Selective hydrolysis of a mitochondrial pool of sphingomyelin induces apoptosis.

Our previous results have indicated that the major cellular pool of sphingomyelin present on the outer leaflet of the plasma membrane is not involved in the ceramide pathway of apoptosis. Thus, in this study we aimed at defining which intracellular pools of sphingomyelin and ceramide are involved in cell death. The bacterial sphingomyelinase (SMase) gene fused with green fluorescent protein was subcloned into mammalian vectors containing sequences that target the fusion proteins to cytoplasm, plasma membrane, mitochondria, Golgi apparatus, endoplasmic reticulum, or nucleus. Transfection of MCF7 breast cancer cells showed for all constructs an increase in SMase activity ranging from 2- to 60-fold, concomitant with an increase in total cellular ceramide levels (10-100%) as compared with vector-transfected cells. Next, the effect of overexpression of the SMase on cell death was examined. Results demonstrate that only when bacterial SMase was targeted to mitochondria did cells undergo apoptosis; its targeting to the other intracellular compartments was ineffective. Further, the results show that apoptosis induced by mitochondrial targeting of bacterial SMase requires SMase catalytic activity, is prevented by the overexpression of Bcl-2, and is mediated by inducing cytochrome c release. These results demonstrate that ceramide induces cell death specifically when generated in mitochondria. The results highlight the significance of compartment-specific lipid-mediated cell regulation, and they offer a novel general approach for these studies.

Structural requirements of ceramide and sphingosine based inhibitors of mitochondrial ceramidase.

The effects of structural analogues of ceramide on rat brain mitochondrial ceramidase (mt-CDase) were investigated. Design of target compounds was mainly based on modifications of the key elements in ceramide and sphingosine, including stereochemistry, the primary and secondary hydroxyl groups, the trans double bond in the sphingosine backbone, and the amide bond. Mt-CDase was inhibited by (1) all stereoisomers of D-erythro-ceramide (D-e-Cer) with an IC50 of 0.11, 0.21, and 0.26 mol % for the L-threo, D-threo, and L-erythro isomers, respectively; (2) all stereoisomers of sphingosine with IC50 ranging from 0.04 to 0.14 mol %, N-methyl-D-erythro-sphingosine (N-Me-Sph, IC50 0.13 mol %); and (3) D-erythro-urea-C16-ceramide (C16-urea-Cer IC50 0.33 mol %). The enzyme was not inhibited by N-methyl ceramide (N-Me-C16-Cer), 1-O-methyl ceramide (1-O-Me-C16-Cer), 3-O-methyl ceramide (3-O-Me-C16-Cer), cis-D-erythro ceramide (cis-D-e-C16-Cer) and 3-O-methyl-D-erythro-sphingosine (3-O-Me-Sph). It was less potently inhibited by D-erythro-sphinganine (D-e-dh-Sph, IC50 0.20 mol %), D-erythro-dehydro sphingosine (D-e-deh-Sph, IC50 0.25 mol %), (2S)-3-keto-sphinganine (3-keto-dh-Sph, IC50 0.34 mol %), (2S) 3-keto-ceramide (3-keto-C16-Cer, IC50 0.60 mol %), and ceramine (C18-ceramine, IC50 0.62 mol %), 1-O-methyl-D-erythro-sphingosine (1-O-Me-Sph), cis-D-erythro-sphingosine (cis-D-e-Sph), (2S)-3-ketosphingosine (3-keto-Sph), (2S)-3-keto-dehyrosphingosine (3-keto-deh-Sph), and N,N-dimethyl-D-erythrosphingosine (N,N-diMe-Sph) were weak inhibitors whereas ceramide-1-phosphate (Cer-1-P) and sphingosine-1-phosphate (Sph-1-P) stimulated the enzyme. Thus, for inhibition, the enzyme requires the primary and secondary hydroxyl groups, the C4-C5 double bond, the trans configuration of this double bond, and the NH-protons from either the amide of ceramide or the amine of sphingosine. Therefore, these results provide important information on the requirements for ceramide-enzyme interaction, and they suggest that ligand interaction with the enzyme occurs in a high affinity low specificity manner, in contrast to catalysis which is highly specific for D-erythro-ceramide (D-e-Cer) but occurs with a lower affinity. In addition, this study identifies two competitive inhibitors of mt-CDase; urea-ceramide (C16-urea-Cer) and ceramine (C18-ceramine) that may be further developed and used to understand the mechanism of mt-CDase in vitro and in biologic responses.

Biochemical characterization of the reverse activity of rat brain ceramidase. A CoA-independent and fumonisin B1-insensitive ceramide synthase.

We have previously purified a membrane-bound ceramidase from rat brain and recently cloned the human homologue. We also observed that the same enzyme is able to catalyze the reverse reaction of ceramide synthesis. To obtain insight into the biochemistry of this enzyme, we characterized in this study this reverse activity. Using sphingosine and palmitic acid as substrates, the enzyme exhibited Michaelis-Menten kinetics; however, the enzyme did not utilize palmitoyl-CoA as substrate. Also, the activity was not inhibited in vitro and in cells by fumonisin B1, an inhibitor of the CoA-dependent ceramide synthase. The enzyme showed a narrow pH optimum in the neutral range, and there was very low activity in the alkaline range. Substrate specificity studies were performed, and the enzyme showed the highest activity with d-erythro-sphingosine (Km of 0.16 mol %, and Vmax of 0.3 micromol/min/mg), but d-erythro-dihydrosphingosine and the three unnatural stereoisomers of sphingosine were poor substrates. The specificity for the fatty acid was also studied, and the highest activity was observed for myristic acid with a Km of 1.7 mol % and a Vmax of 0.63 micromol/min/mg. Kinetic studies were performed to investigate the mechanism of the reaction, and Lineweaver-Burk plots indicated a sequential mechanism. Two competitive inhibitors of the two substrates were identified, l-erythro-sphingosine and myristaldehyde, and inhibition studies indicated that the reaction followed a random sequential mechanism. The effect of lipids were also tested. Most of these lipids showed moderate inhibition, whereas the effects of phosphatidic acid and cardiolipin were more potent with total inhibition at around 2.5-5 mol %. Paradoxically, cardiolipin stimulated ceramidase activity. These results define the biochemical characteristics of this reverse activity. The results are discussed in view of a possible regulation of this enzyme by the intracellular pH or by an interaction with cardiolipin and/or phosphatidic acid.

Molecular cloning and characterization of a human mitochondrial ceramidase.

We have recently purified a rat brain membrane-bound nonlysosomal ceramidase (El Bawab, S., Bielawska, A., and Y. A. Hannun (1999) J. Biol. Chem. 274, 27948-27955). Using peptide sequences obtained from the purified rat brain enzyme, we report here the cloning of the human isoform. The deduced amino acid sequence of the protein did not show any similarity with proteins of known function but was homologous to three putative proteins from Arabidospis thaliana, Mycobacterium tuberculosis, and Dictyostelium discoideum. Several blocks of amino acids were highly conserved in all of these proteins. Analysis of the protein sequence revealed the presence at the N terminus of a signal peptide followed by a putative myristoylation site and a putative mitochondrial targeting sequence. The predicted molecular mass was 84 kDa, and the isoelectric point was 6.69, in agreement with rat brain purified enzyme. Northern blot analysis of multiple human tissues showed the presence of a major band corresponding to a size of 3.5 kilobase. Analysis of this major band on the blot indicated that the enzyme is ubiquitously expressed with higher levels in kidney, skeletal muscle, and heart. The enzyme was then overexpressed in HEK 293 and MCF7 cells using the pcDNA3. 1/His-ceramidase construct, and ceramidase activity (at pH 9.5) increased by 50- and 12-fold, respectively. Next, the enzyme was characterized using lysate of overexpressing cells. The results confirmed that the enzyme catalyzes the hydrolysis of ceramide in the neutral alkaline range and is independent of cations. Finally, a green fluorescent protein-ceramidase fusion protein was constructed to investigate the localization of this enzyme. The results showed that the green fluorescent protein-ceramidase fusion protein presented a mitochondrial localization pattern and colocalized with mitochondrial specific probes. These results demonstrate that this novel ceramidase is a mitochondrial enzyme, and they suggest the existence of a topologically restricted pathways of sphingolipid metabolism.

Purification and characterization of a membrane-bound nonlysosomal ceramidase from rat brain.

We have purified a membrane bound ceramidase 22,300-fold to apparent homogeneity. The purification scheme included Triton X-100 extraction of membranes followed by Q-Sepharose, blue Sepharose, phenyl-Sepharose, and MonoS column chromatography. The purified enzyme showed an apparent molecular mass of 90 kDa as estimated by SDS-polyacrylamide gel electrophoresis under reducing conditions and 95 kDa by chromatography on Superose 12. Using C(16)-ceramide as substrate, the enzyme showed a broad pH optimum in the neutral to alkaline range. A mixed micelle assay was developed, and using Triton X-100/ceramide mixed micelles, the enzyme exhibited classical Michaelis-Menten kinetics, with a K(m) of 1.29 mol % and a V(max) of 4.4 micromol/min/mg. When dihydroceramide was used as substrate, these values were 3.84 mol % and 1.2 micromol/min/mg, respectively, indicating that the enzyme hydrolyzes ceramides preferentially. The activity of the purified ceramidase did not require cations, and it was inhibited by reducing agents. Phosphatidylcholine and sphingomyelin were without effect on the enzyme activity, whereas phosphatidic acid and phosphatidylserine stimulated the activity 3-fold. Sphingosine acted as a competitive inhibitor with an IC(50) of 5-10 microM. These results indicate that the purified enzyme is a novel ceramidase.

Relationships between phosphatidic acid and cyclic nucleotide phosphodiesterases in activated human blood mononuclear cells.

We have previously shown that mitogenic activation of human PBMC rapidly increases both the intracellular phosphatidic acid (PA) level and cyclic nucleotide phosphodiesterase (PDE) activity, with time-course responses, suggesting a causative relationship between the two events. PA also directly stimulated cAMP-PDE activity in acellular systems. Thus the mitogenic properties of PA night be due to its ability to lower the level of cAMP, a negative effector of lymphocyte activation, through PDE activation. In this study, human PBMC were stimulated either with the mitogenic lectin ConA, the anti-CD3 mAb OKT3, or the phorbol ester TPA. All three agonists increased the radiolabeled PA level and the PA mass in treated cells and simultaneously increased cytosolic and particulate cAMP- and cGMP-PDE activities, with significant positive correlations between PA accumulation and PDE activities. Furthermore, the ConA-induced PDE activation was dose-dependently reduced by treatment of PBMC with the diacylglycerol-kinase inhibitor R59022. This compound also dose-dependently lowered the PA level and inhibited the proliferative response to ConA. In addition, TPA-induced PDE activation was totally abolished by ethanol, which strongly reduced PA accumulation in response to the phorbol ester. These data suggest that PA increase may be linked to mitogen-induced PDE activation. Experiments performed in the presence of rolipram indicated that ConA and TPA stimulated both the rolipram-sensitive PDE4 and the rolipram-insensitive PDE activities, OKT3 being more active on PDE4. All three agonists stimulated the cGMP-specific PDE5. These results suggest that PA is an important component of the mechanisms that maintain a low level of cyclic nucleotides, which is a prerequisite for an optimal lymphoproliferative response.

Highly sensitive measurement of lipid molecular species from biological samples by fluorimetric detection coupled to high-performance liquid chromatography.

As the molecular species composition of glycerophospholipids provides more valuable information than the corresponding fatty acid composition, we have applied a fluorimetric detection (360 and 460 nm for excitation and emission wavelengths, respectively) of anthroyl derivatives of diradylglycerol species to minor phospholipid classes and subclasses from biological samples. Diacylglycerol species were obtained by phospholipase C treatment of phosphatidylcholine subclasses and phosphatidic acid extracted from rat thymocytes. Subpicomole measurements of molecular species from the minor subclass alkenylacylglycerophosphocholine could be achieved (e.g. 0.4 pmol of the 18:1/20:5 species). Such a sensitivity allowed study of the molecular species composition of another minor phospholipid, phosphatidic acid, and to evaluation of its alteration in mitogen-stimulated thymocytes as compared to unstimulated ones. Finally, we report that such a measurement is also applicable to other minor bioactive lipids with a hydroxyl group available, namely hydroxyeicosatetraenoates (HETEs), with a similar gain of sensitivity over conventional UV detection. Overall, these measurements, especially those of phospholipid molecular species, are sensitive, reliable and meaningful for precursor-product relationship between phospholipids.

Selective stimulation of a cAMP-specific phosphodiesterase (PDE4A5) isoform by phosphatidic acid molecular species endogenously formed in rat thymocytes.

We have previously reported that concanavalin A (ConA) stimulation of rat thymocytes induces an increase in the cellular phosphatidic acid mass as well as a change in its fatty acid composition. An increase in phosphodiesterase (PDE) activity, mostly due to cAMP-specific (PDE4) isoforms, has also been observed in thymocytes stimulated by ConA. Furthermore, phosphatidic acid was able to stimulate PDE4 activity in vitro. In the present study, cAMP levels have been shown to decrease upon ConA stimulation of thymocytes. Decreasing phosphatidic acid level using diacylglycerol kinase inhibitors induced a parallel decrease of the ConA-stimulated cAMP-specific PDE activity in these cells. Analyses of phosphatidic acid molecular species in cells stimulated for 5 min by ConA revealed a significant increase in 1-stearoyl-2-arachidonoyl-sn-glycerol-3-phosphate and a relative decrease in the other molecular species of phosphatidic acid, mainly species containing palmitate. On the other hand, phosphatidic acid extracted from ConA-stimulated cells activated more efficiently the recombinant PDE4A5 isoform in vitro, as compared to phosphatidic acid extracted from unstimulated cells. In addition, phosphatidic acid species containing unsaturated fatty acids were stimulatory, while those containing two saturated fatty acids had only a marginal effect on the enzyme activity. Taken together, these data suggest that the mitogenic stimulation of thymocytes is accompanied by the synthesis of peculiar phosphatidic acid molecular species able to activate a PDE4 isoform. This activation might be of physiological relevance since cAMP is a major negative effector of the mitogenic response.

Contribution of phosphoinositides and phosphatidylcholines to the production of phosphatidic acid upon concanavalin A stimulation of rat thymocytes.

Stimulation of rat thymocytes by concanavalin A (Con A) results in a very early increase of the cellular level of phosphatidic acid (PA), while that of diacylglycerol (DAG) was not affected. As the biological activity of PA is very likely to be determined by its molecular species composition, the present study aims to investigate the pathways leading to the production of PA in Con A-stimulated rat thymocytes. Prelabeling the cells with [3H]arachidonic acid, [3H]myristic acid, [3H]choline, or [14C]lysophosphatidylcholine allowed us to determine that PA is formed by both phosphoinositide (PIs) and phosphatidylcholine (PC) hydrolysis. We then investigated whether PA derived from PC was formed by phospholipase C (PLC) or phospholipase D (PLD) hydrolysis. In the presence of 1-butanol, the production of phosphatidylbutanol was only observed in tetradecanoyl phorbol acetate (TPA)-stimulated cells. The use of a specific PC phospholipase C inhibitor resulted in a decrease of Con A-stimulated PA production in cells labeled with [3H]myristate. When cells were labeled with [3H]choline, only TPA stimulation induced a release of labeled choline. All together, these experiments suggest that PA is originated from two phospholipid sources, predominantly PI via PLC hydrolysis and to a lesser extent PC, by PLC hydrolysis also. Molecular species analyses by reverse phase HPLC are in agreement with this hypothesis, as diacyl-GP molecular species composition is similar to that of diacyl-GPC and DAG in resting cells, but resembles that of diacyl-GPI in Con A-treated cells. Thus, in stimulated cells, the amount of 18:0/20:4 species doubled while those of saturated and monounsaturated species decreased.

Time-course changes in content and fatty acid composition of phosphatidic acid from rat thymocytes during concanavalin A stimulation.

Several studies have shown the potential role of phosphatidic acid (PA) as a second messenger in different cell types. Thus, PA has been shown to mimic physiological agonists leading to various cellular responses, such as neurotransmitter and hormone release, cell proliferation by modulating DNA or RNA synthesis, the expression of several proto-oncogenes and growth factors, and the stimulation of enzyme activities such as phospholipase C (PLC), protein kinases and cyclic AMP (cAMP) phosphodiesterase. Stimulation of [3H]arachidonate-labelled rat thymocytes with the mitogen lectin concanavalin A (con A) resulted in enhanced production of radiolabelled PA after only 5 min of activation. The radiolabelled PA increase corresponded to a real increase in PA mass as determined by GLC quantification of its fatty acid content. In the presence of ethanol (0.5%), formation of phosphatidylethanol was not observed after 5 min of con A activation. Pretreatment of cells with R 59022 (10 microM), a diacylglycerol (DAG) kinase inhibitor, showed an inhibition in the formation of radiolabelled PA and in PA mass. These results suggest that the PLC-DAG kinase may be the pathway for PA synthesis in the first minutes of mitogenic thymocyte activation. A detailed analysis of the fatty acid composition showed that the relative amount of unsaturated fatty acids was increased in PA from stimulated cells concomitantly with a decrease in saturated ones; in particular, arachidonic acid was increased approximately 2-fold only 2 min after con A addition whereas palmitic acid was decreased for the whole period investigated (20 min). These changes favour the hydolysis of phosphoinositides rather than phosphatidylcholines by PLC. As PA remains a minor phospholipid, these changes are unlikely to affect cell membrane fluidity; but PA being now well recognized as a potential second messenger, its increased content as well as its increased unsaturation in the fatty acyl moiety might modulate several signalling pathways or the activity of enzymes such as cyclic nucleotide phosphodiesterase, controlling in this way the cellular level of cAMP, a negative regulator of blastic transformation.