Protein Kinase C-induced Activation of a Ceramide/Protein Phosphatase 1 Pathway Leading to Dephosphorylation of p38 MAPK*

Recently we showed that, in human breast cancer cells, activation of protein kinase C by 4β-phorbol 12-myristate 13-acetate (PMA) produced ceramide formed from the salvage pathway (Becker, K. P., Kitatani, K., Idkowiak-Baldys, J., Bielawski, J., and Hannun, Y. A. (2005) J. Biol. Chem. 280, 2606-2612). In this study, we investigated intracellular signaling events mediated by this novel activated pathway of ceramide generation. PMA treatment resulted in transient activation of mitogen-activated protein kinases (ERK1/2, JNK1/2, and p38) followed by dephosphorylation/inactivation. Interestingly, fumonisin B1 (FB1), an inhibitor of the salvage pathway, attenuated loss of phosphorylation of p38, suggesting a role for ceramide in p38 dephosphorylation. This was confirmed by knock-down of longevity-assurance homologue 5, which partially suppressed the formation of C16-ceramide induced by PMA and increased the phosphorylation of p38. These results demonstrate a role for the salvage pathway in feedback inhibition of p38. To determine which protein phosphatases act in this pathway, specific knock-down of serine/threonine protein phosphatases was performed, and it was observed that knock-down of protein phosphatase 1 (PP1) catalytic subunits significantly increased p38 phosphorylation, suggesting activation of PP1 results in an inhibitory effect on p38. Moreover, PMA recruited PP1 catalytic subunits to mitochondria, and this was significantly suppressed by FB1. In addition, phospho-p38 resided in PMA-stimulated mitochondria. Upon PMA treatment, a mitochondria-enriched/purified fraction exhibited significant increases in C16-ceramide, a major ceramide specie, which was suppressed by FB1. Taken together, these data suggest that accumulation of C16-ceramide in mitochondria formed from the protein kinase C-dependent salvage pathway results at least in part from the action of longevity-assurance homologue 5, and the generated ceramide modulates the p38 cascade via PP1.

Several metabolic routes contribute to ceramide formation, and the two best studied involve activation of sphingomyelinases (17,18) or the de novo pathway (11,19). Recently, we also described activation of the "salvage pathway," which re-utilizes long-chain sphingoid bases (sphingosine and sphinganine) formed by degradation of (dihydro)ceramide or complex sphingolipids (15). The results showed that PMA, a protein kinase C (PKC) activator, stimulated the generation of ceramide, which was inhibited by treatment with fumonisin B1 (FB1) (20), a (dihydro)ceramide synthase inhibitor, but not by myriocin (21), an inhibitor of de novo synthesis. This distinguishes the de novo from the salvage pathway and suggests the utility of FB1 (versus myriocin) in elucidating biological functions of the salvage pathway-derived ceramide in PKC-dependent cell responses. Based on such approaches, ceramide formed from the salvage pathway has been shown to be involved in the inhibition of juxtanuclear translocation of PKC-␤II upon PMA treatment, implying a role of ceramide in intracellular signal transduction (15).
The activation of the salvage pathway by PKC suggested possible roles for ceramide in regulating some PKC-mediated responses (15). Activation of PKC results in phosphorylation and regulation of a myriad of substrates (32,33). Some of the best studied are members of the mitogen-activated protein kinases (MAPKs), which include c-Jun N-terminal kinases 1/2 (JNK1/2), extracellular signal-regulated kinase 1/2 (ERK1/2), and p38 (34,35). Ceramide has been reported to exert various effects on the MAPK cascades (22, 36 -43). For example, ceramide has been shown to activate kinase suppressor of Ras (22) and TAK1 (40), which act upstream of the MAPK cascades. In contrast, some studies showed that ceramide down-regulated activation of MAPKs (38,43). These considerations led us to investigate the role of ceramide, specifically generated from the salvage pathway, in regulating protein phosphorylation/dephosphorylation of members of the MAPK family.
In this study, we examined specific roles of ceramide formed from the salvage pathway in signal transduction in PMA-stimulated human breast cancer cells (MCF-7). The results demonstrate that 1) PMA activation of the salvage pathway contributes to selective increases in C 16 -ceramide, which significantly occurred in mitochondria and were at least partly mediated by the activity of longevity-assurance homologue 5 (Lass5), 2) one of the CAPPs, PP1, as well as p38 were relocalized to mitochondria where C 16 -ceramide was enriched, and 3) this ceramide/PP1 pathway functioned as a negative regulator of the p38 cascade.
Subcellular Fractionation by Differential Centrifugation and Mitochondrial Isolation by Continuous Centrifugation Using a Percoll Gradient-Cells were incubated in buffer containing 250 mM sucrose, 10 mM Hepes (pH 7.4), 1 mM EDTA, and 0.5 mM phenylmethylsulfonyl fluoride for 30 min on ice. Cells were disrupted by passage through a 25-gauge needle for 5 strokes and then were centrifuged at 1,000 ϫ g for 10 min, 10,000 ϫ g for 10 min, and 100,000 ϫ g for 60 min, for collection of the nuclear fraction, mitochondria-enriched heavy membrane fraction, and light membrane fraction, respectively. These membrane fractions were washed twice with the above buffer, and ceramide levels were measured. To isolate mitochondria, the washed heavy membrane fractions were further loaded onto a 30% Percoll solution and centrifuged at 100,000 ϫ g for 2 h. Each fraction (fraction numbers 1-12) was subjected to immunoblotting with voltage-dependent anion channel (porin) or calnexin to determine fractions of mitochondria or endoplasmic reticulum, respectively.
Western Blotting-Cells were washed three times with icecold phosphate-buffered saline (PBS) supplemented with Halt TM phosphatase inhibitor mixture and then lysed using Laemmli buffer. Proteins (10 g) were subjected to 10% SDS-PAGE. Proteins were electrophoretically transferred to nitrocellulose membranes, blocked with PBS/0.1% Tween 20 (PBS-T) containing 5% nonfat dried milk, washed with PBS-T, and incubated for 12 h with primary antibody in PBS-T containing 5% nonfat dried milk. The blots were washed with PBS-T and incubated with secondary antibody in PBS-T containing 5% nonfat dried milk. Detection was performed using enhanced chemiluminescence.
Quantitative Real-time PCR-One g of total RNA, isolated using an RNA isolation kit (Qiagen), was used in reverse transcription reactions as described (44). The resulting total cDNA was then used in the quantitative real-time PCR to measure the mRNA levels using TaqMan gene expression kit (Applied Biosystems) as described by the manufacturer with ABI 7300 Q-PCR system. The mRNA levels of rRNA were used as internal control.
Transfection with FLAG-tagged LASS5 Expression Vector-Cells, growing on glass coverslips, were transfected with 1 g of pCMVexSVneo-LASS5 plasmid DNA using an Effectine transfection kit (Qiagen) according to the manufacturer's instructions.
Direct Immunofluorescence-Cells, growing on glass coverslips, were fixed for 10 min at room temperature with 4% formaldehyde in PBS and washed with PBS. Next, 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 for PP1c, Lamin B, HSP60, phospho-p38, calreticulin, FLAG, or cytochrome c were diluted in PBS containing 2% human serum, and incubated for 90 min at room temperature. Samples were washed with PBS, and TRITC-or fluorescein isothiocyanate-conjugated anti-IgG antibodies were applied for 1 h in PBS containing 2% human serum. Confocal laser microscopy was performed using an LSM510 microscope (Carl Zeiss, New York).
LC/MS-Analysis of ceramide species in lipid extract was performed by LC/MS as described in Becker et al. (15).
Statistical Analysis-Comparison between two groups was carried out by unpaired or paired Student's t test.

Transient Activation of MAPKs and Ceramide-dependent
Inactivation upon PMA Treatment-Many reports have shown that PMA induces acute activation of MAPKs in a PKC-dependent manner (32,35,45); this activation is mostly transient, probably because of the action of various protein phosphatases.
To determine the contribution of ceramide selectively derived from the salvage pathway to the regulation of MAPKs, MCF-7 cells were initially treated with PMA for up to 8 h, and then MAPK activation was monitored by immunoblotting with antibodies to the phosphorylated/active forms of JNK1/2, ERK1/2, and p38. As shown in Fig. 1, activation of JNK1/2, ERK1/2, and p38 was observed 30 min after PMA treatment and was transient such that by 2 h there was loss of PMA-induced phosphorylation, indicating action of protein phosphatases. To determine if the ceramide generated in response to PMA participates in this dephosphorylation phase of MAPKs, ceramide production was inhibited by the (dihydro)ceramide synthase inhibitor, FB1. The results showed that FB1 enhanced p38 activation induced by PMA in a dose-dependent manner ( Fig. 2A), and densitometric analysis showed that the effect was statistically significant (Fig. 2B). In contrast, FB1 did not exert a significant effect on the phosphorylation status of JNK1/2 and ERK1/2 ( Fig. 2A). In support of these results, exogenous C 6 -ceramide was also found to exert an inhibitory effect on p38 activation induced by PMA (Fig. 2C).
Because ceramide synthesis through the salvage pathway presumably requires the involvement of members of the LASS family of (dihydro)ceramide synthases, which are inhibitable by FB1, we next evaluated their participation in PMA-induced ceramide formation and ceramide-dependent dephosphorylation of p38. Specific knock-down of the LASS family members, LASS1 (47) or LASS5 (48,49), was achieved using siRNAs. Each siRNA treatment was confirmed to decrease the levels of the respective mRNA (Fig. 3A). Knock-down of LASS1 or LASS5 evoked a slight increase in basal levels of phospho-p38. Importantly, phosphorylation of p38 upon PMA treatment was markedly increased at the indicated time points in LASS5 knock-down cells as compared with the use of scrambled RNA or LASS1 siRNA (Fig. 3B). It should be noted that LASS5 knock-down had no effects on either the phosphorylation of ERK1/2 or JNK1/2 (data FIGURE 1. Activation of JNK1/2, ERK1/2, and p38. MCF-7 cells were stimulated with 100 nM PMA for the indicated periods. Whole cell lysates were prepared and subjected to immunoblot analysis with anti-phospho-JNK1/2, anti-phospho-ERK1/2, and anti-phospho-p38 antibodies. Equal amounts of protein were loaded in each lane. The results are representative of two to three independent experiments. FIGURE 2. FB1 enhancement of p38 phosphorylation/activation upon PMA treatment. A, mCF-7 cells were pretreated with the indicated concentration of FB1 for 2 h and then stimulated with 100 nM PMA for 1 h. Whole cell lysates were prepared and subjected to immunoblot analysis with antibodies for specific active forms of JNK1/2, ERK1/2, and p38. B, cells were pretreated with 100 M FB1 for 2 h and then stimulated with 100 nM PMA for 1 h. Amounts of phospho-p38 were estimated by measuring the density of bands of phospho-p38 and expressed as arbitrary units. Equal amounts of protein were loaded in each lane. The data represent mean Ϯ S.E. of three values. *, significantly different compared with FB1-untreated control (p Ͻ 0.004). **, significantly different compared with PMA stimulation (p Ͻ 0.02). C, cells were pretreated with 20 M C 6 -ceramide for 2 h, and then stimulated with 100 nM PMA for 1 h. Whole cell lysates were prepared and subjected to immunoblot analysis with antibodies for specific active form of p38. not shown). In addition, the partial silencing of LASS5 resulted in a commensurate partial inhibition of the generation of C 16 -ceramide after PMA stimulation (Fig. 3C). Taken together, the results suggest that LASS5 is at least in part involved in the salvage pathway and that LASS5-mediated generation of ceramide then acts to enhance dephosphorylation of p38.
Involvement of the PP1 Serine/Threonine Protein Phosphatase in p38 Dephosphorylation-The above results suggested that endogenous ceramide might activate a protein phosphatase in the p38 response. The serine/threonine phosphatases PP1 and PP2A have been implicated as CAPPs in vitro and as mediators of actions of ceramide on protein dephosphorylation in cells (1,4,27). To assess their roles in the ceramide effect on p38, individual isoforms of the catalytic subunits of CAPPs were knocked down by treatment with specific siRNA. The specific effectiveness of the knock-down was confirmed by immunoblotting with antibodies for PP1c-␣, PP1c-␤, PP1c-␥, and PP2Ac (Fig. 4A). Next, the effects of the siRNA on the PMAinduced phosphorylation of the MAPKs were evaluated. Interestingly, knock-down of each of the PP1c isoforms resulted in mild enhancement of basal phosphorylation of p38, with PP1c-␤ showing the least effect (Fig. 4B). In PMA-stimulated cells, each of PP1c siRNAs displayed a significant elevation in p38 phosphorylation as well (Fig. 4B), implying that PP1c acts as a negative regulator of the p38 pathway. In contrast, knockdown of PP1c-␣ and PP1c-␤ did not show much effect on phosphorylation of JNK1/2 or ERK1/2 induced by PMA, but PP1c-␥ knock-down showed modest enhancement of phosphorylation of JNK1/2 and ERK1/2. In contradistinction to PP1 isoforms, knock-down of PP2Ac-␤ did not modulate the effects of PMA on the phosphorylation status of p38 or the other MAPKs (Fig. 4B). Taken together, these results show that PP1c regulates the dephosphorylation of p38 following PMA-induced phosphorylation.
Relocalization of PP1c to Mitochondria in Response to PMA-To investigate the spatial association of PP1c with PMA-induced increases in ceramide, the effects of PMA on the intracellular localization of PP1c were examined. For these studies, PP1c was detected by direct immunofluorescence confocal microscopy using an antibody (E9), which recognizes all three isoforms of PP1c (PP1c-␣, PP1c-␤, and PP1c-␥), and colocalization was determined using antibodies for Lamin B as a nuclear membrane marker or HSP60 as a mitochondrial marker. As shown in Fig. 5, PP1c was present initially in a diffuse cytosolic pattern, but became relocalized to the perinuclear region in response to 1 h of treatment with PMA. Treatment of cells with PMA caused partial colocalization of PP1c with Lamin B; however, the pattern was such that PP1c localized in a ring like pattern partly overlapping but surrounding Lamin B, suggesting extra-nuclear localization (Fig. 5A). On the other hand, PMA not only induced a significant degree of colocalization of most of PP1c with HSP60, but it also induced a redistribution of HSP60, a mitochondrial matrix protein, into the same perinuclear pattern of PP1c (Fig. 5B). Staining with calreticulin, an endoplasmic reticulum protein, exhibited tubules as well as ring-shape structure around the nucleus, but PMA failed to recruit PP1c to the tubular endoplasmic reticulum (Fig. 5C). Taken together, the results demonstrate dynamic regulation of PP1c localization and association with a perinuclear pool of mitochondria in PMA-stimulated MCF-7 cells.
Ceramide-dependent Relocalization of PP1c-To determine if ceramide formation is involved in the relocalization of PP1c, the effects of FB1 were examined, because under these conditions FB1 specifically inhibited ceramide formation in response to PMA. As shown in Fig. 6A, FB1 treatment significantly diminished the PMA-induced relocalization of PP1c, and the percentages of cells with perinuclear PP1c was reduced from 78 Ϯ 10% to 36 Ϯ 3% with FB1 treatment (Fig. 6B). Taken together, PP1c relocalization induced by PMA requires, at least in part, ceramide formation.
Localization of p38 and LASS5 upon PMA Treatment-The above results on the ceramide-dependent relocalization of PP1 and dephosphorylation of p38 suggested that p38 may also localize, at least in part, to mitochondria. Therefore, direct immunofluorescence was performed to determine the intracellular compartment in which active p38 localizes. In contrast to a small amount of phospho-p38 observed in unstimulated cells, PMA induced significant increases in phospho-p38, and the majority of the staining was colocalized with cytochrome c (Fig. 7A). Consistent with the above immunocytochemical analysis, phospho-p38 increased with PMA stimulation in heavy membrane fraction (Fig. 7B). Because the negative regulation of p38 by PP1 is controlled by ceramide synthesis at least in part through the action of LASS5, the localization of this enzyme was also investigated. Direct immunofluorescence was performed in MCF-7 cells transfected with an expression vector of FLAG-tagged LASS5. Consistent with a previous study regarding LASS5 localization (48), FLAG-LASS5 was detected on the nuclear envelope and as a reticular structure in the perinuclear and cytoplasmic regions in unstimulated cells (Fig. 7C). Interestingly, whereas PMA treatment did not affect LASS5 localization, it did induce clustering of mitochondria around the nuclear envelope in close proximity to LASS5 (Fig. 7C).

Increase of C 16 -Ceramide in Heavy and Light Membrane Fractions and in Mitochondria in PMA-stimulated MCF-7
Cells-Next, it became important to determine the spatial relation of ceramide generation to that of PP1/p38. Therefore, the intracellular distribution of ceramide and the specific species of ceramide involved were evaluated by LC/MS. The levels of ceramide in the nuclear fraction acquired by centrifugation at 1,000 ϫ g were not significantly altered in response to PMA (Fig. 8A), whereas significant increases in ceramide were observed in the heavy membrane fraction (mitochondria-enriched fraction) acquired by centrifugation at 10,000 ϫ g (Fig. 8B). A light membrane fraction containing microsomes acquired by centrifugation at 100,000 ϫ g also displayed a PMA-induced elevation of ceramide (Fig.  8C). Notably, there were hardly any increases in C 24 -ceramide in response to PMA. However, there were significant increases in dihydroC 16 -ceramide in all thee fractions, but  the absolute levels were significantly lower than those of C 16 -ceramide (Fig. 8). Interestingly, C 16 -ceramide in the heavy membrane fraction displayed the highest concentration relative to total protein (Fig. 8B). Moreover, PMA-induced elevation of total ceramide levels in whole cells, and the ceramide in both the light and heavy membrane fractions was diminished by pre-treatment with FB1, an inhibitor of (dihydro)ceramide synthase (Fig. 9A). As shown in Fig. 9B, the specific increases in C 16 -ceramide in the heavy membrane fraction in response to PMA stimulation were significantly diminished by treatment with FB1. In contrast, PMA had no effects on C 24 -ceramide levels, and FB1 reduced the basal levels of this ceramide in total homogenate as well as in most fractions. Taken together, PMA caused a specific accumulation of C 16 -ceramide observed in both light and heavy membranes but not in the nuclear fraction.
To determine ceramide levels in mitochondria specifically, the heavy membrane fraction was subjected to Percoll centrifugation, and ceramide levels in the isolated mitochondria were determined. As expected, the heavy membrane fraction acquired by centrifugation at 10,000 ϫ g contained porin, a protein highly enriched in mitochondria and had some contamination with calnexin, an endoplasmic reticulum protein (Fig. 10A). Further centrifugation of the heavy membrane fractions using a 30% Percoll solution separated mitochondria from endoplasmic reticulum (Fig. 10B). Mitochondria acquired from three fractions (fraction number 8 -10) were subjected to LC/MS analysis. As shown in Fig. 10C, PMA induced a signifi-   Localization of p38 and LASS5. A, MCF-7 cells were treated with or without 100 nM PMA for 30 min, and cells were then fixed and immunostained with antibodies raised against phospho-p38 (green) and cytochrome c (red). The imaging was performed by confocal microscopy. B, MCF-7 cells were stimulated with 100 nM PMA for 30 min. Biochemical fractionation by differential centrifugation was performed as described under "Experimental Procedures." Heavy membrane fractions were subjected to immunoblot analysis with antibodies specific for phospho-p38 or porin. Equal amounts of protein were loaded in each lane. The results are representative of two independent experiments. C, MCF-7 cells were transfected with 1 g of pCMVexSVneo-LASS5 for 24 h followed by treatment with or without 100 nM PMA for 1 h. Fixed cells were immunostained with antibodies raised against FLAG (red) and HSP-60 (green). The imaging was performed by confocal microscopy.
cant elevation of C 16 -ceramide of 1.4-fold as compared with control mitochondria.

DISCUSSION
In this study, we demonstrate that the generation of C 16 -ceramide via the salvage pathway resulted in dephosphorylation of p38 through the action of PP1c isoforms. Moreover, the results revealed that LASS5 at least partly contributed to ceramide synthesis, and in turn LASS5 regulated ceramide-and PP1c-dependent dephosphorylation of p38. This study has implications for the subcellular localization of ceramide action, the role of protein phosphatases, and the emerging functions of the regulated salvage pathway of ceramide generation.
Our recent study demonstrated that the sphingoid base salvage represented the predominant pathway of induction of ceramide synthesis in response to PMA stimulation of MCF-7 cells (15). The salvage pathway most likely operates by salvaging sphingoid bases arising in lysosomes from hydrolysis of complex sphingolipids. Presumably, the sphingoid bases can traverse lysosomes and then re-enter sphingolipid biosynthetic pathways via ceramide synthases, probably LASS family members (52)(53)(54).
The results from this study demonstrate that the re-synthesis of ceramide in response to PMA leads to partial accumulation of ceramide in (or close to) mitochondria (e.g. in mitochondria-associated membranes (MAMs)). Thus, FB1-sensitive elevation of C 16 -ceramide was observed in the mitochondria-enriched heavy membrane fraction (Fig. 9). In addition, C 16 -ceramide accumulated in the mitochondrial fraction acquired from continuous centrifugation using a Percoll gradient (Fig. 10). Interestingly, ceramide synthase activity (55) has been observed in MAMs where free sphingoid bases are converted to (dihydro)ceramide by (dihydro)ceramide synthases. Moreover, ceramide-enriched mitochondria upon PMA treatment were seen to local-    (47,48). These spatial changes in subcellular compartments (mitochondria, endoplasmic reticulum, and MAMs) may play a key role in the accumulation of C 16 -ceramide in mitochondria and/or MAMs.
The results from this study also begin to implicate a specific (dihydro)ceramide synthase in the salvage pathway in response to PMA stimulation. LASS1 and LASS5 are known to preferentially produce C 18 -ceramide and C 16 -ceramide, respectively (47,48). Thus, inhibition of C 16 -ceramide formation by knock down of LASS5 (Fig. 3C) is likely to indicate a link of LASS5 to the salvage pathway. Therefore, the salvage pathway is suggested to at least in part contribute to C 16 -ceramide production in mitochondria. Notably, the results also showed elevation of ceramide in non-mitochondrial fractions, but not in nuclei (Figs. 8 and 9).
The subcellular location of ceramide generation (56 -59) is likely to play an important role in dictating its downstream targets and thereby the biological response of the cell to this mediator. As shown in Figs. 2 and 6, inhibition of ceramide accumulation by FB1 reversed the PMA-induced relocalization of PP1 to mitochondria and enhanced p38 activation, strongly indicating the potential of ceramide to mediate PP1 relocalization to mitochondria and subsequent effects on p38. Notably, the ceramide signaling is unlikely to link to the pathways for ERK1/2 or JNK1/2. In addition to pharmacological approaches using FB1 (Fig.  2), inhibition of C 16 -ceramide synthesis by LASS5 knock down upon PMA treatment resulted in enhancement of p38 phosphorylation (Fig. 3). These approaches targeting (dihydro)ceramide synthases, the LASS family members, strongly suggest an involvement of LASS5 in the ceramide signaling capable of influencing PP1 action and subsequent p38 dephosphorylation.
Interestingly, phorbol esters were shown to stimulate PP1 activity (60), implying an association of PKC activation with PP1 action. Further, a potent inhibitor of CAPPs (both PP1 and PP2A), okadaic acid, has been shown to increase p38 activation (61,62), which supports the ceramide/PP1-dependent dephosphorylation of p38 demonstrated in this study. Although studies on the involvement of PP1 in the p38 pathway have been restricted because of the lack of a PP1 isoform-specific inhibitor, knock-down by siRNA specific for PP1c isoforms shown in this study (Fig. 4) was able to support specific roles of PP1 isoforms in p38 dephosphorylation. However, further investigation is required for determining if the ceramide/PP1 pathway inactivates p38 directly or indirectly by acting on upstream kinases that regulate p38 phosphorylation.
In this context of mitochondrial action of ceramide, it is noteworthy that multiple studies have demonstrated strong links between ceramide and mitochondria in the regulation of cell death. Selective hydrolysis of a mitochondrial pool of sphingomyelin by bacterial sphingomyelinase targeted to the mitochondrial matrix resulted in apoptosis, whereas production of ceramide in the plasma membrane, endoplasmic reticulum, nucleus, and Golgi apparatus by bacterial sphingomyelinase targeted to these compartments exerted little effect on cell viability (63). Mitochondrial ceramide has also been proposed to play roles in mediating cell responses to UV irradiation (57) or tumor necrosis factor-␣ (56). Mechanistically, accumulation of mitochondrial ceramide was shown to trigger relocalization of Bax to the mitochondrion and subsequent cell death (56). Moreover, D609 (an inhibitor for sphingomyelin synthase) inhibition of UV-induced mitochondrial ceramide generation significantly prevented both disruption of mitochondrial transmembrane potential and release of cytochrome c from mitochondria, which resulted in suppression of apoptosis (57). Another recent study showed that targeting positively charged A, MCF-7 cells were harvested and subjected to biochemical fractionation. Cell lysates were centrifuged at 1,000 ϫ g for 10 min and 10,000 ϫ g for 10 min for collection of supernatant (Sup) of 1,000 ϫ g and 10,000 ϫ g, and the heavy membrane fraction (HM). These fractions were subjected to immunoblotting with antibodies specific for calnexin and porin. B, the heavy membrane fraction was further loaded onto a 30% Percoll solution and centrifuged as described under "Experimental Procedures." Twelve fractions were acquired, and the fractions enriched in mitochondria and endoplasmic reticulum were determined by immunoblotting with antibodies specific for calnexin and porin. C, cells were stimulated with (PMA) or without (Control) 100 nM PMA for 1 h, and then subjected to mitochondrial isolation. The levels of individual ceramide species in mitochondria acquired from three fractions (fraction numbers 8 -10) were determined as described under "Experimental Procedures" by LC/MS. C14-Cer, C 14 -ceramide; C16-Cer, C 16 -ceramide; C18-Cer, C 18 -ceramide; C18:1-Cer, C 18:1ceramide; C24-Cer, C 24 -ceramide; C24:1, C 24:1 -ceramide. The data represent mean Ϯ S.E. of three values.
ceramides to the mitochondrion resulted in reduction of cell viability of HepG cell and MCF-7 cells (64). Moreover, mitochondrial ceramides have been shown to not only inhibit mitochondrial respiratory chain complex III (65) but also stimulate formation of ceramide channels at the outer mitochondrial membranes that leads to increase the permeability of the mitochondrial outer membranes to small proteins (66,67). Thereby, the mitochondrial ceramide is likely to act as a mediator to modulate mitochondrial functions. On the other hand, a recent study showed activation of p38 by relatively high concentrations of ceramide, and this was present in the mitochondrial fraction, mediating loss of mitochondrial transmembrane potential (68). These recent studies imply not only a local action of ceramide in mitochondria but also the ability of mitochondrial ceramide to initiate/modulate cell death signals. In this study, PMA failed to induce apoptotic events, including lack of effects on release of cytochrome c and translocation of Bax to mitochondria (data not shown); however, mitochondrial ceramide played an important role as a negative regulator of p38 though PP1 recruitment to mitochondria.
Indeed, several studies have observed multiple effects of ceramides on the phosphorylation and activation of ERK1/2, JNK1/2, or p38 (39, 42, 68 -71). In contrast to the ability of ceramide to potentiate these kinase cascades, exogenous ceramide was reported to suppress not only p38 phosphorylation stimulated with cross-linking of Fc⑀RI receptor (38) but also lipopolysaccharide activation of p38 (43). In light of our present study, inhibition of p38 by ceramide might be due to ceramidedependent activation of PP1.
In the present study, we propose that LASS5 not only contributes to the synthesis of C 16 -ceramide through the salvage pathway but also subsequently to the ceramide signal resulting in PP1-dependent regulation of p38 (Fig. 11).
In summary, the present study provides evidence that ceramide formation is pathway-specific and compartment-specific. In mitochondria, ceramide exhibits the potential to drive PP1 to the mitochondria as well as to cause mitochondrial clustering around the nucleus. Activation of PP1 results in an inhibitory effect on p38 cascade; thus, demonstrating intricate crosstalk of the PKC, phosphatase, and MAPK pathways.