Ceramide Stimulates Epidermal Growth Factor Receptor ...

3 downloads 0 Views 6MB Size Report
Ceramide Stimulates Epidermal Growth Factor Receptor. Phosphorylation in A43 1 Human Epidermoid Carcinoma Cells. EVIDENCE THAT CERAMIDE MAY ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

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 CarcinomaCells EVIDENCETHATCERAMIDE

MAY MEDIATESPHINGOSINE ACTION* (Received for publication, February 5, 1991)

Tzipora GoldkornS, Kenneth A. DresslerS, Josephia MuindiS, Norman S. RadinQ,John MendelsohnS, David Menaldinoll, Dennis Liottaq, and RichardN. KolesnickS From the $Division of Molecular Pharmacology and Therapeutics, The Sloan Kettering Institute, Cornell University Medical College, New York, New York 10021, the §Mental HealthResearch 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- fective concentrations, both compounds elevated celduction pathway involving sphingomyelin and deriv- lular ceramide but not sphingosine levels. Additional atives (Kolesnick, R.N. (1989) J. Biol. Chern. 264, studies performed with['Hlsphingosine demonstrated 7617-7623). The present studies compare effects of that cells containsubstantially less N,N-dimethylceramide, sphingosine, and N,N-dimethylsphingosine sphingosine than free sphingosine and, during short on epidermalgrowthfactor (EGF) receptor phos- term incubation, convertless than 6%of added sphinphorylationinA431 human epidermoid carcinoma gosine to N,N-dimethylsphingosine.These studies procells. To increase ceramidesolubility, a ceramide con- vide evidence that ceramide may have bioeffector taining octanoic acid at the second position (CS-cer) properties andsuggest sphingosine may act in part by was synthesized. CS-cer induced time- and concentra- conversion to ceramide. tion-dependent EGF receptor phosphorylation. This event wasdetectable by 2 min and maximal by 10 min. As little as 0.1 ~ L CS-cer M was effective, and 3 p~ CSRecent investigations from this and otherlaboratories have cer induced maximal phosphorylation to 1.9-fold of suggested the existence of a signal transduction pathway control. EGF (20 nM) increased phosphorylation to 2.1- involving sphingomyelin and its derivatives (1-6). Such a fold of control. Sphingosine stimulated receptor phosphorylation over thesame concentration range (0.03- pathway was initially proposed by this laboratory to involve 3 p ~ and ) to the same extent (1.8-fold of control) as sphingomyelin degradation to ceramide via the action of a ceramide. The effectsof CS-cer and sphingosine were sphingomyelinase. Evidence was presented that ceramide was similar by three separate criteria,phosphoamino acid converted to sphingoid bases, potential inhibitors of protein analysis, anti-phosphotyrosine antibody immunoblot- kinase C (7). Under these conditions, phorbol ester-induced ting, and phosphopeptide mapping by high perform- activation of protein kinase C was antagonized and differenoc- tiation of human leukemia (HL-60) cells into macrophages ance liquid chromatography. Phosphorylation curred specifically on threonine residues. N,N-Di- was prevented. Subsequently, it has been shown that vitamin methylsphingosine, a potential derivative of sphingo- D, interferon-y, and tumor necrosis factor-astimulatea sine, wasless effective. Since sphingosine and ceram- sphingomyelinase in HL-60 cells and that this event may, in ide are interconvertible, the level of each compound part, mediate monocyte differentiation via these agents (4). was measured under conditions sufficientfor EGF Further support for a potential pathway from sphingomyelin receptor phosphorylation. CS-cer (0.1-1 WM) induced derives from data thatdemonstrate the existence of ceramide 1-phosphate (8).This phosphorylated form of ceramide was dose-responsive elevation of cellular ceramide from generated selectively from ceramide derived from sphingo132 to 232 pmol* 10'cells". In contrast, cellular sphingosine levels did not rise. This suggests that CS-cer myelin but notglycosphingolipids.These studiesprovide early acts without conversion to sphingosine. Exogenous evidence for a biological pathway involving sphingomyelin ) increased cellular ceram- and its derivatives. sphingosine (0.1-1 p ~ also The purpose of the present studies is to determine whether ide levels to 227 pmol*10' cells-', but did not increase its own cellular level of 12 pmol.10' cells-'. Higher ceramide may serve as a bioeffector molecule. In these studies, sphingosine concentrations that induced no further in- the effects of ceramide are compared to sphingosine and crease inEGF receptor phosphorylation produced very N,N-dimethylsphingosine.Recent investigations have demlarge elevations in cellular sphingosine. Hence, at ef- onstrated that sphingoid bases may induce a variety of biologic events independent of inhibition of protein kinase C. These include inhibition of thyrotropin-releasing hormone * This work was supported by National Institutes of Health Grant R01-CA-42385 and American Cancer Society Grant FRA-345 (to R. binding to GH, pituitary cells (9, lo), inhibition of several N. K.), United States Public Health Service Grants NS 03192 and calmodulin-dependent enzymes ( l l ) , and phosphorylation of HD 07406 (to N. S. R.), National Science Foundation Grant DCB- the epidermal growth factor (EGF)' receptor from A431 hu8710283 (to D. L.), and National Institutes of Health Grant CA42060 (to J. M.). K. A. D. was supported by Clinical Scholar Grant CA-09512-06 (Brian Piccolo Cancer Research Fund) from the NationalInstitutes 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 18U.S.C. Section 1734 solelyto indicate this fact.

The abbreviations used are: EGF, epidermal growth factor; C8cer, N-octanoyl sphingosine; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic 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 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 fattyacids at the second position and arequite 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-diacylglyceroleffects on protein kinase C . This problem was overcome by the use of synthetic diacylglycerols containingfatty acids of shorter chain length (16). Similarly, the presentstudies utilize a synthetic ceramide containing octanoic acid at the second position and compare the effects of this compound to sphingosine 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 similar effects and had similar concentration dependencies. Further, both compounds selectively induced threonine phosphorylation and generated identical phosphopeptide maps. N,N-Dimethylsphingosine was slightly less effective. At effective concentrations, virtually all of the sphingosine was converted 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 bioeffector molecule. Further, these studies suggest that some of the biologic effects of exogenous sphingoid bases maybe 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 ketone-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 PontNew England Nuclear and repurified by TLC prior to utilization. Protein A-Sepharose (PAS) CL-4B was from Pharmacia LKB Biotechnology Inc. Fetal and neonatal calf serum were from Gibco Laboratories. Ceramide containing octanoic acid at the second position (N-octanoyl sphingosine termed C8-cer) was prepared from Dsphingosine as previously described (17). Anti-phosphotyrosine antibody 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 belowwere 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 consistentwith the results obtained. OH

HO

OH

2

HO OH

n 0

HO OH

4

3 N-Butoxycarbonyl dehydrosphingosine 1 was prepared as described 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) wasadded slowly, and the reaction mixture was refluxed for 24 h or until the reaction was complete. The reaction was monitored

16093

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 (approximately 100 ml), washed with water and brine, dried over anhydrous magnesium sulfate, and concentrated in uacuo. The resulting carbamate 3 in 90% yield was purified by flash column chromatography using 79% methylene chloride, 20% methanol, 1%of a 10% ammonium hydroxide solution as thesolvent system. Carbamate 3 (1.0 eq, 0.22 mmol, 100 mg) wasdissolved 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 andquenched with a minimal amount of water. Ethyl acetate (approximately 100 ml) was then added followed by anhydrous magnesium sulfate. The reaction mixture 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 thesolvent 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 maintainedin monolayer culture in a mixture of Dulbecco'smodified 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 resuspended 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,Ndimethylsphingosine, or C8-cer was added. Control incubations received diluent (0.1% EtOH). A t the indicated times, incubations were discontinued at 4 "C by removing media, washing twice with phosphate-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). Antibody (50 fig) was first complexed to PAS (40 mg) by co-incubation in 20 mM HEPES, pH 7.5 for 1 h a t 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 0.25mlof lysis buffer at 4 "C with portions of cell lysates containing equal quantities of trichloroacetic acid-precipitable radioactivity. 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-mercaptoethanol, 3% SDS, 62.5 mM Tris-HC1 buffer, pH 6.8) containing bromphenol 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 Receptor-Phosphoamino acid analysis of immunoprecipitated EGF receptor was performed by partial acid hydrolysis (1 h at 110 "C in 6 M HC1) and thinlayer electrophoresis by the method of Hunter and coworkers (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 materials 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 a n 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 Westernsolubilization buffer (20 mM HEPES, 1%Triton X-100,5mM 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 SDSPAGE. Proteins were transferred to nitrocellulose a t 100 mA utilizing a Trans-Blot apparatus(Bio-Rad). The stateof tyrosine phosphorylation of the EGF receptors was investigated by using a monoclonal anti-tyrosine phosphate antibody (PY-69) and '"I-protein A as described (21). Total EGF receptor was detected with RKII polyclonal antibodies as described (22). EGF Receptor Phosphopeptkie Analysis by High Performunce Liquid Chromutography(HPLC)-The bands containingEGF receptors, isolated by immunoprecipitation and polyacrylamide gel electrophoresis as described above, were cut from the gel, ground in 0.1 M NH4HCOn,and digested with trypsin for 18 h. The "'P-phosphopeptides obtained were analyzed by reverse-phase HPLC using a C18 reverse-phase column (Dynamax 4.6mm inside diameter, Rainin) equilibrated with 0.1% trifluoroacetic acid containing 0.05% triethylamine 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 mlof 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) toremove 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 Noctanoyl 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 "Cfor 1h (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; 2ml/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 determined 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. RESULTSANDDISCUSSION

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 measurable amount of radiolabeled EGF receptor. C8-cer (0.3-10 p ~ induced ) concentration-dependent incorporation of radiolabel at 15 min of stimulation. The lowerpanel shows a compilation of data from fourexperiments 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-stimulatedEGF receptor phosphorylation was 30% completedby 2 min and maximal by 10 min of stimulation. The level remained elevated for at least 30 min. These studies clearly demonstrate that ceramide stimulates phosphorylation 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 thelevel induced by C8-cer. The concentration dependence of this effect is also similar to that stimulated byC8-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-dimethylsphingosine. A controversy exists asto whether N,N-dimethylsphingosine, a potential derivative of sphingosine, A

200-

92.5-

-

Conlrol,.03 .I

.3 I 3 IO 30,EGF

Cer (pM)

Cer ( p M )

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 anexperiment, cells were resuspended in Dulbecco's modified Eagle's/F-12 medium containing 5% neonatal calf serum and transferred to 6 well plates(3 X IO5 cells/well). On the day of the experiment, media were replaced with the same medium without phosphate containing "Pi (0.4 mCi/ml). After 5 h, C8-cer, EGF (20 nM), or diluent (0.1% EtOH) was added. After 15 min, incubations were discontinued a t 4 "C by removing media, washing twice with PBS, resuspending wells in lysis buffer, and scraping cells into microcentrifuge tubes. Cell lysates were centrifuged a t 14,000 X g for 15 min, and the particulate fraction was discarded. Equal quantities of trichloroacetic acid-precipitable counts from cell lysates were incubated a t 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 autoradiography and quantified by liquid scintillation counting.

Ceramide-inducedEGF Receptor Phosphorylation

16095 TABLE I

N 10

3000 I

1

0 .03

I

I

.3 3 Sphingosine ( p M )

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.

Effect of exogenous ceramide and sphingosine on cellular ceramide and sphingosine levels Cells, prepared as described in Fig. 1, were incubated with sphingosine, 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 a t 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) asdescribed (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 andan isocratic elution with MeOH, 5 mM KP04, pH7.5 (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

12

20 FIG. 3. Comparison of sphingosine (SS)and N,N-dimethylsphingosine (DMSS)on phosphorylation of theEGF receptor. 326 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-dimethylsphingosine 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 effectively induced phosphorylation of the EGF receptor, at all concentrations it was less effectivethan 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-dimethylsphingosineconcentration did not further enhance phosphorylation. Similarly, Nmethylsphingosine and stearylamine were less effective than sphingosine. N-Methylsphingosine (0-1 p ~ was ) equally effective as N,N-dimethylsphingosine,and stearylamine (0-1 p ~ was ) about half as effective. These studies clearly demonstrate sphingosine to be more effective than these other compounds inintact 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 phosphorylation 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 ophthalaldehyde. 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 sphingosinelevels overthis same concentration range at any time between 1 and 15 min of stimulation.

Control Cer, 0.1 p~ Cer, 1 p~ Cer, 10 p~ ss,0.1 p M ss,0.3 p M

ss, 1 pM ss, 10 p M

Cellular ceramide Cellular pmol. IO6 cells"

132 183 232 306 145 162 227 232

sphingosine

15

13 25 12 12

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-foldof control to 227 pmol. 10' cells", a value similar to that achieved with maximal ceramide? Higher concentrations of sphingosine that induced no further increase in EGF receptor phosphorylation produced the very large elevations in cellular sphingosine levels observedin other cell types (28). These studies support the notion that sphingosine may act, in part, by conversion to ceramide. In contrast, N,N-dimethylsphingosinedid not affect cellular ceramide levels. Studies were also performed to assess conversion of sphingosine to N,N-dimethylsphingosine.These studies utilized [3H]sphingosineto probe these pools. Sphingosine and N,Ndimethylsphingosine were isolated by TLC as described by Igarashi et al. (27). In cellslabeledfor 4 dayswith [3H] sphingosine, the free sphingoid base contained %fold as much radioactivity as N,N-dimethylsphingosine(13,000 dpmversus 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-dimethylsphingosine (n = 3). These studies demonstrate that N,N-dimethylsphingosineis 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, phosphoamino acid analysis of the EGF receptor was performed. It is shown in Fig. 4 andTable 11, and in agreement with previously reported studies (12), that sphingosine primarily induced an increase in threonine phosphorylation. Treatment with C8-

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

Ceramide-induced EGF Phosphorylation Receptor

16096 B.

A.

C

0. Control SphingosineCeramlde FIG. 4. Phosphoamino acidtwo-dimensional analysis of the EGF receptor. A431 cells were labeled in phosphate-free medium with [:"PP]orthophosphateand 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 solubilizedand the EGFreceptor was isolated by immunoprecipitation and 7% polyacrylamide gel electrophoresis. 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 identified by ninhydrin staining of the phosphoserine (S), phosphothreand phosphotyrosine ( Y ) .The amino acids were scraped onine (T), and radioactivity was assayed by scintillation counting. The quantitative 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 phosphoamino acid to thethree totalphosphoamino 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

served by ceramide and sphingosine treatment. Only EGF caused an increase in tyrosine phosphorylation of EGF receptors, 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-phosphopeptides obtained were analyzedby reverse-phase HPLC (Fig. 6). The maps obtained for the EGF receptor from cellstreated 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 weresuperimposable (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 phosphoryl2 1.9 1.a

-

0.a 0.6 -

A

1.7 1.6

1 s 1.4 1.3

1.2 1.1

1 -

0.9 0.7

D x 0

FIG. 5. Western blot analysis of tyrosine-specific phosphorylation of EGF receptor. A431 cells were untreated ( A ) or treated with 20 nM EGF ( B ) ,3 p~ ceramide ( C ) ,or 3 p M sphingosine

'-1

CONTROL

+

1C.lMiItil.

CIJ-CEWIOE

( 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 antiEGF receptor antibody (RK-11) (not shown). Equal quantities of total protein were separated by 7% SDS-PAGE, transferred to nitrocellulose, immunoblotted, andthen detected with '"'I-protein A. The nitrocellulose membrane was autoradiographed for 24 h, and the hands were cut andquantified 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 byeither ceramide or sphingosine treatment (Fig. 4 and Table 11). The effects of ceramide and sphingosine on EGF receptor tyrosine phosphorylationwere further substantiatedby Western 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 immunodecorated with RK-I1 anti-EGF receptor polyclonal antibodies. Equivalent levels of EGF receptor were detected by this method (data not shown). The other half was immunodecorated with PY-69, an anti-phosphotyrosine monoclonal antibody. A slight decrease in tyrosine phosphorylation was ob-

0

SPHINCOSINE

A

B

600

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-phosphopeptides were analyzed by reverse-phase HPLC anddetected after elution from a CIScolumn 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 ceramide and sphingosine phosphopeptide maps. Similar results were obtained in three separate experiments. The small differences between HPLC profiles of ceramide and sphingosine were not reproducible and likely represent some variability in sample collection.

Ceramide-induced EGF Receptor Phosphorylation ation sites of the EGF receptor. Further, these studies demonstratethat theseagentsenhancethreonine and reduce tyrosine phosphorylation. The present studies demonstrate that ceramide, sphingosine, 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 stimulate EGF receptor phosphorylation via elevation of the cellular 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 phosphorylation, the cellular sphingosine level rises substantially. This is consistent with the sphingoid base load overwhelming the capacity of the system for acylation. N,N-Dimethylsphingosine was also found to be effective, although less so than ceramide or sphingosine. Since this compound was not generated within the time frame of these studies, it is unlikely that it mediates sphingosine-induced EGF receptor phosphorylation. These studiesdo not rule out arole for free sphingoid bases as direct biomodulators independent of their role to inhibit protein kinase C. Not only do N-methylsphingosine, N,Ndimethylsphingosine and stearylamine possess some sphingosine-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 WI38 cells (12, 13) and inhibition of thyrotropin-releasing hormone binding to GHs cells (9, 10). Interestingly, in both instances, the effects of sphingoid bases in uitro only partially replicated the effect in intactcells. 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 inhibitedthyrotropin-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 wouldbe unlikely. Additional studies will have to be performed 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 present 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 sphingosine 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, theceramide elevation is within the range achieved during lq25-dihy-

16097

droxyvitamin DB,tumor necrosis factor-a, interferon-y (4-6), and phorbol ester3-induced monocyte/macrophage differentiation in HL-60 cells. These studies suggest that ceramide might act, under the appropriate conditions, asa second messenger. Inthis 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 RKIIpolyclonal antibodies, and Margaret Priddle and Mary Hemer for expert technical assistance.

REFERENCES 1. Kolesnick, R. N. (1987) J. Biol. Chem. 262, 16759-16762 2. Kolesnick, R. N., and Clegg, S. (1988) J. Biol. Chem. 263,65346537 3. Kolesnick, R. N. (1989) J. Biol. Chem. 2 6 4 , 7617-7623 4. Okazaki, T., Bell, R. M., and Hannun,Y. A. (1989) J. Biol. Chem. 264,19076-19080 5. Okazaki, T., Bielawaska, A., Bell, R.M., and Hannun, Y. A. (1990) J. Biol. Chem. 2 6 5 , 15823-15831 6. Kim, M.-Y., Linardic, C., Obeid, L., and Hannun, Y. (1991) J. Biol. Chem. 266,484-489 7. Hannun, Y., and Bell, R. M. (1987) Science 235,670-674 8. Dressler, K. A., and Kolesnick, R.N. (1990) J. Biol. Chem. 2 6 5 , 14917-14921 9. Winicov, I., and Gershengorn. . M. C. (1988) J. Bwl. Chem. 263. 12179L12182 10. Winicov. I.. Corv. R. N.. and Gersheneorn. M.C. (1990) Endocrimlogy'126,"1668-1672 11. Jefferson, A. B., and Schulman, H.(1988) J. Biol. Chem. 2 6 3 , 15241-15244 12. Faucher, M., Girones, N., Hannun, Y. A,, Bell, R. M., and Davis, R. J. (1988) J. Biol. Chem. 263,5319-5327 13. Davis, R. J., Girones, N., and Faucher, M. (1988) J. Biol. Chem. 263,5373-5379 14. Northwood, I. C., and Davis, R. J. (1988) J. Biol. Chem. 263, 7540-7543 15. Kishimoto, Y. (1983) in The Enzymes (Boyer, P.D., ed) pp. 357407, Academic Press, New York 16. Nishizuka, Y. (1984) Nature 308, 693-698 17. Vunnam, R. R., and Radin, N. S. (1979) Biochim. Biophys. Acta 5 7 3 , 73-82 18. Sunada, H., MacLeod, C., and Mendelsohn, J. (1985) Anal. Biochem. 149,438-447 19. Nimkar, S., Menaldino, D., Jamison, W. C. L., Merrill, A. H., Jr., and Liotta, D. (1988) Tetrahedron Lett. 2 9 , 3037-3040 20. Cooper, J. A., Septon, B. B., and Hunter, T. (1983) Methods Enzymol. 99, 387-402 21. Zipperl, R., Sturani, E., Toschi, L., Naldini, L., Alberguna, L., and Comoglio, P. M. (1986) Biochim. Biophys. Acta 881, 5461 22. Kris, R. M., Lax, I., Gullick, W., Waterfield, M. D., Ullrich, A., Fridkin, M., and Schlessinger, 3. (1985) Cell 4 0 , 619-625 23. Kaller, S. (1961) Biochem. 2. 334,451-456 24. Slife, C. W., Wang, E., Hunter, R., Wang, S., Burgess, C., Liotta, D. C., and Merrill, A. H., Jr. (1989) J. Bwl. Chem. 264,1037110377 25. Merrill, A. H., Jr., Wang, E., Mullins, R. E., Jamison, W. C. L., Nimkar, S., and Liotta, D. C. (1988) Anal. Biochem. 171,373~I

381

26. Schneider, E. G., and Kennedy, E. P. (1973) J. Biol. Chem. 2 4 8 , 3739-3741 27. Igarashi, Y., Kitamura, K., Toyokuni, T., Dean, B., Fenderson, B., Ogawa, T., and Hakomori, S.-I. (1990) J. Bwl. Chem. 2 6 5 , 5385-5389 28. 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. 263,9304-9309 -R. N. Kolesnick and K. A. Dressler, unpublished observations.