0 1991 by The American Society for Biochemistry and Molecular Biology, Inc. THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 266, No. 24, Issue of August 25, pp. 16092-16097,1991

Printed in U.S.A.

Ceramide Stimulates Epidermal Growth Factor Receptor Phosphorylation in A43 1 Human Epidermoid Carcinoma Cells EVIDENCE THAT CERAMIDE MAY MEDIATE SPHINGOSINE ACTION*

(Received for publication, February 5, 1991)

Tzipora GoldkornS, Kenneth A. DresslerS, Josephia MuindiS, Norman S. RadinQ, John MendelsohnS, David Menaldinoll, Dennis Liottaq, and Richard N. KolesnickS From the $Division of Molecular Pharmacology and Therapeutics, The Sloan Kettering Institute, Cornell University Medical College, New York, New York 10021, the §Mental Health Research Institute, University of Michigan, Ann Arbor, Michigan 48104, and the lIDepartment of Chemistry, Emory University, Atlanta, Georgia 30322

Recent studies suggest the existence of a signal trans- duction pathway involving sphingomyelin and deriv- atives (Kolesnick, R. N. (1989) J. Biol. Chern. 264, 7617-7623). The present studies compare effects of ceramide, sphingosine, and N,N-dimethylsphingosine on epidermal growth factor (EGF) receptor phos- phorylation in A431 human epidermoid carcinoma cells. To increase ceramide solubility, a ceramide con- taining octanoic acid at the second position (CS-cer) was synthesized. CS-cer induced time- and concentra- tion-dependent EGF receptor phosphorylation. This event was detectable by 2 min and maximal by 10 min. As little as 0.1 ~ L M CS-cer was effective, and 3 p~ CS- cer induced maximal phosphorylation to 1.9-fold of control. EGF (20 nM) increased phosphorylation to 2.1- fold of control. Sphingosine stimulated receptor phos- phorylation over the same concentration range (0.03- 3 p ~ ) and to the same extent (1.8-fold of control) as ceramide. The effects of CS-cer and sphingosine were similar by three separate criteria, phosphoamino acid analysis, anti-phosphotyrosine antibody immunoblot- ting, and phosphopeptide mapping by high perform- ance liquid chromatography. Phosphorylation oc- curred specifically on threonine residues. N,N-Di- methylsphingosine, a potential derivative of sphingo- sine, was less effective. Since sphingosine and ceram- ide are interconvertible, the level of each compound was measured under conditions sufficient for EGF receptor phosphorylation. CS-cer (0.1-1 WM) induced dose-responsive elevation of cellular ceramide from 132 to 232 pmol* 10' cells". In contrast, cellular sphin- gosine levels did not rise. This suggests that CS-cer acts without conversion to sphingosine. Exogenous sphingosine (0.1-1 p ~ ) also increased cellular ceram- ide levels to 227 pmol*10' cells-', but did not increase its own cellular level of 12 pmol.10' cells-'. Higher sphingosine concentrations that induced no further in- crease in EGF receptor phosphorylation produced very large elevations in cellular sphingosine. Hence, at ef-

* This work was supported by National Institutes of Health Grant R01-CA-42385 and American Cancer Society Grant FRA-345 (to R. N. K.), United States Public Health Service Grants NS 03192 and HD 07406 (to N. S. R.), National Science Foundation Grant DCB- 8710283 (to D. L.), and National Institutes of Health Grant CA- 42060 (to J. M.). K. A. D. was supported by Clinical Scholar Grant CA-09512-06 (Brian Piccolo Cancer Research Fund) from the Na- tional Institutes of Health and in part by the Charles A. Dana Foundation. 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 accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

fective concentrations, both compounds elevated cel- lular ceramide but not sphingosine levels. Additional studies performed with ['Hlsphingosine demonstrated that cells contain substantially less N,N-dimethyl- sphingosine than free sphingosine and, during short term incubation, convert less than 6% of added sphin- gosine to N,N-dimethylsphingosine. These studies pro- vide evidence that ceramide may have bioeffector properties and suggest sphingosine may act in part by conversion to ceramide.

Recent investigations from this and other laboratories have suggested the existence of a signal transduction pathway involving sphingomyelin and its derivatives (1-6). Such a pathway was initially proposed by this laboratory to involve sphingomyelin degradation to ceramide via the action of a sphingomyelinase. Evidence was presented that ceramide was converted to sphingoid bases, potential inhibitors of protein kinase C (7). Under these conditions, phorbol ester-induced activation of protein kinase C was antagonized and differen- tiation of human leukemia (HL-60) cells into macrophages was prevented. Subsequently, it has been shown that vitamin D, interferon-y, and tumor necrosis factor-a stimulate a sphingomyelinase in HL-60 cells and that this event may, in part, mediate monocyte differentiation via these agents (4). Further support for a potential pathway from sphingomyelin derives from data that demonstrate the existence of ceramide 1-phosphate (8). This phosphorylated form of ceramide was generated selectively from ceramide derived from sphingo- myelin but not glycosphingolipids. These studies provide early evidence for a biological pathway involving sphingomyelin and its derivatives.

The purpose of the present studies is to determine whether ceramide may serve as a bioeffector molecule. In these studies, the effects of ceramide are compared to sphingosine and N,N-dimethylsphingosine. Recent investigations have dem- onstrated that sphingoid bases may induce a variety of bio- logic events independent of inhibition of protein kinase C. These include inhibition of thyrotropin-releasing hormone binding to GH, pituitary cells (9, lo), inhibition of several calmodulin-dependent enzymes ( l l ) , and phosphorylation of the epidermal growth factor (EGF)' receptor from A431 hu-

The abbreviations used are: EGF, epidermal growth factor; C8- cer, N-octanoyl sphingosine; HEPES, 4-(2-hydroxyethyl)-l-pipera- zineethanesulfonic acid PAS, Protein A-Sepharose; HPLC, high performance liquid chromatography; THF, tetrahydrofuran; PBS, phosphate-buffered saline; EGTA, [ethylenebis(oxyethylenenitrilo)] tetraacetic acid SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis.

16092

Ceramide-induced EGF Receptor Phosphorylation 16093

man epidermoid carcinoma cells (12-14). Since sphingosine may be acylated to form ceramide, the question arose as to whether some of the effects of exogenous sphingosine on cells might be mediated by ceramide. Natural ceramides contain long chain saturated and monounsaturated fatty acids at the second position and are quite hydrophobic (15). Hence, these compounds would not be readily available to cells when added directly into an aqueous medium. A similar problem was encountered during the study of 1,2-diacylglycerol effects on protein kinase C . This problem was overcome by the use of synthetic diacylglycerols containing fatty acids of shorter chain length (16). Similarly, the present studies utilize a synthetic ceramide containing octanoic acid at the second position and compare the effects of this compound to sphin- gosine and N,N-dimethylsphingosine. These studies utilize EGF receptor phosphorylation to perform this comparison. Ceramide, like sphingosine, stimulated phosphorylation of the EGF receptor. Both compounds induced quantitatively simi- lar effects and had similar concentration dependencies. Fur- ther, both compounds selectively induced threonine phos- phorylation and generated identical phosphopeptide maps. N,N-Dimethylsphingosine was slightly less effective. At effec- tive concentrations, virtually all of the sphingosine was con- verted to ceramide. In contrast, ceramide was not converted to sphingoid bases. The cellular level of ceramide increased with either compound to within 2-fold of basal. These studies provide initial evidence that ceramide may serve as a bioef- fector molecule. Further, these studies suggest that some of the biologic effects of exogenous sphingoid bases may be mediated via conversion to ceramide.

EXPERIMENTAL PROCEDURES

Materials-Ceramides (Type 111, prepared from bovine brain sphingomyelin), sphingosine, o-phthalaldehyde, protease inhibitors, and HEPES were from Sigma. Tosylphenylalanyl chloromethyl ke- tone-treated trypsin was from Worthington. [32P]Orthophosphate (carrier-free) was from ICN Radiochemicals. [4,5-3H]Sphingosine (8.5 Ci/mmol) was kindly provided by Dr. David Ahern from Du Pont- New England Nuclear and repurified by TLC prior to utilization. Protein A-Sepharose (PAS) CL-4B was from Pharmacia LKB Bio- technology Inc. Fetal and neonatal calf serum were from Gibco Laboratories. Ceramide containing octanoic acid at the second posi- tion (N-octanoyl sphingosine termed C8-cer) was prepared from D- sphingosine as previously described (17). Anti-phosphotyrosine anti- body PY-69 was from ICN. EGF receptor monoclonal antibody 528 was produced as described (18). Reagents were HPLC grade and from Fisher.

Synthesis of N,N-Dimethylsphingosine-All compounds listed be- low were fully characterized using a combination of the following techniques: infrared spectroscopy, nuclear magnetic spectroscopy, mass spectroscopy, and high pressure liquid chromatography. All structural assignments are consistent with the results obtained.

HO OH HO OH

2

HO OH HO OH

n 0 3

4

N-Butoxycarbonyl dehydrosphingosine 1 was prepared as de- scribed by Nimkar et al. (19). Compound I (1.26 mmol, 500 mg) was dissolved in 18 ml of dry tetrahydrofuran (THF) and stirred under Nz atmosphere. Solid lithium aluminum hydride (6.2 eq, 20.4 mmol, 775 mg) was added slowly, and the reaction mixture was refluxed for 24 h or until the reaction was complete. The reaction was monitored

by TLC using 79% methylene chloride, 20% methanol, 1% of a 10% ammonium hydroxide solution as the solvent system. The work-up procedure consisted of quenching the reaction while cooling in an ice bath with equal amounts of 1 N sodium hydroxide and water, dilution with ethyl acetate (approximately 100 ml), addition of anhydrous magnesium sulfate, filtration, and concentration of the crude product in uucuo to afford N-methyl sphingosine (2, white solid) in 80% yield. Purification was done using standard flash column chromatography techniques with 79% methylene chloride, 20% methanol, 1% of a 10% ammonium hydroxide solution as the solvent system.

N-Methyl sphingosine 2 (1.0 eq, 0.32 mmol, 100 mg) was dissolved in 25 ml of dry THF at room temperature under N, atmosphere. Sodium carbonate (1.5 eq, 48 mmol, 50 mg) and benzyl chloroformate (1.5 eq, 48 mmol, 82 mg) were then added, and the reaction mixture was stirred overnight. The THF was evaporated in uacuo, and the remaining residue was diluted with methylene chloride (approxi- mately 100 ml), washed with water and brine, dried over anhydrous magnesium sulfate, and concentrated in uacuo. The resulting carba- mate 3 in 90% yield was purified by flash column chromatography using 79% methylene chloride, 20% methanol, 1% of a 10% ammo- nium hydroxide solution as the solvent system.

Carbamate 3 (1.0 eq, 0.22 mmol, 100 mg) was dissolved in 25 ml of dry THF at room temperature under N2 atmosphere. Solid lithium aluminum hydride (2.0 eq, 0.44 mmol, 17 mg) was added, and the reaction mixture was stirred at room temperature for 6 h. The reaction mixture was then cooled in an ice bath and quenched with a minimal amount of water. Ethyl acetate (approximately 100 ml) was then added followed by anhydrous magnesium sulfate. The reaction mix- ture was then filtered, and the solvents were evaporated in uucuo yielding N,N-dimethylsphingosine 4 in 82% yield. Purification was achieved by flash column chromatography using 79% methylene chloride, 20% methanol, 1% of a 10% ammonium hydroxide solution as the solvent system. 'H NMR (CDC1,) 6.88 (t, J = 6.6 Hz, 3H), 1.26 (br.s, 22 H), 1.30-1.41 (m, 2 H), 2.0-2.1 (m, 2 H), 2.38-2.42 (m, 1 H), 2.45 (s, 6 H), 2.47 (br.s, 2 H), 3.77 (d, J = 5.3 Hz, 2 H), 4.39 (t, J = 5.4 Hz, 1 H), 5.51-5.57 (m, 1 H), 5.73-5.81 (m, 1 H). Low resolution FAB mass spectrum; m/e 327.

Cell Culture-A431 human epidermoid carcinoma cells were main- tained in monolayer culture in a mixture of Dulbecco's modified Eagle's and Ham's F-12 medium (l:l, v/v) containing 10% fetal calf serum. Cells were passaged biweekly by trypsinization before reaching confluence.

Cell Studies-On the day prior to an experiment, cells were resus- pended in the above medium containing 5% neonatal calf serum and transferred to 6 well plates (3 X lo5 cells/well). On the day of the experiment, the medium was replaced with the same medium without phosphate containing '*Pi (0.4 mCi/ml). After 5 h, sphingosine, N,N- dimethylsphingosine, or C8-cer was added. Control incubations re- ceived diluent (0.1% EtOH). A t the indicated times, incubations were discontinued at 4 "C by removing media, washing twice with phos- phate-buffered saline (PBS), resuspending each well in lysis buffer (1.5 mM MgCl,, 1 mM EGTA, 50 mM HEPES, pH 7.5, 1% Triton (v/ v), 10% glycerol (v/v), 1 mM phenylmethylsulfonyl fluoride, 2 mM Na3V04, 10 pg/ml each of leupeptin and aprotinin), and scraping cells into microcentrifuge tubes. Cell lysates were centrifuged at 14,000 X g for 15 min and, after the particulate fraction was discarded, were stored at -70 "C.

Immunoprecipitation-Immunoprecipitating antibody 528 to the EGF receptor was utilized as adapted from Sunada et al. (18). Anti- body (50 fig) was first complexed to PAS (40 mg) by co-incubation in 20 mM HEPES, pH 7.5 for 1 h at 22 "C. After 1 h, the PAS.Ab complex was washed three times with buffer (HNTG: 20 mM HEPES, 150 mM NaC1, 0.1% Triton, 10% glycerol, pH 7.5) and incubated (3 mg) in 0.25 ml of lysis buffer at 4 "C with portions of cell lysates containing equal quantities of trichloroacetic acid-precipitable radio- activity. After 1 h, the PAS. Ab complex containing bound EGF receptor was washed three times with HNTG. EGF receptor was released into sample buffer (10% glycerol (v/v), 0.7 M P-mercaptoeth- anol, 3% SDS, 62.5 mM Tris-HC1 buffer, pH 6.8) containing brom- phenol blue by boiling at 110 "C for 10 min. EGF receptor was separated by SDS-polyacrylamide gel electrophoresis as described (18). Radiolabeled EGF receptor was visualized by autoradiography and quantified by liquid scintillation counting.

Phosphoamino Acid Two-dimensional Analysis of the EGF Recep- tor-Phosphoamino acid analysis of immunoprecipitated EGF recep- tor was performed by partial acid hydrolysis (1 h at 110 "C in 6 M HC1) and thin layer electrophoresis by the method of Hunter and co- workers (20). After removal of the acid by drying under vacuum,

16094 Ceramide-induced EGF Receptor Phosphorylation

hydrolysates were resuspended in 250 pl of HzO and applied to a Dowex AGl-X8 column (Bio-Rad). The absorbed "2P-labeled mate- rials were eluted with 0.5 M HCI and lyophilized. The recovery of radioactivity through the procedure was about 65%. "P-Labeled phosphoamino acids were analyzed by thin layer electrophoresis of an aliquot of each digest, as described (20). Individual phosphoamino acids were detected by ninhydrin staining of carrier phosphoamino acids, and radioactivity within each phosphoamino acid was measured via liquid scintillation chromatography.

Tyrosine Phosphorylation of the EGF Receptor in Intact Cells by Western Blot Analysis-Cells, treated with either C8-cer, sphingosine, or EGF as described above, were solubilized in Western solubilization buffer (20 mM HEPES, 1% Triton X-100,5 mM MgC12,120 mM KCI, 10% glycerol, 2 mM Na3V04, 1 mM phenylmethylsulfonyl fluoride, 10 pg/ml aprotinin, 10 pg/ml leupeptin). Lysates were mixed with sample buffer, and equal amounts of total protein were loaded onto SDS- PAGE. Proteins were transferred to nitrocellulose at 100 mA utilizing a Trans-Blot apparatus (Bio-Rad). The state of tyrosine phosphoryl- ation of the EGF receptors was investigated by using a monoclonal anti-tyrosine phosphate antibody (PY-69) and '"I-protein A as de- scribed (21). Total EGF receptor was detected with RKII polyclonal antibodies as described (22).

EGF Receptor Phosphopeptkie Analysis by High Performunce Liq- uid Chromutography (HPLC)-The bands containing EGF receptors, isolated by immunoprecipitation and polyacrylamide gel electropho- resis as described above, were cut from the gel, ground in 0.1 M NH4HCOn, and digested with trypsin for 18 h. The "'P-phosphopep- tides obtained were analyzed by reverse-phase HPLC using a C18 reverse-phase column (Dynamax 4.6 mm inside diameter, Rainin) equilibrated with 0.1% trifluoroacetic acid containing 0.05% trieth- ylamine as described (12). "'P-Phosphopeptides were eluted with a linear gradient (I%/min) of acetonitrile a t a flow rate of 1 ml/min. Fractions (0.5 ml) were collected, and the "P-phosphopeptides were detected by measuring associated Cerenkov radiation.

Lipid Studies-Cells were handled as above except media did not contain "'Pi. A t the end of each study, cells were washed with PBS and harvested by trypsinization, and lipids were extracted from cell pellets with 1 ml of ch1oroform:methanokHCI (100:1001, v/v) and 0.3 ml of a buffered saline solution (135 mM NaC1, 4.5 mM KCI, 1.5 mM CaCIz, 0.5 mM MgCl', 5.6 mM glucose, 10 mM HEPES, pH 7.2) containing 10 mM EDTA (1). Portions of the organic phase were subjected to mild alkaline hydrolysis (0.1 M methanolic KOH a t 37 "C for 1 h) to remove glycerophospholipids (1). Ceramides were resolved by TLC using Silica Gel G plates and CHC1,:MeOH:HAc (65:2.5:4, v/v) as solvent (3). The R, values for endogenous ceramide and N- octanoyl sphingosine were 0.6 and 0.5, respectively. Ceramides were detected by iodine vapor staining, eluted from the silica gel with CHCln:MeOH (l:l, v/v), and dried under Nz. The ceramide fractions were combined and deacylated to sphingoid bases in 6 N HCkbutanol (1:1, v/v) a t 100 "C for 1 h (23) and desiccated under reduced pressure prior to derivatization (24). Sphingoid base levels were determined by HPLC after derivatization with o-phthalaldehyde as described (25). Derivatized sphingoid bases were resolved by isocratic elution with MeOH:5 mM KP04, pH 7.5 (9010, v/v; 2 ml/min) using a reverse-phase C18 column and detected by fluorescence spectroscopy as described (25). Recoveries were assessed by standards carried throughout the isolation procedures. Individual values were deter- mined by comparison to a standard curve run concomitantly. The levels of ceramide obtained by this procedure were similar to that obtained by a radioenzymatic method utilizing diacylglycerol kinase (8,26).

Studies with P'H/Sphingosine-For long term labeling studies, cells were incubated in ['Hlsphingosine (0.14 pCi/ml medium) for 4 days prior to lipid isolation as above. Incorporation of "H into individual lipids was determined by TLC using CHC13:MeOH:NH40H (80202, v/v) as solvent as described (27). The R, values in this system are: lactocerebroside, 0.04; galactocerebroside, 0.20; glucocerebroside, 0.28; sphingosine, 0.43; N,N-dimethylsphingosine, 0.61; ceramide, 0.78. Short term labeling studies were performed similarly except cells received 1.4 pCi/ml ['Hlsphingosine.

Statistics-Statistical analysis was performed by t test and linear regression analysis by the method of least squares.

RESULTS AND DISCUSSION

Initial studies determined whether ceramide, like sphingoid bases (12-14), might serve to induce phosphorylation of the EGF receptor. These studies utilized cells resuspended for 5

h in media without phosphate containing "Pi (0.4 mCi/ml). Fig. 1 (upper panel) demonstrates an autoradiogram of a typical experiment. As shown, resting cells contain a meas- urable amount of radiolabeled EGF receptor. C8-cer (0.3-10 p ~ ) induced concentration-dependent incorporation of radi- olabel at 15 min of stimulation. The lower panel shows a compilation of data from four experiments quantified by liquid scintillation counting. As little as 0.1 pM C8-cer was effective, and a maximal effect to 1.9-fold of control occurred with 10 p~ C8-cer. Similar stimulation was achieved with a maximally effective EGF concentration (20 nM) to 2.1-fold of control ( n = 3). Ceramide-stimulated EGF receptor phosphorylation was 30% completed by 2 min and maximal by 10 min of stimula- tion. The level remained elevated for at least 30 min. These studies clearly demonstrate that ceramide stimulates phos- phorylation of the EGF receptor in A431 cells.

Subsequent studies assessed the effect of sphingosine on EGF receptor phosphorylation. Fig. 2 shows that sphingosine stimulated EGF receptor phosphorylation to a level 1.8-fold of control at 15 min of stimulation. This is similar to the level induced by C8-cer. The concentration dependence of this effect is also similar to that stimulated by C8-cer. These studies demonstrate that ceramide and sphingosine have very similar effects on EGF-receptor phosphorylation. Faucher et al. (12) similarly reported that sphingosine stimulated EGF receptor phosphorylation over 5-20 min in intact A431 cells.

Fig. 3 compares the effect of sphingosine to N,N-dimethyl- sphingosine. A controversy exists as to whether N,N-di- methylsphingosine, a potential derivative of sphingosine,

A 200-

92.5- - Conlrol,.03 . I .3 I 3 IO 30,EGF

Cer (pM)

Cer (pM)

FIG. 1. Effect of ceramide on phosphorylation of the EGF receptor. Upper panel, autoradiogram of an individual experiment. Lower panel, compilation of data from four experiments. On the day prior to an experiment, cells were resuspended in Dulbecco's modified Eagle's/F-12 medium containing 5% neonatal calf serum and trans- ferred to 6 well plates (3 X IO5 cells/well). On the day of the

phosphate containing "Pi (0.4 mCi/ml). After 5 h, C8-cer, EGF (20 experiment, media were replaced with the same medium without

nM), or diluent (0.1% EtOH) was added. After 15 min, incubations were discontinued at 4 "C by removing media, washing twice with PBS, resuspending wells in lysis buffer, and scraping cells into microcentrifuge tubes. Cell lysates were centrifuged at 14,000 X g for 15 min, and the particulate fraction was discarded. Equal quantities of trichloroacetic acid-precipitable counts from cell lysates were in- cubated at 4 "C with EGF receptor monoclonal antibody 528 bound to PAS in 0.25 ml of lysis buffer. After 1 h, the PAS. Ab complex containing bound EGF receptor was washed three times with PBS. EGF receptor was released into sample buffer by boiling at 110 "C for 10 min and resolved by SDS-polyacrylamide gel electrophoresis (18). Radiolabeled EGF receptor was visualized by auto- radiography and quantified by liquid scintillation counting.

Ceramide-induced EGF Receptor Phosphorylation 16095

N 10

3000 I 1 I I

0 .03 .3 3 Sphingosine (pM)

FIG. 2. Effect of sphingosine on the phosphorylation of the EGF receptor. These studies were performed as described in Fig. 1 except that cells received sphingosine. These data (mean) represent data from three experiments.

FIG. 3. Comparison of sphingosine (SS) and N,N-dimethyl- sphingosine (DMSS) on phosphorylation of the EGF receptor. Representative autoradiogram of one of three experiments.

might be more potent than sphingosine itself and perhaps a physiologic mediator of EGF receptor phosphorylation (27). This is based on the observation that N,N-dimethylsphingo- sine far more potently stimulated tyrosine phosphorylation of the EGF receptor in an in Vitro assay. Fig. 3 shows an autoradiogram of a typical experiment that compares the effect of sphingosine and N,N-dimethylsphingosine at three concentrations. Although N,N-dimethylsphingosine effec- tively induced phosphorylation of the EGF receptor, at all concentrations it was less effective than sphingosine (n = 3). At 0.3 p ~ , dimethylsphingosine enhanced phosphorylation by 20% and a maximal effect with 1 PM was only 31% above control, or less than half of the effect with sphingosine. Further increases in the N,N-dimethylsphingosine concentra- tion did not further enhance phosphorylation. Similarly, N- methylsphingosine and stearylamine were less effective than sphingosine. N-Methylsphingosine (0-1 p ~ ) was equally ef- fective as N,N-dimethylsphingosine, and stearylamine (0-1 p ~ ) was about half as effective. These studies clearly dem- onstrate sphingosine to be more effective than these other compounds in intact cells.

Since sphingosine and ceramide are interconvertible, an additional set of studies attempted to assess which compound might be the more likely regulator of EGF receptor phos- phorylation in these studies. These studies compared the cellular level of each compound attained during EGF receptor phosphorylation. Ceramide levels were measured by HPLC after deacylation to sphingosine and derivatization by o- phthalaldehyde. Table I shows that A431 cells contain 12 pmol. 10' cells" of sphingosine and 132 pmol- 10' cells" of ceramide. Similar levels have been reported in a variety of other cells for these compounds (1, 3, 4). C8-cer induced a dose-responsive increase in the cellular ceramide level to 232 pmol.10' cells" over the range of concentrations found to effectively stimulate EGF receptor phosphorylation. C8-cer failed to increase sphingosine levels over this same concentra- tion range at any time between 1 and 15 min of stimulation.

TABLE I Effect of exogenous ceramide and sphingosine on cellular ceramide

and sphingosine levels Cells, prepared as described in Fig. 1, were incubated with sphin-

gosine, C8-cer, or diluent (0.1% EtOH). After 10 min, cells were washed twice with PBS, harvested by trypsinization, and placed on ice. Cells were counted by a Coulter Counter, centrifuged at 1,000 x g for 5 min, and extracted with 1 ml of CHCI,:MeOH (l:l, v/v) and 0.3 ml of buffered saline solution containing 10 mM EDTA. Portions of the organic phase were subjected to mild alkaline hydrolysis (0.1 M methanolic KOH a t 37 "C for 1 h) as described (1). Samples were re-extracted and derivatized with o-phthaldehyde as described (25). For studies that measured ceramide levels, ceramide was isolated by TLC and deacylated by acid butanol hydrolysis as described (23) prior to derivatization. Derivatized samples were resolved by HPLC using a C18 reverse-phase column and an isocratic elution with MeOH, 5 mM KP04, pH 7.5 (9010, v/v; 2 ml/min) as described (25). Sphingosine (SS) levels were determined from a standard curve run concomitantly, and recovery was assessed by carrying standards throughout the isolation procedure. These data (mean) represent values from four separate experiments.

Stimulant Cellular ceramide Cellular sphingosine pmol. IO6 cells"

Control 132 12 Cer, 0.1 p~ 183 15 Cer, 1 p~ 232 13 Cer, 10 p~ 306 25 ss, 0.1 pM 145 12 ss, 0.3 pM 162 12 ss, 1 pM 227 20 ss, 10 pM 232 326

Hence, these studies suggest that ceramide itself stimulates EGF receptor phosphorylation. Sphingosine, within the range of effective concentrations for EGF receptor phosphorylation (0.1-1.0 p ~ ) , failed to substantially increase its own level but instead increased the ceramide level. The maximally effective sphingosine concentration (1.0 p ~ ) increased the ceramide level 1.7-fold of control to 227 pmol. 10' cells", a value similar to that achieved with maximal ceramide? Higher concentra- tions of sphingosine that induced no further increase in EGF receptor phosphorylation produced the very large elevations in cellular sphingosine levels observed in other cell types (28). These studies support the notion that sphingosine may act, in part, by conversion to ceramide. In contrast, N,N-di- methylsphingosine did not affect cellular ceramide levels.

Studies were also performed to assess conversion of sphin- gosine to N,N-dimethylsphingosine. These studies utilized [3H]sphingosine to probe these pools. Sphingosine and N,N- dimethylsphingosine were isolated by TLC as described by Igarashi et al. (27). In cells labeled for 4 days with [3H] sphingosine, the free sphingoid base contained %fold as much radioactivity as N,N-dimethylsphingosine (13,000 dpm versus 4,600 dpm. 10' cells", respectively). Further, during the short term incubations (5-15 min) sufficient for EGF receptor phosphorylation, less than 5% of the cellular uptake of ["HI sphingosine (0.06-10 PM) was converted to ['H]N,N-di- methylsphingosine (n = 3). These studies demonstrate that N,N-dimethylsphingosine is neither as potent as sphingosine in EGF receptor phosphorylation nor a major metabolite of sphingosine in intact A431 cells.

To determine the changes in amino acid phosphorylation in response to ceramide or sphingosine treatment, phospho- amino acid analysis of the EGF receptor was performed. It is shown in Fig. 4 and Table 11, and in agreement with previously reported studies (12), that sphingosine primarily induced an increase in threonine phosphorylation. Treatment with C8-

* Conversion of 1 p~ sphingosine into ceramide occurred with the following kinetics: 20% at 1 min, 60% a t 10 min, and 100% at 15 min.

16096 Ceramide-induced EGF Receptor Phosphorylation

A. B. C

0. Control Ceramlde Sphingosine

FIG. 4. Phosphoamino acid two-dimensional analysis of the EGF receptor. A431 cells were labeled in phosphate-free medium with [:"PP]orthophosphate and were untreated ( A ) or treated with 3 p~ ceramide ( B ) or 3 p~ sphingosine ( C ) for 15 min as described in Fig. 1. Subsequently, cells were solubilized and the EGF receptor was isolated by immunoprecipitation and 7% polyacrylamide gel electro- phoresis. The isolated receptors were eluted from the gel and subjected to partial acid hydrolysis. The "P-phosphoamino acids were resolved by two-dimensional thin layer electrophoresis as described under "Experimental Procedures," localized by autoradiography, and iden- tified by ninhydrin staining of the phosphoserine (S), phosphothre- onine (T), and phosphotyrosine ( Y ) . The amino acids were scraped and radioactivity was assayed by scintillation counting. The quanti- tative data are presented in Table 11.

TABLE I1 Phosphoamino acid two-dimensional analysis

Phosphoamino acid analysis of immunoprecipitated EGF receptors was performed as described in Fig. 4. The proportion of each phos- phoamino acid to the three total phosphoamino acids was determined for each group. The values in parentheses represent the effect of treatment on this proportion expressed as a percentage of control.

Experiment Total P-Tyr P-Thr P-Ser cprn" cpm CPm cPm

Control 2832 340 (100%) 566 (100%) 1926 (100%) Ceramide 3824 234 (50%) 1090 (140%) 2570 (97%) Sphingosine 3540 212 (50%) 1168 (165%) 2159 (90%) EGF 6607 1520 (192%) 2180 (165%) 2907 (65%) " :'jP counts/min are normalized for a total yield of 12 ? 1%.

A B C D

FIG. 5. Western blot analysis of tyrosine-specific phos- phorylation of EGF receptor. A431 cells were untreated ( A ) or treated with 20 nM EGF ( B ) , 3 p~ ceramide ( C ) , or 3 p M sphingosine (D) for 15 min, as described in Fig. 1. Lysates were split in half for blotting with anti-P-Tyr antibody (PY-69) and with polyclonal anti- EGF receptor antibody (RK-11) (not shown). Equal quantities of total protein were separated by 7% SDS-PAGE, transferred to nitrocellu- lose, immunoblotted, and then detected with '"'I-protein A. The nitrocellulose membrane was autoradiographed for 24 h, and the hands were cut and quantified via a y-counter.

cer also caused a major increase, and to a similar extent as with sphingosine, in threonine phosphorylation of the EGF receptor. Interestingly, tyrosine phosphorylation of the EGF receptor was reduced by either ceramide or sphingosine treat- ment (Fig. 4 and Table 11).

The effects of ceramide and sphingosine on EGF receptor tyrosine phosphorylation were further substantiated by West- ern blot analysis with anti-phosphotyrosine antibodies as shown in Fig. 5. After 15 min of incubation with either 3 PM sphingosine, 3 PM C&cer, or 20 nM EGF, A431 cells were lysed, and equal total protein was subjected to SDS-PAGE. Half of the blot after the Western transfer was immuno- decorated with RK-I1 anti-EGF receptor polyclonal antibod- ies. Equivalent levels of EGF receptor were detected by this method (data not shown). The other half was immunodeco- rated with PY-69, an anti-phosphotyrosine monoclonal anti- body. A slight decrease in tyrosine phosphorylation was ob-

served by ceramide and sphingosine treatment. Only EGF caused an increase in tyrosine phosphorylation of EGF recep- tors, as expected.

To further characterize the sites of phosphorylation of the EGF receptor as affected by ceramide and sphingosine, the phosphorylation state of the receptor was investigated after EGF receptors were digested with trypsin. The "P-phospho- peptides obtained were analyzed by reverse-phase HPLC (Fig. 6). The maps obtained for the EGF receptor from cells treated with either ceramide or sphingosine showed an increase in phosphorylation of peptides eluted at 17% acetonitrile and a decrease in the phosphorylation of fractions eluted at 23-25% acetonitrile (Fig. 6A). Moreover, the HPLC peptide maps obtained by ceramide or sphingosine were superimposable (Fig. 6B).

Hence, two-dimensional phosphoamino acid analysis, Western blotting with an anti-phosphotyrosine antibody, and HPLC analysis of tryptic peptides all showed that ceramide and sphingosine induced similar changes in the phosphoryl-

2 1.9

1.a - 1.7 - 1.6 - 1 s

1.4 - 1.3

1.2 1.1

- - - -

1 -

0.9 - 0.a - 0.7 - 0.6 -

A

0 CONTROL + CIJ-CEWIOE 0 SPHINCOSINE

x 1C.lMiItil.

'-1 600 B A

0 5 IO 15 20 25 30

X Acetonitrile

+ UI-CERIUIDE 0 SPHlNC3SINE

FIG. 6. Characterization of EGF receptor phosphopeptides by HPLC. [:"P]Phosphate-labeled EGF receptors were prepared as described in the legend to Fig. 1. The gel slices containing EGF receptors were excised and digested for 18 h with 100 pg of trypsin as described under "Experimental Procedures." The "P-phosphopep- tides were analyzed by reverse-phase HPLC and detected after elution from a CIS column with a linear gradient of acetonitrile by measuring the associated Cerenkov radiation. Panel A, control, ceramide- and sphingosine-treated EGF receptor. Panel B, superimposition of cer- amide and sphingosine phosphopeptide maps. Similar results were obtained in three separate experiments. The small differences be- tween HPLC profiles of ceramide and sphingosine were not repro- ducible and likely represent some variability in sample collection.

Ceramide-induced EGF Receptor Phosphorylation 16097

ation sites of the EGF receptor. Further, these studies dem- onstrate that these agents enhance threonine and reduce tyrosine phosphorylation.

The present studies demonstrate that ceramide, sphingo- sine, and N,N-dimethylsphingosine stimulate EGF receptor phosphorylation in intact A431 cells. The effects of ceramide and sphingosine appear quite similar. Ceramide appears to act independently of conversion to sphingosine. In contrast, sphingosine is rapidly converted to ceramide under conditions of EGF receptor phosphorylation. The most parsimonious explanation for these results is that both agents act to stim- ulate EGF receptor phosphorylation via elevation of the cel- lular level of ceramide. However, direct evidence for this supposition will have to await identification of the protein kinase involved in EGF receptor phosphorylation. At levels that induce no additional increase in EGF receptor phos- phorylation, the cellular sphingosine level rises substantially. This is consistent with the sphingoid base load overwhelming the capacity of the system for acylation. N,N-Dimethylsphin- gosine was also found to be effective, although less so than ceramide or sphingosine. Since this compound was not gen- erated within the time frame of these studies, it is unlikely that it mediates sphingosine-induced EGF receptor phos- phorylation.

These studies do not rule out a role for free sphingoid bases as direct biomodulators independent of their role to inhibit protein kinase C. Not only do N-methylsphingosine, N,N- dimethylsphingosine and stearylamine possess some sphin- gosine-like properties in the present studies, but in a t least two prior instances the effects of sphingoid bases in intact cells have been replicated, in part, in isolated membranes. This includes EGF receptor phosphorylation of A431 and WI- 38 cells (12, 13) and inhibition of thyrotropin-releasing hor- mone binding to GHs cells (9, 10). Interestingly, in both instances, the effects of sphingoid bases in uitro only partially replicated the effect in intact cells. In the case of EGF receptor phosphorylation, the predominant effect of sphingoid bases in intact cells, as in the present study, was to phosphorylate threonine residues. In contrast, the effect in uitro was almost exclusively to induce tryosine phosphorylation. Similarly, sphingoid bases inhibited thyrotropin-releasing hormone binding to intact cells by a mechanism that involved changes in both receptor affinity and number whereas the effect in membranes from these cells was only to decrease the apparent binding affinity. These effects in uitro presumably occur by direct action of the sphingoid bases since these studies are performed under conditions in which acylation to ceramide would be unlikely. Additional studies will have to be per- formed to compare the effects of these compounds in intact cells and in uitro and to fully determine conditions for their interconversion. Some of the differences in the effects of sphingosine and N,N-dimethylsphingosine between the pres- ent studies, those of Davis et al. (12-14) and Hakomori and co-workers (27), may reflect whether studies were performed in intact cells or in uitro.

In sum, these studies demonstrate that ceramide and sphin- gosine induce similar phosphorylation of the EGF receptor. Since at effective concentrations these compounds result in elevation of cellular ceramide but not sphingoid base levels, these studies suggest that ceramide may mediate some of the biologic effects of free sphingoid bases. Further, the ceramide elevation is within the range achieved during lq25-dihy-

droxyvitamin DB, tumor necrosis factor-a, interferon-y (4-6), and phorbol ester3-induced monocyte/macrophage differen- tiation in HL-60 cells. These studies suggest that ceramide might act, under the appropriate conditions, as a second messenger. In this regard, investigations are underway to identify the ceramide-activated protein kinase which mediates EGF receptor phosphorylation.

Acknowledgments-We would like to thank Dr. J. Schlessinger for kindly providing the RKII polyclonal antibodies, and Margaret Prid- dle and Mary Hemer for expert technical assistance.

1. 2.

3. 4.

5.

6.

7. 8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

23. 24.

25.

26.

27.

28.

-

REFERENCES Kolesnick, R. N. (1987) J. Biol. Chem. 262, 16759-16762 Kolesnick, R. N., and Clegg, S. (1988) J. Biol. Chem. 263,6534-

Kolesnick, R. N. (1989) J. Biol. Chem. 264 , 7617-7623 Okazaki, T., Bell, R. M., and Hannun, Y. A. (1989) J. Biol. Chem.

Okazaki, T., Bielawaska, A., Bell, R. M., and Hannun, Y. A.

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

Hannun, Y., and Bell, R. M. (1987) Science 235,670-674 Dressler, K. A., and Kolesnick, R. N. (1990) J. Biol. Chem. 265 ,

Winicov, I., and Gershengorn. M. C. (1988) J. Bwl. Chem. 263.

6537

264,19076-19080

(1990) J. Biol. Chem. 265 , 15823-15831

Biol. Chem. 266,484-489

14917-14921

12179L12182 - .

Winicov. I.. Corv. R. N.. and Gersheneorn. M. C. (1990) Endo- crimlogy'126,"1668-1672

~I

Jefferson, A. B., and Schulman, H. (1988) J. Biol. Chem. 263 ,

Faucher, M., Girones, N., Hannun, Y. A,, Bell, R. M., and Davis,

Davis, R. J., Girones, N., and Faucher, M. (1988) J. Biol. Chem.

Northwood, I. C., and Davis, R. J. (1988) J. Biol. Chem. 263,

Kishimoto, Y. (1983) in The Enzymes (Boyer, P.D., ed) pp. 357-

Nishizuka, Y. (1984) Nature 308, 693-698 Vunnam, R. R., and Radin, N. S. (1979) Biochim. Biophys. Acta

Sunada, H., MacLeod, C., and Mendelsohn, J. (1985) Anal. Biochem. 149,438-447

Nimkar, S., Menaldino, D., Jamison, W. C. L., Merrill, A. H., Jr., and Liotta, D. (1988) Tetrahedron Lett. 29 , 3037-3040

Cooper, J. A., Septon, B. B., and Hunter, T. (1983) Methods Enzymol. 99, 387-402

Zipperl, R., Sturani, E., Toschi, L., Naldini, L., Alberguna, L., and Comoglio, P. M. (1986) Biochim. Biophys. Acta 881, 54- 61

Kris, R. M., Lax, I., Gullick, W., Waterfield, M. D., Ullrich, A., Fridkin, M., and Schlessinger, 3. (1985) Cell 40 , 619-625

Kaller, S. (1961) Biochem. 2. 334,451-456 Slife, C. W., Wang, E., Hunter, R., Wang, S., Burgess, C., Liotta,

D. C., and Merrill, A. H., Jr. (1989) J. Bwl. Chem. 264,10371- 10377

Merrill, A. H., Jr., Wang, E., Mullins, R. E., Jamison, W. C. L., Nimkar, S., and Liotta, D. C. (1988) Anal. Biochem. 171,373- 381

15241-15244

R. J. (1988) J. Biol. Chem. 263,5319-5327

263,5373-5379

7540-7543

407, Academic Press, New York

573 , 73-82

Schneider, E. G., and Kennedy, E. P. (1973) J. Biol. Chem. 248 , 3739-3741

Igarashi, Y., Kitamura, K., Toyokuni, T., Dean, B., Fenderson, B., Ogawa, T., and Hakomori, S.-I. (1990) J. Bwl. Chem. 265 ,

Wilson, E., Wang, E., Mullins, R. E., Uhlinger, D. J., Liotta, D. C., Lambeth, J. D., and Merrill, A. H., Jr. (1988) J. Biol. Chem.

5385-5389

263,9304-9309 - R. N. Kolesnick and K. A. Dressler, unpublished observations.