the time course of these studies, PKC␣ levels in both cell lines did not change appreciably (Fig. 2B). Therefore, these results disclose a defect in the ability of PMA to induce phosphorylation at residue Thr-638/641 in MCF-7 cells.
Involvement of Ceramide in Regulating Phosphorylation of PKC␣/␤II-The above results raised the possibility that the defect in phosphorylation of PKC␣/␤II may be the result of ceramide generation (which is induced in MCF-7 cells but is not observed in HEK-293 cells) and that this phosphorylation may be required for translocation of PKC␣/␤II to the pericentrion. First, to address the potential role of ceramide in regulating PKC phosphorylation, the ability of FB1 to restore it was evaluated. Immunoblotting using an antibody specific against phospho-Thr-638/641 of PKC␣/␤II revealed that FB1 stimulated the phosphorylation of PKC␣/␤II at Thr-638/641 upon PMA treatment in MCF-7 cells (Fig. 3A). Moreover, densitometric analysis demonstrated significant effects of FB1 on PKC␣ phosphorylation. It should be noted that FB1 failed to increase the phosphorylation of PKC␣/␤II at another autophosphorylation site, Ser-657/660 (Fig. 3B). These results suggested that formation of ceramide from the salvage pathway, which is inhibitable by FB1, contributed specifically to the diminished phosphorylation of PKC␣/␤II at Thr-638/641 in MCF-7 cells.
To determine whether an increase in ceramide was sufficient to attenuate phosphorylation of PKC␣/␤II at Thr-638/641, HEK-293 cells were treated with bacterial sphingomyelinase or C 6 -ceramide, and the effects on PKC␣/␤II phosphorylation were examined. Bacterial sphingomyelinase treatment suppressed phosphorylation of PKC␣/␤II at Thr-638/641 induced by PMA in HEK-293 cells (Fig. 4A). Moreover, C 6 -ceramide also exerted an inhibitory effect on this phosphorylation (Fig.  4B). On the other hand, exogenously added C 6 -ceramide is hydrolyzed to sphingosine and then reacylated to form longchain ceramide (ceramide recycling pathway) (51). Blocking recycling of C 6 -ceramide to long chain ceramide by FB1 partially restored the inhibitory effect of C 6 -ceramide on the phosphorylation of PKC (Fig. 4C). Therefore, ceramide appears to be both necessary and sufficient for preventing/reversing the phosphorylation of Thr-638/641.
Regulation of PKC␣/␤II Phosphorylation by Specific Ser/Thr Protein Phosphatases in MCF-7 Cells-The above results raised the possibility that the action of ceramide may involve activation of ceramide-activated protein phosphatases that would then oppose the phosphorylation of PKC␣/␤II in response to PMA. At first, pharmacological approaches for inhibiting Ser/ Thr protein phosphatases (PP1 and PP2A) were employed to evaluate if ceramide-dependent prevention of PKC phosphorylation in MCF-7 cells could be overcome by inhibiting these protein phosphatases. Okadaic acid (52), an inhibitor for PP2A and PP1 with preference for PP2A, has been shown to potently inhibit PP2A activity as compared with tautomycin (53), which displays the opposite preference. As shown in Fig. 5A, treatment with tautomycin but not okadaic acid resulted in a significant increase in basal phosphorylation of PKC␣/␤II at Thr-638/641. Upon PMA stimulation, okadaic acid failed to restore the levels of PKC phosphorylation. These results implicated PP1, but not PP2A, in regulation of phosphorylation on Thr-638/641. Because PP1 is composed of several catalytic isoforms, it became important to dissect the effects of specific PP1 iso-  forms on PKC phosphorylation. Therefore, in a complementary approach, siRNAs directed against individual protein phosphatases were employed to determine whether specific protein phosphatases are required for preventing/reversing this phosphorylation. siRNAs directed toward each of the isoforms of PP1c (␣, ␤, and ␥) and toward PP2Ac-␤ were confirmed to specifically knock down the target protein (23). As shown in Fig.  5B, knockdown of PP1c-␣ and PP1c-␤ provoked an increase in the basal status of PKC␣/␤II phosphorylation at Thr-638/641 in MCF-7 cells, but siRNAs specific for PP1c-␥ and PP2Ac-␤ had no effects on this basal phosphorylation. Importantly, the silencing of PP1c-␣ and PP1c-␤, but not PP1c-␥, significantly enhanced phosphorylation of Thr-638/641 upon PMA treatment. It should be noted that the silencing of PP2Ac-␤ in some experiments was observed to exert a modest effect on restoring this phosphorylation upon PMA treatment (data not shown). Therefore, specific isoforms of PP1c, PP1c-␣, and PP1c-␤ are strongly suggested to mediate dephosphorylation of PKC␣/␤II at Thr-638/641 in MCF-7 cells.
Regulation of PKC␣ Sequestration to the Pericentrion by Specific Ser/ Thr Protein Phosphatases in MCF-7 Cells-Given the above results demonstrating the action of specific protein phosphatases in regulating dephosphorylation of Thr-638/641, it became important to address the question of the role of protein phosphatases in regulating the translocation of PKC␣/␤II to the pericentrion. Pretreatment of MCF-7 cells with okadaic acid at 10 nM partially restored translocation of GFP-PKC␣ to the pericentrion; however, tautomycin pretreatment displayed more potent and more significant effects on translocation than okadaic acid (Fig. 6A). These results suggested that PP1s are more likely the protein phosphatases involved in inhibiting PKC translocation.
To determine the specific contribution of PP1c isoforms or PP2Ac to inhibition of PKC␣ translocation to the pericentrion, siRNA was employed. Cells transfected with the indicated siRNAs were stimulated with PMA for 1 h after transfection with GFP-PKC␣. Representative photographs on the effects of siRNA on GFP-PKC␣ localization are shown in Fig.  6B, and the number of cells showing GFP-PKC␣ translocation to the pericentrion was counted as shown in Fig. 6C. As shown in Figs. 6, B and C, PMA was unable to induce translocation of GFP-PKC␣ to the pericentrion in MCF-7 cells treated with scrambled RNA. On the other hand, down-regulation of PP1c-␣ and PP1c-␤ enabled significant induction of GFP-PKC␣ translocation to the pericentrion. Interestingly silencing of either PP1c-␥ or PP2Ac failed to influence translocation of GFP-PKC␣ to the pericentrion. Taken together these data indicate that PP1c-␣ and PP1c-␤ are involved in inhibition of GFP-PKC␣ translocation to the pericentrion.

Requirement for Dual Phosphorylation of PKC␣/␤II at Thr-638/641 and Ser-657/660 for Translocation to the Pericentrion-
Taken together, the data shown above implicate protein phosphatases both in prevention of translocation to the pericentrion and in dephosphorylation of Thr-638/641 in PKC␣/␤II. Therefore, it became important to investigate the role of this phosphorylation in the translocation of PKC to the pericentrion. To address this question, both phosphorylation sites (Thr-641 or Ser-660) in GFP-PKC␤II were individually mutated to alanine  to mimic the non-phosphorylated protein, and then the effects of the mutations were evaluated. Both mutants of GFP-PKC␤II (T641A and S660A), which correspond to Thr-638 and Ser-657 in PKC␣, respectively, were localized in the cytoplasm in HEK-293 cells (Figs. 7, B and C), which was identical to wild type GFP-PKC␤II (Fig. 7A). Upon treatment of cells with PMA, both mutants translocated only to the plasma membrane but not to the pericentrion (Figs. 7, B and C). These results indicate a requirement for dual phosphorylation on both residues for the subsequent translocation of PKC to the pericentrion.
These results also suggested that failure of translocation of PKC␣/␤II in the MCF-7 cells might arise from loss of phosphorylation on Thr-638/641. To address this question, point mutations to glutamate, which mimics phosphorylation, T641E or S660E, were employed. Interestingly, whereas PMA failed to induce translocation of wild type GFP-PKC␤II to the pericentrion in MCF-7 cells (Fig. 8A), it was able to cause significant overall internalization of GFP-PKC␤II (T641E) into an endosome-like compartment, which in some cells was sequestrated to the pericentrion (Fig. 8B). In contrast, the S660E mutation displayed no translocation to the pericentrion upon PMA treatment (Fig. 8C); however, inhibition of ceramide synthesis by FB1 restored the defect in translocation of GFP-PKC␤II (S660E) to the pericentrion (Fig. 8D). These results provide spe-cific evidence on the role of the phosphorylation of PKC␣/␤II at Thr-638/641 in translocation to the pericentrion and that the defect in this phosphorylation in response to ceramide accounts at least in part for the lack of translocation in MCF-7 cells.
Sustained activation of PKC␣/␤II leads to their translocation to the pericentrion, resulting in the cPKCdependent sequestration of recycling components into this compartment. Recycling molecules including transferrin, GM1, and CD59 are sequestered to the pericentrion in a cPKC-dependent manner (3)(4)(5). To determine the consequences of inhibition of PKCdependent sequestration of recycling components by ceramide, the dynamic regulation of transferrin and CD59 by ceramide was evaluated. Treatment of HEK-293 with PMA led to sequestration of GFP-PKC␤II and AlexaFluor-tagged transferrin, and this effect was significantly blocked by pretreatment with C 6 -ceramide (Fig. 9A). In contrast, sequestration of transferrin was not observed in PMA-stimulated MCF-7 cells, but inhibition of ceramide synthesis by FB1 induced this sequestration (Fig. 9B). In addition, the dynamic regulation of another recycled molecule, CD59, by ceramide was examined. In HEK-293 cells, both CD59 and mCherry-tagged PKC␤II were sequestered to the pericentrion upon PMA treatment. This sequestration was also inhibited by C 6 -ceramide (Fig. 10A). Similar to the effects of ceramide on transferrin sequestration, inhibition of ceramide synthesis by FB1 stimulated sequestration of CD59 to the pericentrion in MCF-7 cells (Fig. 10B). Moreover, GM1 trafficking underwent similar ceramide regulation (data not shown). The above data show that generation of ceramide inhibits dynamic sequestration of recycling components observed upon sustained activation of PKC.

DISCUSSION
The results from this study reveal that in PMA-stimulated MCF-7 cells there was a defect in the ability of PKC␣/␤II to become phosphorylated on Thr-638/641. Moreover, phosphorylation on this carboxyl-terminal site of cPKC (as well as on Ser-657/660) was required for the translocation of PKC to the pericentrion. The defect was mediated by ceramide, generated specifically from the salvage pathway, and the ceramide effects were in turn mediated by specific isoforms of PP1c (␣ and ␤). These results, therefore, demonstrate that the defect in phosphorylation of PKC␣/␤II at Thr-638/641 in MCF-7 cells plays a key role in preventing the ability of PMA to induce PKC translocation to the pericentrion. These studies shed important light on the mechanisms involved in translocation of cPKC and on the operation of specific ceramide-mediated pathways of cell signaling. Importantly, ceramide-mediated signaling results in dephosphorylation of cPKC and counteracts PKC-dependent sequestration of recycling components.
The translocation of PKC to the pericentrion is also seen with agonists that cause sustained stimulation of PKC. In a previous report, platelet-derived growth factor-bb was shown to induce translocation of PKC␣ to the pericentrion (4). In our ongoing studies we find that in A549 cells expressing G-protein-coupled bradykinin receptor, bradykinin also induced PKC␤II translocation to the pericentrion and that short chain ceramide pretreatment reduced this translocation (data not shown).
Phosphorylation/dephosphorylation plays a central role in regulating the function of protein kinases (46,54). All PKC isoenzymes are known to undergo post-translational regulation with phosphorylation on specific sites (2). Autophosphoryla-  tion of PKC␣/␤II on Thr-638/641 and/or Ser-657/660 increases negative charge at those sites, which has been proposed to affect thermal stability (47,55,56), catalytic competence (44,45,56,57), and/or subcellular localization (57)(58)(59). Mutagenesis approaches in HEK-293 cells (Fig. 7) revealed that mutation of either of Thr-641 or Ser-660 to Ala in PKC␤II abolished translocation of cPKC to the pericentrion but not to the plasma membrane. Thus, PKC␣/␤II has to be phosphorylated on both sites for it to translocate to the pericentrion. Interestingly, Prevostel et al. (60) also showed that phorbol ester action induces the transport of phosphorylated/activated PKC␣ from the plasma membrane to endosomes which are likely the precursor vesicles that lead to the formation of the pericentrion. This process required PKC activity and catalytically competent PKC, consistent with our results (5). On the other hand, the same group had shown in an earlier study (61) that treatment of COS-7 cells with phorbol ester induced dephosphorylation of the kinase, and this preceded degradation of cPKC. These results fit the current emerging model whereby sequestration of cPKC in the pericentrion was proposed to protect cPKC from degradation (5). One could then predict that dephosphorylation (e.g. in response to PMA-induced ceramide) prevents localization to the pericentrion and may then enhance degradation.
On the other hand, the study by Hu and Exton (7) showed that treatment of COS-7 cells with phorbol ester induced phosphorylation of PKC␣ and translocation of the kinase to the peri-centrion. The reasons for the discrepancy between these two studies on the effects of phorbol ester on phosphorylation (7) versus dephosphorylation (61) are not clear but could be the result of effects on distinct phosphorylation sites or a result of acquired variation in the cell line (as we see in this study comparing MCF-7 and HEK-293).
In the former study (7) it was also observed that inhibitors of PKC (Ro-31-8220 and bisindolylmaleimide) increased basal phosphorylation but prevented the translocation to the pericentrion. The requirement for PKC activity for its translocation is consistent with our results. However, the effects on phosphorylation appear different from ours. It should be noted, however, that in that study (7) the observed phosphorylation is most likely not due to direct PKC activity (increased not inhibited by Ro-31-8220 and bisindolylmaleimide) and, therefore, must be distinct from the PKC-dependent autophosphorylation examined in the current study.
Inhibition of PP1c activities or knockdown of PP1c-␣ and PP1c-␤ isoforms restored cPKC translocation to the pericentrion and prevented the dephosphorylation (or poor phosphorylation) of PKC␣/␤II in PMA-stimulated MCF-7 cells. Therefore, the results suggest that either one or both PP1c isoforms (PP1c-␣ and PP1c-␤) mediate the action of ceramide on the dephosphorylation of cPKC on the autophosphorylation sites, resulting in prevention of cPKC translocation to pericentrion. Interestingly, in PMA-stimulated MCF-7 cells, mitochondrial ceramide-dependent dephosphorylation of p38 is mediated by all three isoforms of PP1c (␣, ␤, and ␥) (23). These studies suggest that the specific signaling effects of ceramide account for selective regulation of individual down-stream targets, possibly due to their spatial compartmentalization.
Exogenous ceramide (35) and ceramide endogenously formed in response to tumor necrosis factor-␣ (42) have been shown to cause dephosphorylation of PKC␣, especially on Thr-497 with resultant inactivation of PKC, consistent with a key role for phosphorylation of this residue in the catalytic competence of PKC (62,63). These effects of ceramide occurred via okadaic acid-sensitive Ser/Thr protein phosphatases, most likely PP2A. These results are distinct from the current ones which show selective effects of ceramide on phosphorylation of PKC␣ on Thr-638, which appear to be mediated by PP1c-␣ and PP1c-␤. These differences are likely to be due to the selective ability of PMA to induce ceramide via the salvage pathway. Thus, different pathways of ceramide generation (de novo versus sphingomyelinase versus salvage), which most likely occur in distinct subcellular compartments, launch distinct signaling pathways involving distinct mediators.
Newton and co-workers (44) reported that PP2A but not PP1 dephosphorylated PKC␤II on Ser-660, whereas the residue Thr-641 was selectively dephosphorylated by PP1. Consistently, our pharmacological approaches (Fig. 5A) also support the role of PP1c to dephosphorylate the residue Thr-638/641 of PKC␣/␤II. Moreover, genetic approaches (Fig. 5B) reveal the contribution of specific isoforms (␣ and ␤) of PP1c to this dephosphorylation. Although knockdown of PP1c-␥ failed to exert significant effects on PKC␣ dephosphorylation (Fig. 5B), the subcellular localization of PP1c-␥, which resides mostly in the nucleus or nucleoli (64), may restrict the role of PP1c-␥ to contribute to specific ceramide-dependent effects.
In addition, PLD activity is also involved in recruitment of cPKC to the pericentrion (5). Ceramide is known to inhibit PLD activation both directly or indirectly through unknown mechanisms (39 -41). However, salvage-generated ceramide had no effects on PMA-induced PLD activation (Fig. 1A). Moreover, perturbation of PLD signaling by 1-butanol had no effects on the autophosphorylation of PKC␣ in HEK-293 cells (data not shown). Therefore, it appears that ceramide formed from the salvage pathway is unable to inhibit PMA-induced activation of PLD, unlike the action of exogenous ceramides (39). Therefore, two independent pathways involved in cPKC relocalization to pericentrion are suggested to exist, PLD signaling and ceramide/PP1c signaling. Another interesting conclusion emerging from the autophosphorylation studies indicates that autophosphorylation of PKC may not be required for activation of PLD in response to PMA.
Sustained activation of cPKC results in a dynamic sequestration of recycling components. Ceramide is generated upon various stimuli, leading to initiation of ceramide signaling, with ceramide-activated protein phosphatases (PP1 and/or PP2A) functioning as effecter downstream molecules (23,28,34). The present study demonstrates that cPKC is a target for ceramide/ PP1c and that the dephosphorylation of cPKC by ceramide counteracts PKC-dependent sequestration of recycling compo- nents ( Figs. 9 and 10). Thus, ceramide formed from the salvage pathway is proposed to account for counteracting PKC-regulated cellular trafficking.
Taken together these results indicate that PKC␣/␤II has to be dually phosphorylated on Thr-638/641 and Ser-657/660 so as to allow translocation to the pericentrion. Also, the results suggest that ceramide generated upon PMA treatment in MCF-7 cells is involved in dephosphorylation of PKC␣/␤II and, thus, prevents its sequestration and sequestration of the PKCdependent subset of recycling components.