detection and characterization of ceramide- 1-phosphate

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 29, Issue of October 15, pp. 22150-22155, 1993 Printed in U. S. A.

Detection and Characterization of Ceramide- 1-phosphate Phosphatase Activity in Rat Liver Plasma Membrane*

(Received for publication, March 31, 1993, and in revised form, June 21, 1993)

Olga Boudker and Anthony H. FutermanS From the Department of Membrane Research and Biophysics, Weizmann Institute of Science, Rehovot 76100, Israel

A calcium-dependent ceramide (Cer) kinase was recently detected in human leukemia (HL-60) cells (Ko- lesnick, R. N., and Hemer, M. R. (1990) J. Biol. Chem. 265, 18803-18808) where it may function in terminating the regulatory effects of Cer, and in synaptic vesicles (Bajjalieh, S. M., Martin, T. F. J., and Floor, E. (1989) J. Biol. Chem. 264,14354-14360). We now demonstrate that the addition of both Cer- 1-phosphate (Cer-1-P) and a short-acyl chain analog of Cer-1-P, N- hexanoylsphingosine-1-phosphate (Ca-Cer-1-P) to cultured cells and a variety of subcellular fractions results in rapid degradation to Cer and Ca-Cer, respectively. The Cer-1-P phosphatase activity is enriched in a rat liver plasma membrane fraction and appears to be distinct from the phosphatase that hydrolyzes phosphatidic acid (PA), PA phosphohydrolase, as shown by the difference in sensitivity of Cer-1-P and P A hydrolysis to propranolol, detergent, and heat treatment. More- over, the K , of Cer-l-P hydrolysis is 10-fold lower than the K,,, of P A hydrolysis in plasma membrane. PA is a noncompetitive inhibitor of Cer-1-P hydrolysis, with an inhibition constant 1-1.5-fold higher than the K , of Cer-1-P hydrolysis. In contrast, Cer-1-P does not inhibit PA hydrolysis. Finally, we describe the synthesis of a novel analog of Cer-1-P which is not hydrolyzed in vitro and in vivo and is internalized in cultured cells by endocytosis. These results are dis- cussed in relation to the possible roles of Cer-1-P in regulating intracellular levels of Cer.

A number of sphingolipid-derived products are believed to play roles in cell regulation (for review see Hannun and Bell, 1989; Kolesnick, 1991), among them ceramide (Cer)’ (Koles- nick, 1992). A variety of bioactive agents, including 1,a-25- hydroxyvitamin DS (Okazaki et al., 1989), and interferon-y and tumor necrosis factor-a (Kim et al., 1991), stimulate a

* This work was partly supported by the Basic Research Founda- tion of the Israel Academy of Sciences and Humanities, by the Johns Hopkins University-Weizmann Institute Cooperative Research Pro- gram, and by the Josef Cohn Centre for Biomembrane Research of the Weizmann Institute of Science. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Incumbent of the Recanati Career Development Chair in Cancer Research. To whom correspondence should be addressed. Tel.: 972- 8-342704; Fax: 972-8-344112. E-mail: BMFUTEmWE1ZMANN.- BITNET.

The abbreviations used are: Cer, ceramide; Cer-1-P; ceramide 1- phosphate; CG-Cer, N-hexanoylsphingosine; NBD, 4-nitrobenzo-2- oxa-1,3-diazol; SM, sphingomyelin; N-SMase, neutral-sphingomyelinase; PA, phosphatidic acid DAG, diacylglycerol; NEM, N-ethylmaleimide; BSA, bovine serum albumin; CHO, Chinese hamster ovary; PM, plasma membrane.

neutral sphingomyelinase (N-SMase) activity, resulting in the degradation of sphingomyelin (SM) to Cer, as an early event in the action of these agents. The addition of cell-permeable analogs of Cer to human leukemia HL-60 cells mimics the biological effects of these agents, including inhibition of cell proliferation and induction of monocyte differentiation (Kim et al., 1991; Okazaki et al., 1990). Cer also stimulates a novel protein kinase that phosphorylates the epidermal growth factor receptor (Goldkorn et al., 1991; Mathias et al., 1991) and stimulates a cytosolic protein phosphatase (Dobrowsky and Hannun, 1992).

Since Cer is involved in cell regulation, its intracellular levels must be carefully controlled. Recently, a novel pathway has been suggested at the plasma membrane (PM) in which Cer produced as a result of N-SMase, but not glucosylceram- idase, is rapidly phosphorylated by a calcium-dependent Cer kinase (Dressler and Kolesnick, 1990; Kolesnick, 1992; Koles- nick and Hemer, 1990). Although the function of ceramide 1- phosphate (Cer-1-P) has not been established, it is possible that Cer phosphorylation terminates the modulatory effects of Cer. A similar calcium-dependent Cer kinase was reported previously in synaptic vesicles (Bajjalieh et al., 1989), which was suggested to be involved in neurotransmitter release.

We now demonstrate that Cer-1-P is rapidly hydrolyzed in cultured cells, and although hydrolyzed in a variety of subcellular fractions from rat liver and brain, hydrolysis is enriched in liver PM and brain synaptosomes. Characterization sug- gests that a novel phosphatase may be responsible for Cer-l- P hydrolysis, and we discuss the possible function of Cer-1-P hydrolysis in regulating cellular levels of Cer and Cer-1-P and other biologically active metabolites of sphingolipids.

EXPERIMENTAL PROCEDURES

Materials-Rats (Wistar) were obtained from the Weizmann In- stitute Breeding Center. l-[“C]Hexanoic acid was from American Radiolabeled Chemicals Inc. (St. Louis, MO). [14C]Phosphatidic acid, (1-a-dipalmitoyl [gly~erol-’~C]J was from Du-Pont-New England Nu- clear. N-Hexanoylsphingosine (C6-Cer) was from Matreya (Pleasant Gap, PA). Succinimidyl6-(7-nitrobenzo-2-oxa-1,3-diazol-4-yl)amino- hexanoate, N-[6-[(7-nitrobenzo-2-oxa-1,3-diazol-4-yl)amino]hexanoyl]sphingosylphosphorylcholine (C,-NBD-SM), and N-[6-[(7-nitrobenzo-2-oxa-1,3-diazol-4-yl)-amino]hexanoyl]sphingosine (Cs-NBD- Cer) were from Molecular Probes (Eugene, OR). l-Acyl-2-[6-[(7-

phate (Cs-NBD-PA), l-palmitoyl-2-[6-[(7-nitrobenzo-2-oxa-l,3- nitrobenzo-2-oxa-1,3-diazol-4-yl)amino]caproy~]-sn-glycero-3-phos-

diazol-4-yl)amino]caproyl]-sn-glycero-3-phoephocholine (Cs-NBD-phosphatidylcholine), 1,2-dioleoyl-sn-phosphocholine, and 1,2-dipalmitoyl-sn-glycero-3-phosphate were from Avanti Polar Lipids (Ala- baster, AL). D-Sphingosine (from bovine brakcerebrosides), ceram- ides (type 111 from bovine brain sphingomyelin), sphingomyelinase (from Streptomyces), phospholipase C (type I from Clostridium perfingem), acid phosphatase (type I1 from potato), alkaline phosphatase (type XVI from porcine placenta) and N-ethylmaleimide (NEM) were from Sigma. Diacylglycerol (DAG) kinase (Escherichia coli) was from Lipidex (Westfield, NJ). N,N’-dicyclohexylcarbodiimide was from

22150

Ceramide Phosphate Hydrolysis 22151

Merck, as were Silica Gel 60 plates. All other chemicals and reagents were from Sigma. Solvents (analytical grade) were from Bio-Lab Laboratories Ltd. (Jerusalem, Israel).

Preparation and Analysis of Lipids-N-(l-["C]hexanoyl)-D-ery- thro-sphingosine (Cs-["C]Cer) and Cs-NBD-Cer were prepared by N- acylation of sphingosine using the N-hydroxysuccinimide esters of 1- ["Clhexanoic acid or NBD-hexanoic acid, respectively, as described (Futerman and Pagano, 1991b; Pagano, 1989; Schwarzmann and Sandhoff, 1987). C6-lipids were prepared as a complex with defatted bovine serum albumin (BSA) (molar ratio 5:l) as described (Pagano, 1989) except that 10 mM Hepes, pH 7.4, was used. After 15-20 min at room temperature, complexes were dialyzed against 10 mM Hepes, pH 7.4, for 12-18 h at 4 'C. To minimize loss of material when small volumes of &-lipid. BSA complexes were prepared, dialysis was omit- ted. Complexes were stored at -20 "C.

C&er was phosphorylated by incubating a C6-Cer.BSA complex with DAG kinase in 5 mM ATP, 100 mM NaC1, 1 mM dithiothreitol, 1 mM EDTA, 1 mM EGTA, 10 mM MgClZ, 50 mM Hepes, pH 7.4, for 12-18 h at 37 'C with stirring (Dressler and Kolesnick, 1990; Schnei- der and Kennedy, 1973). Lipids were extracted as described below and analyzed by preparative TLC. Cs-Cer and CB-Cer-1-P have RF values of 0.9 and 0.3, respectively, using ch1oroform:methanolacetic acid15 mM CaClz (6035:2:4, v/v/v/v) as the developing solvent. Natural Cer (prepared from bovine brain SM) or [4,5-3H]dihydrocer- amide (N-palmitoyl-[4,5-3H]dihydrosphingosine; prepared as described in Hirschberg et al., 1993) were emulsified by bath sonication prior to phosphorylation as above.

A nonhydrolyzable analog of C6-NBD-Cer-1-P was prepared by esterification (Barker, 1971) of the 3-hydroxy and 1-phosphate groups of Cer-1-P. Ce-NBD-Cer-l-P (up to 1 pmol) was incubated with N,N" dicyclohexylcarbodiimide in 2 ml of N,N-dimethylformamide in a saturated atmosphere of Nz. After 1-2 days at room temperature, N,N-dimethylformamide was dried under a stream of N2 and lipids extracted (Bligh and Dyer, 1959) prior to resolution by preparative TLC using ch1oroform:methanolacetic acid15 mM CaC1, (60:35:2:4, v/v/v/v) as the developing solvent. The major band with an RF value of 0.55 was C~-NBD-Cer-cyclic-l,3-phosphate (see "Results" and Fig. 4B).

The 3-keto forms of Ce-NBD-Cer and C6-NBD-Cer-1-P were obtained by incubation with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone in dioxane (3%, w/v) in Nz-sealed vials for 24 h at 37 "C with stirring (Iwamori et al., 1975; Kishimoto and Mitry, 1974). Lipid extracts (Bligh and Dyer, 1959) were resolved by analytical TLC using chlo- rof0rm:benzene:ethanol (8040:75, v/v/v) to separate 2,3-dichloro-5, 6-dicyano-1,4-benzoquinone from the phosphosphingolipids, and chloroform:methanoll5 mM CaClz (60:35:8, v/v/v) to resolve 3-keto- lipids, which had a slightly higher RF value than the parent com- pounds.

Stock concentrations of radioactive lipids were determined by liquid scintillation counting in a Packard 1500 Tri-Carb scintillation counter using Lumax:toluene (1:3, v/v) as scintillation fluid. The stock concentration of C6-NBD-labeled lipids was determined using a stock solution of Cs-NBD-Cer, prepared by dissolving 1 mg of Ce- NBD-Cer in 2 ml of ch1oroform:methanol (19:1, v/v). NBD fluorescence (bx = 468; b,,, = 530) was quantified using a Perkin-Elmer- Cetus Instruments LS-5 luminescence spectrometer. Stock concentrations of Cs-NBD and other phospholipids were determined by phosphate analysis (Rousser et al., 1966).

Subcellulnr Fractionation-Rat liver microsomes (Ehrenreich et al., 1973) and a purified PM fraction (Hubbard et al., 1983) were prepared as described (Futerman et al., 1990). Rat liver cytosol was prepared by collecting the supernatant obtained by centrifugation (100,000 X gay, 1 h, 4 "C) of a postnuclear fraction. Marker enzymes were assayed as described (Futerman and Pagano, 1991a; Futerman et al., 1990). Synaptosomes from rat brain were prepared according to Dunkley et al. (1988).

Cell Culture-Human leukemia (HL-60) cells were grown in RPMI 1640 medium containing 10% fetal calf serum, 10 units/ml penicillin, 10 pg/ml streptomycin, and 20 mM glutamine. Chinese hamster ovary (CHO) cells were grown in Dulbecco's modified Eagle's medium with 5% fetal calf serum, 10 units/ml penicillin, 10 pg/ml streptomycin, and 20 mM glutamine. All cells were grown at 37 "C in an atmosphere

Enzyme Assays-Ce-NBD-PA and C6-NBD-Cer-1-P hydrolysis was assayed as follows. 20 p1 of a C6-lipid.BSA complex (containing various amounts of the C6-lipid) were added to 50 p1 of buffer containing 200 mM NaC1,2 mM EGTA, 2 mM EDTA, 100 mM Hepes, pH 6.8. The specific activity of Ce-["CICer-1-P was adjusted by the

Of 5% COz.

addition of unlabeled Cs-Cer-1-P. The total volume was adjusted to 90 pl and the reaction started by the addition of 2.5-10 pg of protein in a volume of 10 pl. The reaction time (10-30 min) and protein concentration were adjusted so that the amount of product formed was linear with respect to time and protein concentration. In assays in which long acyl-chain substrates were used, lipids were added as an emulsion of Cer-1-P or PA with dioleoylphosphatidylcholine (molar ratio 3:2), prepared by bath sonication.

Cer-1-P hydrolysis was measured in cultured cells by adding c6- ["C]- or Ce-NBD-Cer-l-P (5 nmol) as a complex with BSA directly to culture dishes. Cells were incubated at 37 'C for 1-3 h, harvested, and homogenized prior to lipid extraction.

Lipids were extracted as described for phosphatidate extraction (Martin et al., 1991). Reactions were terminated by 3.75 volumes of ch1oroform:methanol (1:2, v/v). Two phases were generated by the addition of 1.25 volumes of chloroform and 1.25 volumes of 2 M KC1 containing 0.2 M Hapod. When large quantities of lipid were extracted, the aqueous phase was washed with an additional volume of chloroform, and the organic phases were combined and dried under N2. Lipids were resolved by TLC on Silica Gel 60 plates using ch1oroform:methanol:acetic acid15 mM CaC1, (6035:2:4, v/v/v/v) as developing solvent. Radioactive lipids were identified after autoradi- ography by comparison with authentic Ce-"C-lipid standards, re- covered from the plates by scraping, and radioactivity determined by liquid scintillation counting. Cs-NBD-lipids were identified by comparison with authentic Cs-NBD-lipid standards. Background radioactivity or fluorescence in the area corresponding to Cer and DAG was determined by incubation of labeled Cer-1-P or PA in 100 pl of the reaction buffer for 10 min at 37 "C in the absence of protein, followed by extraction, separation by TLC, scraping, and subtracting from the radioactivity or fluorescence in the corresponding experimental lanes. The efficiency of extraction and counting of radioactive lipids and the quantum yields of fluorescent lipids were identical for "C- and NBD-labeled Cer, Cer-1-P, DAG, and PA, respectively.

Cell Labeling with Short Acyl-chain Fluorescent Lipid Analogs- Cells were labeled with short acyl-chain fluorescent lipid analogs as described (Pagano, 1989; Pagano and Sleight, 1985; Pagano and Longmuir, 1985). CHO cells were grown on 24-mm acid-washed glass coverslips. Dishes containing the coverslips were cooled at 4 "C for 5 min prior to the addition of 5 nmol of a C6-NBD-lipid. BSA complex. Cells were incubated for 30 min at 4 "C, washed three times with Dulbecco's modified Eagle's medium, and incubated for 30 min at 37 'C. Cs-NBD-lipids were removed by "back-exchange" (Pagano, 1989) from the PM by five sequential incubations of the cells in 3.4 mg/ml BSA for 5 min. Cell labeling was monitored by fluorescence microscopy using an Apochromat 63X/1.40 oil objective of a Zeiss Axiovert 35 microscope.

RESULTS

A calcium-dependent Cer kinase that phosphorylates Cer to Cer-1-P has been detected in rat brain synaptic vesicles (Bajjalieh et al., 1989) and HL-60 cells (Dressler and Koles- nick, 1990; Kolesnick and Hemer, 1990). We have now ex- amined whether Cer-1-P is hydrolyzed by these and other tissues. C6-Cer and long-acyl chain Cer (natural Cer obtained from bovine brain SM) were phosphorylated by purified DAG kinase (see "Experimental Procedures") and incubated with a variety of cultured cells and subcellular fractions. In all cases, Cer-1-P was hydrolyzed, although with different specific activities (Fig. 1). The highest specific activity was observed in rat liver PM (44.4 iZ 0.2 nmol of C6-NBD-Cer formed/min/mg of protein), which was enriched approximately 10-fold over the homogenate (specific activity of 5.0 f 0.9 nmol/min/mg), 5-fold over a microsomal fraction (9.9 f 1.2 nmol/min/mg), and 60-fold over cytosol (0.7 f 0.0 nmol/ min/mg). The specific activities of rat brain homogenate and synaptosomes were 17.4 k 1.1 and 25.9 iZ 0.7 nmol/min/mg, respectively, and the specific activity of CHO cells (14.5 f 3.6 nmol/min/mg) was 6-fold higher than HL-60 cells (2.4 f 0.8 nmol/min/mg) (Fig. 1). In a rat liver homogenate, using either C6-Cer-1-P or Cs-NBD-Cer-1-P as substrate, hydrolysis was linear for up to 50 min and 100 Kg of protein, with a broad pH optima of 5.5-7.0 (not shown).

22152 Cerarnide Phosphate Hydrolysis

4- Cer

4- GlcCer

4- CerP

4- SM

a b c d e f g

FIG. 1. Chromatographic analysis of Ca-NBD-Cer-l-P hydrolysis. A Ce-NBD-Cer-l-P.BSA complex was incubated with the following enzymes or subcellular fractions. Lane a, acid-SMase (50 milliunits) in the same buffer routinely used to analyze Cer-1-P hydrolysis; SMase rapidly hydrolyzed Ce-NBD-SM under these conditions (not shown). Lane b, phospholipase C (80 milliunits) (type I from c. perfingem) in 6.3 mM CaC12, 150 mM (NH4)2S04, 50 mM Tris-HCI, pH 7.4; phospholipase C rapidly hydrolyzed Ca-NBD-phosphatidylcholine under these conditions (not shown). Lane c, 10 pg of protein of a rat liver homogenate. Lane d , 2.5 pg of protein of a purified PM fraction. Lane e, 10 pg of protein of a rat brain homogenate. Lane f , 5 pg of protein of purified synaptosomes. Reactions a-f were performed for 15 min at 37 "C. Laneg, Cs-NBD-Cer-1-P was incubated with cultured CHO cells in vivo for 90 min. Ce-NBD-Cer formed as a result of Ce-NBD-Cer-1-P hydrolysis was subsequently metabolized to Ce-NBD-glucosylceramide and Ce-NBD-SM. The following Ce-NBD-sphingolipids were used as TLC standards as indi- cated: Cer, glucosylceramide (GlcCer), Cer-1-P (CerP), and SM.

To determine whether Cer-1-P is degraded by known phosphohydrolases, C6-NBD-Cer-1-P was incubated with purified sphingomyelinase or purified phospholipase C (Fig. l), under conditions in which these two enzymes hydrolyzed C6-NBD- SM and C6-NBD-phosphatidylcholine, respectively, and with acid phosphatase (0.1 unit/ml) or alkaline phosphatase (0.1 unit/ml) (not shown). None of these phosphohydrolases demonstrated any hydrolytic activity toward C6-NBD-Cer-l-P.

Comparison of Cer-I -P and PA Hydrolysis in Rat Liver PM-Experiments were performed to determine whether Cer- 1-P and PA are hydrolyzed by two distinct enzymes or whether the enzyme that hydrolyzes PA, PA phosphohydrolase, is responsible for Cer-1-P hydrolysis in vitro. Two iso- forms of PA phosphohydrolase have been characterized in rat liver (Jamal et al., 1991): an NEM-insensitive form associated with the PM, and an NEM-sensitive microsomal/cytosolic form.

Similar to PA phosphohydrolase (see Jamal et al., 1991 and Table I), the enzyme that hydrolyzes Cer-1-P was not inhibited by NEM in the PM using either c6- or long-acyl chain Cer-1-P as substrate, whereas long-acyl chain Cer-1-P hydrolysis was strongly inhibited by NEM in cytosol (Table I). These results indicate that there are a t least two activities in rat liver which are able to hydrolyze Cer-1-P. Surprisingly, hydrolysis of the C6-analogs of Cer-1-P and PA was inhibited to a much lesser extent by NEM in cytosol (Table I) than their long acyl-chain counterparts, suggesting that cytosol may contain in addition an NEM-insensitive, nonspecific phosphatase that is able to hydrolyze C6-phospholipid analogs but not long-acyl chain phospholipids. Since Cer-1-P hydrolysis is highly enriched in the PM (Fig. 1) and is not inhibited by NEM (Table I), Cer-1-P hydrolysis was characterized further in this fraction.

No significant difference was observed in the effect of sodium fluoride and sodium vanadate (phosphohydrolase in- hibitors), 2,4-dinitrofluorobenzene (a lysine reagent), and

TABLE I Effect of NEM on Cer-1-P and PA hydrolysk

Rat liver PM or cytosol was preincubated at 37 "C for 15 min with 5 mM NEM. The reaction buffer used to measure lipid hydrolysis contained 1 mM dithiothreitol and 3 mM M&12 (Jamal et al., 1991). Both c8- and long acyl-chain lipids were added as an emulsion with 1,2-dioleoyl-sn-glycero-3-phosphocholine (molar ratio of 3:2), prepared by bath sonication, to give a final concentration of 0.3 mM Cer- 1-P or PA. Results are shown as ? S.D. and are means of at least two independent experiments.

with NEM an a percentage of Activity after preincubation

Lipid substrate the control

PM Cytonol %

Ce-NBD-Cer-1-P 84.8 f 1.6 81.5 f 4.2 113.0 f 8.3 48.9 2 11.9

Long acyl-chain Cer-1-P 87.0 f 1.6 17.4 f 2.9 Long acyl-chain PA ND" 7.0 f 2.0

Ce-NBD-PA

ND, not determined.

TABLE I1 Effect of propranolol on Cer-1-P and PA hydrolysk

Ce-Cer-l-P and Ce-NBD-PA hydrolysis was assayed as described under "Experimental Procedures," in the presence of propranolol. _~

Propranolol Activity an a percentage of the control"

concentration Cs-NBD-PA hydrolysis Cs-Cer-l-P hydrolysis mM % 2 62.5 f 18.3 82.5 f 9.4

4 42.1 2 18.0b 71.5 f 19.2b a Considerable variability in the effect of propranolol on both CS-

NBD-PA and Ce-Cer-l-P hydrolysis was observed between experiments. The results represent the mean 2 S.D. of a t least five independent experiments for each concentration of propranolol and for each lipid.

* p < 0.01 (Student's t test).

sphingosine (an amphiphilic amine) on Ce-Cer-1-P and c6- NBD-PA hydrolysis in rat liver PM. However, the amphiphilic amine propranolol, which inhibits PA phosphohydrolase in the PM by approximately 50% a t 2 mM using phosphatidate as substrate (Jamal et aL, 1991), had a significantly greater inhibitory effect on C6-NBD-PA than on Cfl-Cer-l-P hydrolysis (Table 11). Although C6-Cer-1-P hydrolysis was inhibited to a limited extent, the difference in inhibition of hydrolysis of the two lipids by propranolol may suggest that distinct enzymes are involved in hydrolyzing Cer-1-P and PA. However, since propranolol may inhibit PA phosphohydrolase hydrolysis by forming complexes with PA and preventing it from entering the active site (Pappu and Hauser, 1983), the difference in susceptibility to propranolol may simply reflect differences in the interaction of propranolol with Cfl-Cer-l-P and C6-NBD-PA, respectively.

PA phosphohydrolase in PM is stable to heat treatment and partially activated after 15-20 min at 55 "C using phosphatidate as substrate (Jamal et al., 1991). Similarly, PM PA phosphohydrolase was activated by heat treatment using c6- NBD-PA as substrate, whereas C6-Cer-1-P hydrolysis was slightly inhibited (Fig. 2A). At low concentrations, Triton X- 100 activated C6-NBD-PA hydrolysis (Fig. 28). similar to i t s effect on phosphatidate hydrolysis (Jamal et al., 1991). but Triton X-100 significantly inhibited the hydrolysis of C6-Cer- 1-P in the PM even a t low concentrations (Fig. 2R). The differences in the sensitivity of C6-NBD-PA and Ch-Cer-1-P hydrolysis to heat treatment and to detergent suggest that unique enzymes are responsible for hydrolysis of these two lipids at the PM.


I I 1 10 20 30 0 1 2 3 4

Time (min) Triton X-100 concentration ( X viv)

FIG. 2. Effect of heat treatment and of detergent on Cer-l- P and PA hydrolysis. Panel A, aliquots of PM sheets were preincubated at 55 "C for 5-30 min, cooled on ice for 5 min, and incubated with either C6-Cer-l-P (0) or C6-NBD-PA (W). Panel B , the hydrolysis of C6-Cer-l-P (0) or C6-NBD-PA (W) in PM sheets was assayed in the presence of 0.1-4% Triton X-100. The results are means of three independent experiments.

/

-0.05 L

1 S

FIG. 3. Kinetic analysis of the inhibition of Cer-l-P hydrolysis by PA. Cer-l-P hydrolysis was measured using 5, 50, 100 and 200 pM C6-Cer-l-P as substrate in the absence (W) or presence of 50 (O), 100 (01, or 200 p~ (0) C6-NBD-PA. The data are representative of four similar experiments. Panel A, double-reciprocal plot. Only values for the three highest concentrations of C6-Cer-l-P are shown, although the curves represent the best fit by linear regression using all four concentrations. Panel B, Eadie-Hofstee plot. The curves are best fits by linear regression. Units of velocity (V,) and concentration ( S ) are are nmol of C6-Cer formed/min/mg of protein, and p ~ , respectively.

Inhibition of C6-Cer-1-P and C6-NBD-PA Hydrolysis by c6-

lipids-Further evidence that two separate enzymes are responsible for hydrolyzing PA and Cer-l-P was provided by kinetic analysis. The K,,, of C6-NBD-PA hydrolysis in PM was 700 f 35 PM (n = 4), similar to that reported using phosphatidate as substrate (Jamal et al., 1991). However, the K,,, of C6-Cer-l-P hydrolysis was much lower, varying between 46 and 93 pM with a mean of 55 * 16 p~ (n = 8). In three of four experiments, C6-NBD-PA hydrolysis was not inhibited by C6-Cer-l-P, and in one experiment, slight inhibition was observed with a Ki value 10-fold higher than the K,,, of hydrolysis of C6-NBD-PA (not shown). In contrast, C6-NBD- PA significantly inhibited C6-Cer-l-P hydrolysis, but analysis of these data by the double-reciprocal plot demonstrated that the lines intercepted at l/Km at the x axis, indicating noncompetitive inhibition (Fig. 3A). Analysis of the data by the Eadie- Hofstee plot gave a series of parallel lines (Fig. 3B), confirm- ing that Ca-NBD-PA affects the Vmax of CG-Cer-l-P hydrolysis without affecting K,,,. These data are consistent with a noncompetitive type of inhibition, suggesting that C6-NBD-PA does not compete with Cs-Cer-l-P for the same active site. The inhibition of C6-Cer-l-P hydrolysis by C6-NBD-PA is not caused by a nonspecific effect of the Cs-lipid since C6- NBD-SM had little effect on C6-Cer-l-P hydrolysis, with the K, value 8-fold higher than the K,,, (not shown).

The Ki for the noncompetitive inhibition of Cs-Cer-l-P by

C6-NBD-PA (83 p ~ ) is about 9-fold lower than the K,,, of Cg- NBD-PA hydrolysis (700 p ~ ) , indicating that the enzyme that hydrolyzes C6-Cer-l-P is not responsible for significant hydrolysis of C6-NBD-PA. In addition, C6-Cer-l-P hydrolysis could not be the result of PA phosphohydrolase activity since C,-NBD-PA hydrolysis was not inhibited by C6-Cer-l-P. These kinetic considerations imply that even if Cer-l-P is able to be hydrolyzed by PA phosphohydrolase, and PA by the enzyme that hydrolyzed Cer-l-P, these reactions would constitute only an insignificant part of the overall hydrolysis of these lipids and suggest that two distinct enzymes are responsible for the hydrolysis of PA and Cer-l-P.

Churacterization of Nonhydrolyzable Derivative of Cer-l-P-A nonhydrolyzable analog of C6-Cer-l-P (Fig. 4A) was synthesized by esterification of the 3-hydroxy and 1- phosphate groups of C6-Cer-l-P (see "Experimental Proce- dures"). The compound had the proposed structure of C6-Cer cyclic 1,3-phosphate (Fig. 4B) as demonstrated by the following: (1) the compound was more hydrophobic than Cs-Cer-l- P since it had a higher mobility on TLC using a number of different developing solvents; (2) an equimolar ratio of the NBD-fluorophore and of phosphate was obtained for the new compound; (3) high resolution mass spectrometry revealed that the molecular weight of the new compound appeared to be 18 less than Cs-Cer-l-P; (4) the compound did not contain a 3-hydroxy group that was susceptible to oxidation by 2,3- dichloro-5,6-dicyano-1,4-benzoquinone (Iwamori et al., 1975; Kishimoto and Mitry, 1974).

Whereas C6-NBD-Cer-1-P was rapidly hydrolyzed to cg- NBD-Cer by liver fractions and by CHO and HL-60 cells (see above), resulting in labeling of the Golgi apparatus of cultured CHO cells by C6-NBD-Cer (Fig. 5A), C6-NBD-Cer cyclic-1,3- phosphate was totally resistant to hydrolysis in liver fractions and in cultured cells (not shown) and labeled a population of small vesicles in cultured CHO cells (Fig. 5B). These vesicles were indistinguishable from those labeled by C6-NBD-SM (not shown), which is internalized in CHO cells along the endocytic pathway (Koval and Pagano, 1989), indicating that C6-NBD-Cer cyclic 1,3-phosphate can be used as a marker of the endocytic pathway.

Cs-NBD-Cer cyclic 1,3-phosphate weakly inhibited C6-Cer- l-P hydrolysis with a Ki 8-fold higher than the K , of C6-Cer- l - P hydrolysis. This low level of inhibition might be the result of a nonspecific effect of high concentrations of short acyl- chain lipids in the reaction mixtures, since C6-NBD-cyclic- 1,3-Cer-l-P similarly acted as a weak inhibitor of C6-NBD- PA hydrolysis.

A 0 I

CHFO-P-OH I

OH dH

"""""""""""""""

B " - Y o

FIG. 4. Molecular structure of phosphorylated derivatives of Cer. Panel A, C6-Cer-l-P. Panel B, C6-Cer cyclic 1,3-phosphate.

22154 Ceramide Phosphate Hydrolysis

FIG. 5. Cs-NBD-Cer cyclic 1,3-phosphate labels the endocytic pathway. CHO cells were labeled with either Cs-NBD-Cer-1- P ( p a n e l A ) or Cs-NBD-Cercyclic 1,3-phosphate (pane lB) . Cs-NBD- Cer-1-P is hydrolyzed to C6-NBD-Cer which labels the Golgi apparatus ( p a n e l A ) , whereas C8-NBD-Cer cyclic 1,3-phosphate labels endocytic vesicles (panel B ) and an area that probably corresponds to the microtubule organizing center (Koval and Pagano, 1989).

DISCUSSION

Initial observations concerning the function of sphingolipids in signal transduction lead to the suggestion that sphingoid bases and other lysosphingolipids play major roles in cell regulation (Hannun and Bell, 1989; Hannun et al., 1986). However, the relatively small quantities of sphingoid bases generated from Cer in response to biological activators implied that Cer and its derivatives may also play direct roles in signal transduction (for review see Kolesnick, 1991,1992). Following the discovery of a calcium-dependent Cer kinase in HL-60 cells, a SM pathway was proposed in the PM in which the activation of N-SMase leads to the generation of Cer, which can be subsequently phosphorylated by Cer kinase (Dressler and Kolesnick, 1990; Kolesnick and Hemer, 1990). A similar Cer kinase was found to copurify with rat brain synaptic vesicles (Bajjalieh et al., 1989).

In the current report we demonstrate that Cer-1-P is rapidly hydrolyzed in a variety of tissues and subcellular fractions, with the highest specific activity observed in rat liver PM. Since Cer can be phosphorylated by E. coli DAG kinase (Schneider and Kennedy, 1973), indicating that Cer and DAG share some structural similarities, we considered the possibility that Cer-1-P might be hydrolyzed by PA phosphohydrolase (Pagano and Longmuir, 1985). However, our results indicate that Cer-1-P and PA are hydrolyzed by separate enzymes in the PM. Proof of the existence of a unique Cer-1-P phosphatase awaits purification of an activity that hydrolyzes Cer-1-

P with properties similar to those observed in rat liver PM. The recent purification of PA phosphohydrolase from porcine thymus membranes (Kanoh et al., 1992) will permit exami- nation of the possibility that P A phosphohydrolase can hydrolyze Cer-1-P. However, even if purified PA phosphohydrolase can hydrolyze Cer-1-P, the contribution of PA phosphohydrolase in hydrolyzing Cer-1-P in the PM will need to be evaluated taking into consideration the different K, values of PA and Cer-1-P hydrolysis.

A similar phosphatase activity that hydrolyzes Cer-1-P has also been detected independently in rat brain, where it is enriched in synaptic membranes.' If Cer phosphorylation is indeed involved in synaptic vesicle fusion with the presynaptic membrane (Bajjalieh et al., 1989), then Cer-1-P hydrolysis may be involved in regulating or terminating synaptic vesicle fusion.

In light of the observation that Cer-1-P is rapidly hydrolyzed to Cer, the suggestion that Cer phosphorylation acts to terminate the modulatory effects of Cer in the SM pathway may need to be reevaluated. The physiological function of Cer-1-P hydrolysis will be clarified once the function of Cer- 1-P is elucidated. Cer-1-P may be a short lived intermediate in another metabolic pathway unrelated to the SM pathway. For instance, Cer-1-P could be generated by a putative SMase D and subsequently hydrolyzed to Cer. A similar pathway exists for the generation of DAG from phosphoglycerolipids, in which phospholipase D produces PA, which is subsequently hydrolyzed by PA phosphohydrolase to DAG in the PM (Exton, 1990). The physiological relevance of the noncompetitive inhibition of Cer-1-P hydrolysis by PA is not known. However, evidence is accumulating that glycerolipid and sphingolipid metabolism are coordinately regulated and that this regulation may have implications for the function of each in signal transduction. For example, DAG stimulates SMase (Kolesnick, 1987). sphingosine inhibits PA phosphohydrolase (Lavie et al., 1990; Mullman et al., 1991). sphingosine (Lavie and Liscovitch, 1990), and sphingosine 1-phosphate (Desai et al., 1992), activates phospholipase D, and Cer is a competitive inhibitor of DAG kinase (Younes et af., 1992). Thus, P A may be involved in regulating levels of Cer and/or Cer-1-P at the PM by inhibiting Cer-1-P hydrolysis.

Since Cer-1-P is believed to be generated at the PM from Cer produced by the action of N-SMase, the role of Cer-1-P hydrolysis in microsomes and cytosol is unclear. No evidence exists that Cer synthesized de mu0 in the endoplasmic retic- ulum is a substrate for Cer kinase. However, sphingosine 1- phosphate has a number of regulatory effects, and the N- acylation of sphingosine 1-phosphate prior to Cer-1-P hydrolysis may be an alternative pathway for regulating levels of sphingosine 1-phosphate, in addition to degradation to phosphoethanolamine and hexadecenal (Merrill, 1991). It should be noted that we did not perform kinetic characterization of Cer-1-P hydrolysis in cytosol and microsomes and are therefore unable to determine the potential role of cytosolic and microsomal PA phosphohydrolase in hydrolyzing Cer-1-P.

In summary, we have detected and characterized a novel phosphatase that hydrolyzes Cer-1-P and appears distinct from PA phosphohydrolase. Controlling the levels of Cer and Cer-1-P by the coordinate action of Cer kinase and Cer-1-P phosphatase may be crucial in regulating the various metabolic pathways and regulatory events in which these lipids are involved.

Acknowledgments-We thank Professor Asher Mandelbaum. Mass Spectrometry Center, Technion. Haifa. for performing the high rea-

* S. Bajjalieh, personal communication.


olution mass spectrometry; Dr. Mordechai Liscovitch for critical reading of the manuscript and helpful comments; and Dr. Sandy Bajjalieh for sharing data prior to publication.

REFERENCES Bajjalieh, S. M., Martin, T. F. J., and Floor, E. (1989) J. Biol. Chem. 2 6 4 ,

Barker. R. (1971) in Oreanic Chemistn, of Biolopical CornDoudp. DD. 214-216. 14354-14360

~~~ ~~ . ~~ .~~ Prentice-HalljEngleGood Cliffs, NJ- '

I I .. Bligh, E. G., and Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37,911-917 Desai, N. N., Zhang, H., Olivera, A,, Mattie, M. E., and Spiegel, S. (1992) J.

BioL Chem. 267.23122-23128 Dobrowsky, R. T., &d Hannun, Y. A. (1992) J. Biol. Chem. 267,5048-5051 Dressler, K. A., and Kolesnick, R. N. (1990) J. Biol. Chem. 266,14917-14921 Dunkley, P. R., Heath, J. W., Harrison, S. M., Lavie, P. E., Glenfield, P. J.,

Ehrenreich, J. H., Bergeron, J. J. M., Siekevitz, P., and Palade, G. E. (1973) J.

Futerman, A. H., and Pagano, R. E. (1991a) Biochem. J. 280,295-302 Exton, J. H. (1990) J. Biol. Chem. 2 6 6 , l - 4

Futerman, A. H., and Pagano, R. E. (1991b) Methods Enzymol. 209,437-446 Futerman, A. H., Stieger, B., Hubbard, A. L., and Pagano, R. E. (1990) J. Biol.

Goldkorn, T., Dressler, K. A., Muindi, J., Radin, N. S., Mendelsohn, J., Chem. 265,8650-8657

Menaldino, D., Liotta, D., and Kolesnick, R. N. (1991) J. Bid. Chem. 2 6 6 , 16092-16097

and Roatas, J. A. P. (1988) Brain Res. 441,59-71

Cell Biol. 59,45-72

Hannun, Y. A., and Bell, R. M. (1989) Science 243,500-507 Hannun, Y. A,, Loomis, C. R., Merril, A. H., Jr., and Bell, R. M. (1986) J. Biol.

Hirschberg, K., Rodger, J., and Futerman, A. H. (1993) Biochern. J. 290,751-

Hubbard, A. L., Wall, D. A,, and Ma, A. (1983) J. Cell Biol. 9 6 , 217-229

Chem. 2 6 1 , 12604-12609

157

Iwamori, M., Moser, H. G., McCluer, R. H., and Kishimoto, Y. (1975) Biochim.

Jam$ i , 'Mart in , A., Gomez-Munoz, A,, and Brindley, D. N. (1991) J. Biol.

Kanoh, H., Imai, S.-i., Yamada, K., and Sakane, F. (1992) J. Biol. Chem. 2 6 7 ,

Kim, M.-Y., Linardic, C., Obeid, L., and Hannun, Y. (1991) J. Biol. Chem. 2 6 6 ,

Bio h s Acta 380,308-319

Chem. 266,2986-2996

25309-25314

A m - A R O Kishimoto, Y., and Mitry, M. T. (1974) Arch. Biochem. 161,426-434 Kolesnick. R. N. (1987) J. Biol. Chem. 262, 16759-16762

" ""

Kolesnick. R. N. (19911 Proe. LiDid Res. 30.. 1-38 Kolesnick, R. N. (1992) Trends Cell. Biol. 2 , 232-236 Kolesnick, R. N., and Hemer, M. R. (1990) J. Biol. Chem. 266,18803-18808 Koval, M., and Pagano, R. E. (1989) J. Cell Biol. 108 , 2169-2181 Lavie, Y., and Liscovitch, M. (1990) J. Biol. Chem. 268,3868-3872 Lavie, Y., Piterman. 0.. and Liscovitch, M. (1990) FEES Lett. 277, 7-10

. . ~~ """ ~, ". . ~ "", ~. . ~. ~~~~ - .. ~ ..

Martin, A., Gomez-Munoz, A., Zahirali,.J., and Brindley, D. N. (199i) Methods

Mathias, S., Dressler, K. A., and Kolesnick, R. N. (1991) Proc. Natl. Acad. Sci.

Merrill, A. H. (1991) J. Bioewr Biomem 23,83-104 Mullman. T. J.. Sieeel. M. I.. &an. R. W.. and Billah. M. M. (1991) J. Biol.

Enzymol. 197,553-564

Li. S. A. 8 8 , 10009-10013

Chem. 266,2013~2016 ' I '

19080

Chem. 268.15823-1.5831

. ,

Okazaki, T., Bell, R. M., and Hannun, Y. A. (1989) J. Biol. Chem. 264,19076-

Okazaki, T., Bielawska, A,, Bell, R. M., and Hannun, Y. A. (1990) J. Biol.

Pagano, R. E. (1989) Methods Cell Biol. 29, 75-85 Pagano, R. E., and Sleight, R. G. (1985) Sclence 229,1051-1057 Pagano, R. E., and Longmuir, K. J. (1985) J. Bml. Chem. 260,1909-1916 Pamu. A. S.. and Hauser. G. (1983) Neurochem. Res. 8. 1565-1575

~ . ~ "

Rdisser, B., Siakotos, A.,~and'Fleischer, S. (1966)Lipidi 1,85186 Schneider, E. G., and Kennedy, E. P. (1973) J. Biol. Chem. 248,3739-3741 Schwarzmann, G., and Sandhoff, K. (1987) Methods Enzymol. 138,319-341 Younes, A. Kahn D. W Besterman, J. M., Bittman, R., Byun, H. S., and

Kolesnick, R. NI (1992jb. Biol. Chem. 267,842-847

detection and characterization of ceramide- 1-phosphate

Documents