Ceramide kinase profiling by mass spectrometry reveals a conserved phosphorylation pattern downstream of the catalytic site.

Ceramide kinase (CERK) is essential for production of ceramide-1-phosphate (C1P), a bioactive lipid whose formation critically modulates ceramide levels. To explore how CERK is regulated, we used insect cell-expressed, recombinant hCERK and searched for post-translational modifications, using mass-spectrometry techniques. This led to identification of two phosphorylated serine residues, at positions 340 and 408. Point mutations preventing phosphorylation at either of these sites did not lead to detectable changes in subcellular localization or activity. However, preventing phosphorylation at S340 resulted in CERK instability as revealed by the behavior of the S340A mutant protein under various assay conditions in vitro. Phosphorylation of a cognate serine residue in sphingosine kinases was previously shown to be important. Therefore, phosphorylation within a conserved "regulation loop" downstream of the catalytic domain emerges as a new paradigm for regulation of kinases of the diacylglycerol kinase family. This "regulation loop" is reminiscent of the "activation loop" that controls AGC protein kinases, being a similar distance from the critical ATP binding site determinants in the primary sequence.

Subcellular localization of ceramide kinase and ceramide kinase-like protein requires interplay of their Pleckstrin Homology domain-containing N-terminal regions together with C-terminal domains.

Ceramide kinase (CERK) and the ceramide kinase-like protein (CERKL), two related members of the diacylglycerol kinase family, are ill-defined at the molecular level. In particular, what determines their distinctive subcellular localization is not well understood. Here we show that the Pleckstrin Homology (PH) domain of CERK, which is required for Golgi complex localization, can substitute for the N-terminal region of CERKL and allow for wild-type CERKL localization, which is typified by nucleolar accumulation. This demonstrates that determinants for localization of these two enzymes do not lie solely in their PH domain-containing N-terminal regions. Moreover, we present evidence for a previously unrecognized participation of CERK distal sequences in structural stability, localization and activity of the full-length protein. Progressive deletion of CERK and CERKL from the C-terminus revealed similar sequential organization in both proteins, with nuclear import signals in their N-terminal part, and nuclear export signals in their C-terminal part. Furthermore, mutagenesis of individual cysteine residues of a CERK-specific CXXXCXXC motif severely compromised both exportation of CERK from the nucleus and its association with the Golgi complex. Altogether, this work identifies conserved domains in CERK and CERKL as well as new determinants for their subcellular localization. It further suggests a nucleocytoplasmic shuttling mechanism for both proteins that may be defective in CERKL mutant proteins responsible for retinal degenerative diseases.

Modulation of ceramide metabolism in mouse primary macrophages.

Ceramide kinase (CERK) produces the bioactive lipid ceramide 1-phosphate (C1P) and is, together with glucosylceramide synthase (GCS) and sphingomyelin synthases (SMS-1 and -2), a key regulator of ceramide metabolism. Here, we used a previously validated assay for measuring CERK, GCS, and SMS activities simultaneously, to study the regulation of ceramide metabolism in mouse macrophages. Elicitation of peritoneal macrophages as well as differentiation of bone marrow-derived monocytes into macrophages led to "ceramide anabolic switching" by re-directing ceramide anabolism towards C1P synthesis by CERK. In contrast, macrophage activation by lipopolysaccharide (LPS) evoked a "ceramide anabolic switch" going in the opposite direction, i.e. featuring up-regulation of GCS and SMS and down-regulation of CERK. The LPS effects were partially blocked by dexamethasone, a known macrophage de-activator. Altogether, the data reveal a contrasting regulation of ceramide metabolism enzymes during macrophage biological responses.

Transcriptional repression of ceramide kinase in LPS-challenged macrophages.

Ceramide kinase (CERK) produces the bioactive lipid ceramide 1-phosphate (C1P). Both CERK and C1P have been identified as mediators of cell growth and survival. Recent evidence showed that CERK is down-regulated during M1-type macrophage activation, which is known to promote cell growth arrest. However, the mechanism has not been investigated yet and, in particular, whether growth arrest might be the signal for down-regulation of CERK is currently unknown. Here, we found that LPS-mediated TLR-4 engagement reduces Cerk mRNA levels in mouse primary macrophages. Reporter gene assays in RAW264.7 macrophages showed that LPS inhibits the transcriptional activity of the Cerk proximal promoter. The G1-cell cycle blocker mimosine did not inhibit Cerk transcription, suggesting that transcriptional repression of Cerk by LPS is not a primary consequence of LPS-induced cell cycle blockade.

A secondary assay for ceramide kinase inhibitors based on cell growth inhibition by short-chain ceramides.

We recently reported that ectopic expression of ceramide kinase (CerK) in various cell lines increases their sensitivity to cell death induced by the exogenous addition of short-chain (e.g., C2) ceramides (Cer). Here we show that this higher sensitivity results from CerK catalytic activity and production of C2-ceramide 1-phosphate (C2-C1P). If CerK activity is inhibited by the potent inhibitor NVP-231, C2-C1P is not produced and viability returns to control levels. The EC(50) of NVP-231 in this assay is in the low nanomolar range, consistent with the IC(50) determined in activity assays in vitro using purified CerK. NVP-995, a structurally related but inactive compound, does not protect against C2-Cer-induced cell death. This assay is robust and easy to implement and scale up, thereby providing a valuable secondary screen assay for CerK inhibitors.

Targeting ceramide metabolism with a potent and specific ceramide kinase inhibitor.

Ceramide kinase (CerK) produces the bioactive lipid ceramide-1-phosphate (C1P) and appears as a key enzyme for controlling ceramide levels. In this study, we discovered and characterized adamantane-1-carboxylic acid (2-benzoylamino-benzothiazol-6-yl)amide (NVP-231), a potent, specific, and reversible CerK inhibitor that competitively inhibits binding of ceramide to CerK. NVP-231 is active in the low nanomolar range on purified as well as cellular CerK and abrogates phosphorylation of ceramide, resulting in decreased endogenous C1P levels. When combined with another ceramide metabolizing inhibitor, such as tamoxifen, NVP-231 synergistically increased ceramide levels and reduced cell growth. Therefore, NVP-231 represents a novel and promising compound for controlling ceramide metabolism that may provide insight into CerK physiological function.

A conserved cysteine motif essential for ceramide kinase function.

Ceramide kinase (CerK) is a sphingolipid metabolizing enzyme very sensitive to oxidation; however, the determinants are unknown. We show here that the thiol-modifying agent N-ethyl-maleimide abrogates CerK activity in vitro and in a cell based assay, implying that important cysteine residues are accessible in purified as well as endogenous CerK. We replaced every 22 residues in human CerK, by an alanine, and measured activity in the resulting mutant proteins. This led to identification of a cluster of cysteines, C(347)XXXC(351)XXC(354), essential for CerK function. These findings are discussed based on homology modeling of the catalytic domain of CerK.

A critical beta6-beta7 loop in the pleckstrin homology domain of ceramide kinase.

CerK (ceramide kinase) produces ceramide 1-phosphate, a sphingophospholipid with recognized signalling properties. It localizes to the Golgi complex and fractionates essentially between detergent-soluble and -insoluble fractions; however, the determinants are unknown. Here, we made a detailed mutagenesis study of the N-terminal PH domain (pleckstrin homology domain) of CerK, based on modelling, and identified key positively charged amino acid residues within an unusual motif in the loop interconnecting beta-strands 6 and 7. These residues are critical for CerK membrane association and polyphosphoinositide binding and activity. Their mutagenesis results in increased thermolability, sensitivity to proteolysis, reduced apparent molecular mass as well as propensity of the recombinant mutant protein to aggregate, indicating that this loop impacts the overall conformation of the CerK protein. This is in contrast with most PH domains whose function strongly relies on charges located in the beta1-beta2 loop.

Neutropenia with impaired immune response to Streptococcus pneumoniae in ceramide kinase-deficient mice.

In mammals, ceramide kinase (CerK)-mediated phosphorylation of ceramide is the only known pathway to ceramide-1-phosphate (C1P), a recently identified signaling sphingolipid metabolite. To help delineate the roles of CerK and C1P, we knocked out the gene of CerK in BALB/c mice by homologous recombination. All in vitro as well as cell-based assays indicated that CerK activity is completely abolished in Cerk-/- mice. Labeling with radioactive orthophosphate showed a profound reduction in the levels of de novo C1P formed in Cerk-/- macrophages. Consistently, mass spectrometry analysis revealed a major contribution of CerK to the formation of C16-C1P. However, the significant residual C1P levels in Cerk-/- animals indicate that alternative routes to C1P exist. Furthermore, serum levels of proapoptotic ceramide in these animals were significantly increased while levels of dihydroceramide as the biosynthetic precursor were reduced. Previous literature pointed to a role of CerK or C1P in innate immune cell function. Using a variety of mechanistic and disease models, as well as primary cells, we found that macrophage- and mast cell-dependent readouts are barely affected in the absence of CerK. However, the number of neutrophils was strikingly reduced in blood and spleen of Cerk-/- animals. When tested in a model of fulminant pneumonia, Cerk-/- animals developed a more severe disease, lending support to a defect in neutrophil homeostasis following CerK ablation. These results identify ceramide kinase as a key regulator of C1P, dihydroceramide and ceramide levels, with important implications for neutrophil homeostasis and innate immunity regulation.

Enhanced ceramide-induced apoptosis in ceramide kinase overexpressing cells.

We evaluated how increased levels of ceramide kinase (CerK) would impact the growth of COS-1 fibroblasts and RBL-2H3 basophils. The low CerK activity in these cells was strongly up-regulated upon recombinant expression of CerK. CerK-overexpressing COS-1 cells depended on higher concentrations of serum for their growth and displayed many filipodia. The two CerK-overexpressing cell lines were more sensitive to C2-ceramide-mediated apoptosis, and this correlated with the production of C2-ceramide-1-phosphate by CerK. This study indicates that ceramide kinase may participate in the control of cell growth, and establishes a novel assay that will be valuable for testing ceramide kinase inhibitors.

Effects of ceramide-1-phosphate on cultured cells: dependence on dodecane in the vehicle.

Ceramide-1-phosphate (C1P), the product of ceramide kinase, is a sphingophospholipid with recently recognized signaling properties. In particular, it was reported to be mitogenic and capable of direct stimulation of cytosolic phospholipase A(2alpha). Much of the present knowledge has relied on the use of C1P of various acyl chain lengths, together with diverse protocols to deliver it to cultured cells. A mixture of ethanol (or methanol) with dodecane, as the vehicle, has become popular. However, the contribution of this solvent to the observed effects of C1P has not been documented. Here, we show that addition of C1P in ethanol-dodecane to culture medium leads to irreversible cytotoxic effects. These culminate in mitochondrial swelling, vacuole formation, and cell death. Not only the toxicity of C1P, but also its ability to trigger prostaglandin E2 release, is fully dependent upon addition of a premade C1P-dodecane mixture. Furthermore, we show that these effects are not restricted to C1P. They result from the capacity of dodecane to interact with phospholipids; hence, they go undetected with a vehicle control. This study should raise awareness about the use of dodecane for phospholipid delivery and, in turn, help in unraveling C1P signaling, which is still poorly understood.