THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc

Val. 261, No 32, Issue of Novembet ’ 15, PP 23122-23128,1992 Printed in U. S. A.

Sphingosine- 1-phosphate, a Metabolite of Sphingosine, Increases Phosphatidic Acid Levels by Phospholipase D Activation*

(Received for publication, January 24, 1992)

Naishadh N. Desai, Hong Zhang, Ana Olivera, Mark E. Mattie, and Sarah SpiegelS From the Department of Biochemistry and Molecular Biology, Georgetown University Medical Center, Washington,‘D. C. 20007-2197

Sphingosine and sphingosine- 1-phosphate, metabo- lites of membrane sphingolipids, have recently been shown to stimulate release of calcium from internal sources and to increase proliferation of quiescent Swiss 3T3 fibroblasts (Zhang, H., Desai, N. N., Olivera, A., Seki, T., Brooker, G., and Spiegel, S . (1991) J. Cell Biol. 114, 155-167). The present study demonstrates that mitogenic concentrations of sphingosine induce early increases in sphingosine- 1-phosphate levels which precede the increase in the potent mitogen, phos- phatidic acid. Sphingosine- 1-phosphate itself induces a more rapid increase in phosphatidic acid, thus sug- gesting that it may mediate the effects of sphingosine on phosphatidic acid accumulation. The concentration dependence for the formation of phosphatidic acid in- duced by sphingosine-1-phosphate correlates with its effect on DNA synthesis. Similar to sphingosine, sphin- gosine- 1-phosphate also stimulates the activity of phospholipase D, although a significant effect is ob- served at a much lower concentration. However, in contrast to previous reports with sphingosine, sphin- gosine- 1-phosphate does not inhibit the phosphatidic acid phosphohydrolase activity in cell homogenates. Thus, in addition to its effect on mobilization of cal- cium, sphingosine-l-phosphate can increase the level of phosphatidic acid, most likely via activation of phos- pholipase D. We suggest that sphingosine-1-phosphate mediates the effect of sphingosine on phosphatidic acid accumulation in Swiss 3T3 fibroblasts and may regu- late cellular proliferation by affecting multiple trans- membrane signaling pathways.

Long chain sphingoid bases are the basic building units of sphingolipids, which are important membrane lipid constitu- ents. Recently, sphingolipids have been implicated in the regulation of diverse cellular functions (reviewed in Sweeley, 1985; Hakomori, 1990; Merrill, 1991; Hannun and Bell, 1987). Sphingosine (4-trans-sphingenine), a naturally occurring sphingoid breakdown product of cellular sphingolipids (Wil- son et d. , 1988), has recently emerged as a bioregulatory molecule (Hannun and Bell, 1989; Merrill and Stevens, 1989). Sphingosine was found to inhibit protein kinase C (Hannun et al., 1986; Wilson et al., 1986; Merrill et al., 1986), a pivotal

*This work was supported by Research Grant lROl GM43880 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$To whom correspondence and reprint requests should be ad- dressed Dept. of Biochemistry and Molecular Biology, Georgetown University Medical Center, 357 Basic Science Building, 3900 Reser- voir Road NW, Washington, D. C. 20007-2197. Tel.: 202-687-1432.

regulatory enzyme in many cellular processes, and therefore has been implicated as a negative modulator of transmem- brane signaling, opposing the action of diacylglycerol which stimulates protein kinase C (Hannun and Bell, 1987, 1989; Merrill and Stevens, 1989). Further studies demonstrated that sphingosine has complex biological effects, many of which seem to be protein kinase C-independent (Faucher et al., 1988; Davis et al., 1988; Jefferson and Schulman, 1988; Winicov and Gershengorn, 1988; Krishnamurthi, et al., 1989; Igarashi et al., 1989; Arnold and Newton, 1991; Zhang et al., 1990a). We have previously reported that sphingosine stimulates pro- liferation of quiescent Swiss 3T3 fibroblasts acting in a fun- damentally different, protein kinase C-independent pathway (Zhang et al., 1990a). The mitogenic effect of sphingosine was accompanied by an increase in the levels of phosphatidic acid (PA)’ (Zhang et al., 1990b), which is a potent mitogen for a variety of cell types (Yu et al., 1988; Moolenaar et al., 1986; Zhang et al., 1990b) and may function as an intracellular second messenger (Exton, 1990; Bocckino et al., 1991). Re- cently, we found that, in addition to phosphatidic acid, mito- genic concentrations of sphingosine also induced rapid and sustained increases in sphingosine-1-phosphate levels (Zhang et al., 1991). Sphingosine-1-phosphate has long been known to be produced from sphingosine by the action of a specific kinase in the major sphingolipid degradation pathway (Stoffel et al., 1970, 1973). Sphingosine-1-phosphate is rapidly de- graded to trans-2-hexadecanal and phosphorylethanolamine by the action of a microsomal lyase (Stoffel et al., 19701, reported to be located in the endoplasmic reticulum (Van Veldhoven and Mannaerts, 1991). Recently, sphingosine-1- phosphate has been proposed as a putative messenger in the release of calcium from intracellular stores (Ghosh et al., 1990; Zhang et al., 1991). In viable 3T3 fibroblasts, we found sphin- gosine-1-phosphate to be a potent mitogen by itself and to mediate calcium release (Zhang et al., 1991). Thus, sphingo- sine-1-phosphate may be an important component of the intracellular signalling system that is involved in calcium release and in the regulation of cell growth induced by sphin- gosine. In this paper, we present evidence that sphingosine- 1-phosphate markedly increases phosphatidic acid levels in quiescent cultures of Swiss 3T3 fibroblasts primarily through activation of phospholipase D.

EXPERIMENTAL PROCEDURES

Materials-[32P]Orthophosphate (carrier-free), [T-~’P]ATP (3000 Ci/mmol), and [2-3H]glycerol (1 Ci/mmol) were purchased from Amersham. [“C]Glycerol-PA (166 mCi/mmol) was from Du Pont- New England Nuclear. Phospholipase D (from Streptomyces chrom- ofuscus, type VI, 3000 units/mg), sphingosylphosphorylcholine (free

The abbreviations used are: PA, phosphatidic acid HEPES, 4- (2-hydroxyethyl)-l-piperazineethanesulfonic acid; DAG, diacylglyc- erol; BSA, bovine serum albumin.

23122

Sphingosine-1 -phosphate and Phosphatidic Acid Levels 23123

base), D-erythro-sphingosine (from sphingomyelin), and DL-threo- dihydrosphingosine were from Sigma. The various standard phospho- lipids were from Avanti Polar Lipids. Silica Gel 60 G plates were from EM Sciences. R-59022 was from Janssen Life Sciences Products.

Synthesis of Sphingosine-1 -phosphate-An enzymatic method, de- scribed recently by Bell's group, enabled us to prepare milligram quantities of pure sphingosine-1-phosphate (Van Veldhoven et al., 1989). Briefly, sphingosine-1-phosphate was prepared by enzymatic digestion of sphingosylphosphorylcholine with phospholipase D as previously described (Van Veldhoven et al., 1989; Zhang et al., 1991). Sphingosine-1-phosphate was quantified by the malachite green phos- phate method (Hess and Derr, 1975) and stored under N, in methanol at -20 "C.

"'P-Labeled sphingosine-1-phosphate was prepared by a method which utilizes Escherichia coli diacylglycerol kinase to phosphorylate ceramide as was recently described (Bajjalieh et al., 1989; Dressler and Kolesnick, 1990). Briefly, bovine brain type I11 ceramides (0.5 mg) were solubilized in 0.4% Triton X-100 and added to a reaction mixture (50 mM Tris, pH 7.4, 50 mM NaC1, 10 mM MgCL, 2 mM EGTA, 1 mM dithiothreitol) containing E. coli diacylglycerol kinase (0.88 unit/ml) and 10 /IM [-y-32P]ATP (10 Ci/mmol). After 30 min at 30 "C, lipids were extracted wlth 1 ml of chloroform/methanol/con- centrated HCl (100200:1, v/v) and 0.2 ml of balanced salt solution containing 10 mM EDTA (Dressler and Kolesnick, 1990). The organic phase was dried under nitrogen, resuspended in chloroform/methanol ( l : l , v/v), and ceramide-1-phosphate was resolved by TLC with chloroform/methanol/acetic acid (65:15:5, v/v) (Dressler and Koles- nick, 1990). Radioactive lipids with an RF of 0.23 & 0.05, similar to that described for ceramide-1-phosphate (Dressler and Kolesnick, 1990), were eluted from the silica with chloroform/methanol ( l : l , v/ v) and treated with 6 N HC1/1-butanol (l:l, v/v) for 1 h at 100 "C to produce ["'Plsphingosine-1-phosphate which was further character- ized as previously described (Zhang et al., 1991).

Cell Culture and Labeling with "P&3wiss 3T3 fibroblasts were from the American Type Culture Collection (CCL 92). Cells were routinely cultured as previously described (Spiegel, 1989; Zhang et al., 1990a, 1991). Confluent and quiescent cultures of 3T3 cells were washed with phosphate-free Dulbecco's modified Eagle's medium and incubated with this buffer containing 32Pi (40 pCi/ml) for 24 h. In some experiments, confluent cultures were labeled with [2-3H]glycerol (4 pCi/ml) for the last 3 days of culture, and "Pi (40 pCi/ml) was added for the final 24 h. The cells were treated with sphingosine, sphingosine-1-phosphate, or vehicle alone for various times, the me- dium was rapidly removed, and the cells were scraped from the dish in 1 ml of 0.1 M HC1 (Zhang et al., 1990a, 1990b).

Extraction and Analysis of Lipids-Lipids were extracted with chloroform/methanol/concentrated HCl (100:2001, v/v), and the phases were separated as described (Zhang et al., 1990b). The lipids in the lower chloroform phase were analyzed by two dimensional TLC on Silica Gel 60 G. The plates were developed in chloroform/ methanol/ammonium hydroxide (13:7:1, v/v) in the first dimension and chloroform/methanol/water/acetic acid (3030:2:5, v/v) in the second dimension (Zhang et al., 1991). The lipid standards were visualized by spraying with molybdenum blue. Phospholipids were located by autoradiography, and the radioactivity was quantified by liquid scintillation counting of the corresponding silica gel areas. In some experiments, sphingosine-1-phosphate was analyzed by TLC using butanol/water/acetic acid (3:1:1, v/v) as described (Zhang et al., 1991). Phosphatidic acid was also analyzed by TLC using the organic phase of the mixture of isooctane/ethyl acetate/acetic acid/ water (50110:20:100, v/v). In this system (Pai et al., 1988; Zhang et al., 1990b), phosphatidic acid (RF = 0.1) was well separated from other phospholipids (R, = 0). 3H-Labeled mono-, di-, and triacylglyc- erols were analyzed by TLC using hexane/diethyl ether/methanol/ acetic acid (90:20:3:2, v/v).

Assay of Phospholipase D Activity in 3T3 Fibroblasts-Phospholi- pase D activity was determined by measurement of the transfer of :"P-labeled phosphatidyl moieties to 1-propanol. The production of phosphatidylpropanol appears to be exclusively mediated by phos- pholipase D (Pai et al., 1988; Ben Av and Liscovitch, 1989). Briefly, :"P,-metabolically labeled cells were treated with sphingosine, sphin- gosine-1-phosphate, or vehicle in the presence of 134 mM 1-propanol. After various times, the incubations were terminated, cellular lipids were extracted, and [32P]phosphatidic acid and ["P]phosphatidylpro- panol were analyzed by TLC using the isooctane solvent system described above.

Assay of Phosphatidic Acid Phosphohydrolase Actiuity-Fibroblast homogenates were prepared as described (Mullmann et al., 1991).

Isolated cells (106/ml) were suspended at 4 "C in buffer A (50 mM HEPES, pH 7.5, 1.25 mM EDTA, 3.25 mM MgC1,) and sonicated. Phosphatidic acid phosphohydrolase activity was assayed by meas- urement of the production of ['4C]diacylglycerol from ["Clphospha- tidic acid as described (Mullmann et al., 1991). Liposomes were produced by suspending dried egg phosphatidylcholine and phospha- tidic acid in buffer A by sonication followed by vortexing. Assays (in a total volume of 200 pl) were initiated by the addition of 50 p1 of homogenate (200 pg of protein) to a prewarmed (37 "C) liposome suspension containing 150 mM NaCl, 0.2 mM phosphatidic acid, 0.2 mM phosphatidylcholine, and ["Clphosphatidic acid (100,000 dpm) and incubated in the absence or presence of various concentrations of sphingosine-1-phosphate at 37 "C for 30 min. The reactions were stopped by the addition of 0.5 ml of chloroform/methanol (88:12, v/ v), and the lipids in the chloroform phase were analyzed as described above.

RESULTS

The Formation of Sphingosine-1-phosphate in Response to Sphingosine Precedes the Accumulation of Phosphatidic Acid-Sphingosine significantly stimulated incorporation of 32P into phosphatidic acid and sphingosine-1-phosphate in cells metabolically prelabeled with 32Pi to isotopic equilibrium (Zhang et al., 1990b, 1991 and also Fig. 1). Consistent with our previous study (Zhang et al., 1990b), sphingosine caused a dose-dependent increase in [32P]phosphatidic acid levels (Fig. lA). The accumulation of sphingosine-1-phosphate in

0 1 0 2 0 3 0 4 0 5 0 6 0

Sphingosine (pM)

0 1 1 0 1 0 0 I1

Incubation Time (min) 0 0

FIG. 1. Dose-response and time course for sphingosine-in- duced accumulation of 32P-labeled phosphatidic acid and 32P- labeled sphingosine- 1-phosphate. Quiescent cultures of 3T3 cells were washed with phosphate-free Dulbecco's modified Eagle's me- dium and incubated with this buffer containing 32Pi (40 pCi/ml) for 24 h. The cells were treated with increasing concentrations of sphin- gosine-BSA complex for 1 h (A) or with sphingosine-BSA complex (10 p ~ ) for the indicated time periods ( B ) . Lipids were extracted, separated by TLC, and phospholipids were located by autoradiogra- phy as described under "Experimental Procedures." The silica gel areas containing the labeled phosphatidic acid (0) and sphingosine- 1-phosphate (0) were scraped and counted by liquid scintillation spectrometry. The results are expressed as -fold increase relative to untreated controls. The incorporations of 32P into phosphatidic acid and sphingosine-1-phosphate in untreated cells were 600 2 80 and 150 & 60 cpm, respectively, determined from an aliquot of the phos- pholipid extract containing 150,000 cpm.

23124 Sphingosine-1 -phosphate and Phosphatidic Acid Levels

response to sphingosine was also concentration-dependent, with a maximal effect at a significantly lower concentration (20 pM). Concentrations of sphingosine greater than 50 pM were not studied since they were highly cytotoxic (Zhang et al., 1990a, 1990b). The time courses of 32P-labeled sphingo- sine-1-phosphate and phosphatidic acid accumulation in re- sponse to an optimal mitogenic concentration of sphingosine are shown in Fig. 1B. A significant increase in [32P]sphingo- sine-1-phosphate was detected within 5 min after exposure of the cells to sphingosine. This preceded the increase in for- mation of [32P]phosphatidic acid which could be detected only after 20 min. Thus, the formation of sphingosine-1-phosphate was rapid, reaching nearly maximal levels within 60 min, when 6-fold stimulation over the basal value was observed. The maximum accumulation of phosphatidic acid also oc- curred at approximately the same time. The rapid increase in sphingosine-1-phosphate levels appears to precede the in- crease in accumulation of phosphatidic acid, suggesting that sphingosine-1-phosphate could be mediating the effect of sphingosine on phosphatidic acid accumulation (Zhang et al., 1990b).

Two different approaches were used to further investigate the possibility that sphingosine-1-phosphate mediated the sphingosine-induced accumulation of phosphatidic acid. Cel- lular levels of ATP were depleted by pretreatment of the cells for 15 min with antimycin A (0.5 pM) and oligomycin (1.2 PM) which block the mitochondrial respiratory chain and ATP synthetase, respectively (Mohr and Fewtrell, 1990). These inhibitors completely blocked the phosphorylation of sphin- gosine (20 p ~ ) and also prevented the sphingosine-dependent increase in ["'Plphosphatidic acid levels without affecting the labeling of other phospholipids. However, this approach will result in inhibition of any pathway that utilizes ATP as a phosphorylating agent. To further subtantiate these effects of ATP depletion, we used an inhibitor of sphingosine kinase, DL-threo-dihydrosphingosine (10 p ~ ) , which has been shown t o inhibit the production of sphingosine-1-phosphate in iso- lated platelets (Buehrer and Bell, 1992). The production of sphingosine-1-phosphate from sphingosine was decreased 63 .t 2% by threo-dihydrosphingosine, and the sphingosine-in- duced production of phosphatidic acid was completely elimi- nated.

The Effect of Sphingosine-1 -phosphate on Phosphatidic Acid Accumulation-Addition of pure sphingosine-1-phosphate to confluent 3T3 fibroblasts resulted in a rapid increase in ["PI phosphatidic acid accumulation. An increase of 2-fold in ["PI phosphatidic acid levels was detected within 5 min after exposure of the cells to sphingosine-1-phosphate (Fig. 2). After 30-60 min, the levels decrease and nearly return to the basal value by 2-4 h (data not shown). Since the cells were metabolically labeled to isotopic equilibrium, it is most prob- able that these increases reflect changes in phosphatidic acid mass rather than enhanced radioactive labeling. To further substantiate this point, the cells were double-labeled with both [3H]glycerol and 32Pi to isotopic equilibrium. Mitogenic concentrations of sphingosine-1-phosphate induced identical stimulations of incorporation of [3H]glycerol and 32P into phosphatidic acid, indicating that the changes reflect in- creases in phosphatidic acid mass. The dose-response of phos- phatidic acid accumulation induced by sphingosine-l-phos- phate is shown in Fig. 3. A significant effect was detected at a concentration of sphingosine-1-phosphate as low as 0.1 j" and the maximum effect of more than %fold stimulation over the control was found a t a concentration of 2 pM. Interest- ingly, when compared to the effects of sphingosine on the increase in phosphatidic acid levels, a much lower concentra-

0 2 0 4 0 6 0

Time (min)

FIG. 2. Time course of phosphatidic acid accumulation in response to sphingosine-1-phosphate. Confluent and quiescent cultures of Swiss 3T3 cells were prelabeled with 32Pi for 24 h and stimulated with sphingosine-1-phosphate (2 PM) for the indicated time periods. The lipids were extracted, and the "P-labeled phospha- tidic acid in an aliquot of the phospholipid extract containing 150,000 cpm was determined by TLC as described in the legend to Fig. 1. The results are the means of two independent experiments done in dupli- cate and are expressed as percent of maximal response. Inset, auto- radiogram from a representative experiment demonstrating the sep- aration of phosphatidic acid (upper band, RF = 0.1) from other phospholipids (louler band, RF = 0) at 0, 5, 10, and 30 min after treatment with sphingosine-1-phosphate.

Sphingosine-I-Phosphate (pM)

FIG. 3. Dose-response of sphingosine-1-phosphate-induced formation of 32P-labeled phosphatidic acid. Quiescent cultures of Swiss 3T3 cells, labeled with "Pi, were treated with increasing concentrations of sphingosine-1-phosphate for 1 h, and 32P-labeled phosphatidic acid was quantitated as described in the legend to Fig. 1.

tion of sphingosine-1-phosphate was sufficient to induce a significant increase in phosphatidic acid levels. Furthermore, the concentration dependence for the formation of phospha- tidic acid induced by sphingosine-1-phosphate correlated with its effect on DNA synthesis (compare Fig. 3 to Fig. 4 in Zhang et al., 1991). Since sphingosine-1-phosphate is more potent than sphingosine in stimulating production of phosphatidic acid, it seems most likely that sphingosine-1-phosphate is the active agent and does not require degradation to sphingosine. To examine this point further, we have studied the cellular uptake and internal reconversion of 32P-labeled sphingosine- 1-phosphate to sphingosine. 32P-Labeled sphingosine-l-phos- phate was taken up rapidly by Swiss 3T3 fibroblasts, and, after 10 min of labeling with 5 p~ sphingosine-1-phosphate

Sphingosine-1 -phosphate and Phosphatidic Acid Levels 23125

(8.4 cpm/pmol), there was an incorporation of about 50 pmol/ lo5 cells. After incubations as long as 1 h, when the media and cell extracts were analyzed by TLC, [32P]sphingosine-l- phosphate was the only labeled lipid detected and there was no significant increase in "Pi. These results suggest that the active species was sphingosine-1-phosphate itself and was not due to rapid reconversion to sphingosine by the action of a membrane phosphatase.

Effect of Sphingosine-1-phosphate on Phospholipase D Ac- tivity-We next investigated the source of phosphatidic acid elicited by sphingosine-1-phosphate. The hydrolysis of phos- pholipids by phospholipase D is a major pathway by which sphingosine increases the levels of phosphatidic acid in 3T3 fibroblasts (Lavie and Liscovitch, 1990; Zhang et al., 1990b; Kiss and Anderson, 1990). Phospholipase D has transphos- phatidylase activity with a large variety of alcohol acceptors, producing the corresponding phosphatidyl alkyl ester (Pai et al., 1988). We utilized the production of phosphatidylpro- panol, when 1-propanol was the phosphatidyl group acceptor, as a measure of phospholipase D activity. Sphingosine-1- phosphate treatment markedly increased phospholipase D activity (Figs. 4 and 5). Addition of sphingosine-1-phosphate in the presence of 1-propanol induced the synthesis of ["PI phosphatidylpropanol, in addition to [32P]PA (Fig. 4). The formation of both was rapid and occurred in parallel, reaching a maximum within 5-10 min (Fig. 5A). Sphingosine-1-phos- phate-induced [32P]PA accumulation decreased in the pres- ence of 1-propanol, consistent with a redistribution of phos- phatidyl moieties to the synthesis of phosphatidylpropanol. The dose-response for the increase in phosphatidylpropanol is shown in Fig. 5B and was comparable to the dose-response for the increase in PA levels induced by sphingosine-1-phos- phate found in the absence of 1-propanol (compare Fig. 3 and Fig. 5B). A significant accumulation of [32P]PA and [32P] phosphatidylpropanol was detected a t 0.1 p~ sphingosine-1- phosphate and reached a maximum a t 5 PM. In contrast, higher concentrations of sphingosine (10 p ~ ) only resulted in a very small increase (Fig. 4), and, even a t a concentration of 50 p ~ , the formation of both [32P]PA and [32P]phosphatidyl- propanol did not reach a plateau level (data not shown).

p

PPr - - c i

Propanol - + + + + + + + SPP (pM) - - 0.1 0.5 2 5 10 - SPH (vM) - - - - - - - 10

FIG. 4. The effect of sphingosine-1-phosphate on phospho- lipase D activity determined by 32P-labeled phosphatidylpro- panol formation in quiescent 3T3 fibroblasts. Cells were prela- heled with '*Pi for 24 h and stimulated for 40 min with increasing concentrations of sphingosine-1-phosphate (SPP) or sphingosine (SPH) in the absence or presence of 1-propanol (134 mM). The labeled phospholipids were separated by TLC, and autoradiograms were developed as described under "Experimental Procedures." The upper and lower arrows indicate the positions of phosphatidic acid (PA) and phosphatidylpropanol (PPr) , respectively.

Effect of Sphingosine-1-phosphate on Phosphatidic Acid Phosphohydrolase Activity-Recently, it has been shown that in addition to stimulation of phospholipase D, sphingosine also inhibits the activity of phosphatidic acid phosphohydro- lase (Lavie et al., 1990; Mullmann et al., 1991). Phosphatidic acid phosphohydrolase catalyzes the hydrolysis of phospha- tidic acid to produce diacylglycerol. However, in cells which were double-labeled with both [3H]glycerol and 32Pi to isotopic equilibrium, sphingosine-1-phosphate induced significant in- creases in the levels of phosphatidic acid without correspond- ing decreases in the levels of 'H-labeled diacylglycerol. There was even an apparent trend toward increases in the levels of [3H]diacylglycerol (Fig. 6). This increase in [3H]diacylglycerol lagged behind the elevation in phosphatidic acid levels. Inter- estingly, the levels of [3H]triacylglycerols significantly de- creased after treatment with sphingosine-l-phosphate, while the level of [3H]monoacylglycerol remained unchanged (data not shown).

Previously, it had been demonstrated that sphingosine in- hibited phosphatidic acid phosphohydrolase activity in several other cell types measured in vitro (Lavie et al., 1990; Mull- mann et al., 1991). It was thus of interest to examine the effects of sphingosine-1-phosphate on phosphatidic acid phos- phohydrolase activity in 3T3 cell homogenates. When mixed liposomes containing [14C]phosphatidic acid were incubated with 3T3 cell homogenates, ['4C]diacylglycerol was formed in a time- and dose-dependent manner. Boiled homogenates had no measurable activity. Sphingosine-1-phosphate increased diacylglycerol formation with an ICso value of 140 pM (Fig. 7) which is a much higher concentration than that required for all of the reported biological effects observed in intact cells (Zhang et al., 1991).

Effect of Sphingosine-1 -phosphate on Diacylglycerol Ki- nase-While increases in phosphatidic acid levels are a direct consequence of the hydrolysis of phospholipids by phospho- lipase D and, in some cases, combined with inhibition of phosphatidic acid phosphohydrolase activity (Mullmann et al., 1991), phosphatidic acid can also be formed by alternative pathways, such as diacylglycerol phosphorylation by diacyl- glycerol kinase or by acylation of glycerol 3-phosphate. Re- cently, it has been demonstrated that sphingosine activates the 80-kDa isoenzyme form of diacylglycerol kinase in vitro leading to an increase in phosphatidic acid levels (Sakane et al., 1989; Kanoh et al., 1990). We utilized the compound R- 59022, a specific inhibitor of this diacylglycerol kinase isoen- zyme form (ICso, 10 pM), in an attempt to assess the role of this kinase in the formation of phosphatidic acid in response to sphingosine-1-phosphate. Surprisingly, R-59022, at rela- tively low concentrations (1-10 p ~ ) , potentiated, rather than inhibited, the accumulation of phosphatidic acid induced by sphingosine-1-phosphate (Fig. 8). Furthermore, R-59022 also potentiated long-term cellular proliferation induced by sphin- gosine-1-phosphate or sphingosine (Fig. 8). The concentration range of R-59022 that synergized with sphingosine-1-phos- phate and sphingosine was narrow and resembled the dose- response observed previously for bombesin-induced DNA syn- thesis in these cells (Morris et al., 1988). Strong inhibition of cellular proliferation was observed at concentrations of R- 59022 greater than 10 p ~ , which is a concentration within the range used by other workers to inhibit diacylglycerol kinase activity (Sakane et al., 1989; Kanoh et al., 1990). At higher concentrations (220 p ~ ) , R-59022 also inhibited the accumulation of phosphatidic acid induced by sphingosine-1- phosphate.

23126 Sphingosine-1 -phosphate and Phosphatidic Acid Levels

2.5 - ‘ A B

2.00 4-

- n .

w “7

9 g 3- .- .o - e .; 2 - e 0

P

c

” - C ’

-1.00 c I . . ”

0 7 0 2 0 4 0 6 0 0 2 4 6 8 1 0

lneubnlfon Time (mln) Spbinposine-1-pLospLalc (pM)

FIG. 5. Time course and dose-response of sphingosine-1-phosphate- stimulated formation of [32P]phosphatidylpropanol and [32P]phosphatidic acid in the presence of 1-propanol. 3T3 cells were prelabeled with 32Pi for 24 h and were treated with sphingosine-l- phosphate (5 PM) for the indicated periods ( A ) or with increasing concentrations of sphingosine-1-phosphate for 30 min ( B ) in the presence of 1-propanol (134 mM), and the formation of [32P]PA (0) and [32P]phosphatidylpropanol (0) were determined. The data are expressed as -fold increase over unstimulated cells and are the means of duplicate determinations. In untreated control cells, the amounts of PA and phosphatidylpropanol were 0.53 k 0.07 and 0.08 k 0.01% of the total phospholipids, respectively.

1500 I h E, 1300

Y U

P I / I

1000 -I lL

900 4 0 1 0 2 0 3 0 4 0

T i m (min) FIG. 6. Effect of sphingosine-1-phosphate on diacylglycerol

levels. Quiescent cultures of Swiss 3T3 cells were prelabeled with [2-gH]glycerol (4 pCi/ml) for the last 3 days of culture. Cells were treated with sphingosine-1-phosphate (5 p ~ ) for the indicated time periods, and the 3H-labeled diacylglycerols were measured as de- scribed under “Experimental Procedures.”

DISCUSSION

The turnover of phosphoglycerolipids, often related to the phosphoinositide cycle and more recently to phosphatidylcho- line hydrolysis, plays a pivotal role in the initiation of a variety of cellular responses (Billah and Anthes, 1990; Dennis et al., 1991). Agonist-induced activation of phospholipase C, phospholipase D, and phospholipase A2 produces the crucial intracellular second messengers, diacylglycerol, phosphatidic acid, and arachidonic acid, respectively. A cycle of sphingo- lipids, comparable to the phosphoglycerolipid cycles, with its own messenger transducing products (sphingosine or ceram- ide) has been proposed recently (Hannun and Bell, 1989; Merrill and Stevens, 1989; Merrill, 1991). Sphingosine, a breakdown product of cellular sphingolipids, has appropriate properties that make it a suitable candidate to function as an intracellular second messenger (reviewed in Merrill, 1991): 1) sphingosine elicits different responses in a wide variety of cell

Sphingisine-1-phosphate (pM)

FIG. 7. Effect of sphingosine- 1-phosphate on phosphatidic acid phosphohydrolase activity in cell homogenates. Phospha- tidic acid phosphohydrolase activity was determined in the absence or presence of increasing concentrations of sphingosine-1-phosphate as described under “Experimental Procedures.” Values are the means of duplicate determinations of the formation of “C-labeled DAG from [I4C]PA and are from a representative experiment that was repeated three times.

types; 2) the structural properties of this molecule allow for its incorporation and rapid mobility in membranes and makes it accessible to different effector systems; 3) the level of free sphingosine in cells is very low and can be regulated by some physiological stimuli (Wilson et al., 1988); and 4) the turnover of sphingosine by phosphorylation to form sphingosine-l- phosphate followed by cleavage to a long chain aldehyde and phosphorylethanolamine is extremely rapid (Stoffel et al., 1970).

Although in several systems some of the biological effects of sphingosine appear to be a consequence of the inhibition of protein kinase C, it seems clear that it also has other biochemical targets. Previous studies from our laboratory have shown that sphingosine may play an important role in cell growth regulation acting in a protein kinase C-independ- ent pathway (Zhang et al., 1990a). We have shown that the mitogenic effect of sphingosine is accompanied by an increase in phosphatidic acid (Zhang et al., 1990b), a potent mitogen for Swiss 3T3 cells (Yu et al., 1988; Moolenaar et al., 1986; Zhang et al., 1990b) which has recently been implicated as an intracellular second messenger (Bocckino et al., 1991). This

Sphingosine-1 -phosphate and Phosphatidic Acid Levels 23127

4 A

0 2 4 6 8 10 12 14 16 18 20 22

R 59022 (pM)

5 B

0 2 4 6 8 10 12 14 16 18 20 22

R 59022 (pM)

FIG. 8. Effect of R-59022 on 32P-labeled phosphatidic acid accumulation and DNA synthesis induced by sphingosine-l- phosphate. A, quiescent cultures of Swiss 3T3 cells labeled with 32Pi were treated with sphingosine-1-phosphate (5 PM) in the presence of increasing concentrations of R-59022 for 1 h, and 32P-labeled phos- phatidic acid was quantitated as described under “Experimental Procedures.” B, duplicate cultures were incubated in Dulbecco’s mod- ified Eagle’s medium/Waymouth (1:l) supplemented with BSA (20 pg/ml), transferrin (5 pglml), and insulin (2 pg/ml) and treated with 20 p~ sphingosine (0) or 5 p~ sphingosine-1-phosphate (0) in the presence of inceasing concentrations of R-59022. [3H]Thymidine incorporation was measured as described previously (Spiegel, 1989). Each value is the mean & S.D. of triplicate determinations. The results are expressed as -fold incorporation measured in the absence of R-59022. Similar results were obtained in at least five additional experiments.

finding raises the possibility that sphingosine may play an important role as a positive regulator of cell growth acting through this novel pathway.

Recently, we found that sphingosine also induced a rapid increase in the levels of sphingosine-1-phosphate and that the dose-response for the accumulation of sphingosine-l-phos- phate correlated closely with the dose-response of sphingo- sine-induced stimulation of DNA synthesis (Zhang et al., 1991). Furthermore, similar to sphingosine, sphingosine-l- phosphate caused rapid release of calcium from internal stores and also stimulated DNA synthesis and cell division in quies- cent cultures of Swiss 3T3 fibroblasts (Zhang et al., 1991).

In the present study, we observed that the formation of sphingosine-1-phosphate in response to mitogenic concentra- tions of sphingosine was extremely rapid and preceded the formation of phosphatidic acid, suggesting that sphingosine- 1-phosphate could mediate the effect of sphingosine on phos- phatidic acid accumulation (Zhang et al., 1990b). Indeed, sphingosine-1-phosphate at low concentrations induced a rapid formation of phosphatidic acid which may be one intra- cellular agent that mediates the mitogenic effect of sphingo- sine and sphingosine-1-phosphate. Inhibition of sphingosine- 1-phosphate formation also prevented the accumulation of phosphatidic acid.

The mechanism by which sphingosine-1-phosphate stimu- lates formation of phosphatidic acid was investigated. An increase in cellular phosphatidic acid levels can occur through several known pathways, such as stimulation of phospholipase D, inhibition of the degradation of existing phosphatidic acid pools by phosphatidic acid phosphohydrolase, acylation of glycerol 3-phosphate, and increased phosphorylation of DAG catalyzed by DAG kinase. Recently, sphingosine has been shown to stimulate phospholipase D in several cell types (Lavie and Liscovitch, 1990). It has been observed that sphin- gosine decreases phosphatidylcholine levels in NG108-15 neuroblastoma-glioma hybrid cells (Lavie and Liscovitch, 1990) and in Swiss 3T3 fibroblasts (Zhang et al., 1990b) and also stimulates the hydrolysis of phosphatidylethanolamine in NIH 3T3 fibroblasts (Kiss and Anderson, 1990). Further- more, recent evidence demonstrates that sphingosine has the additional effect of inhibiting phosphatidic acid phosphohy- drolase activity in cell lysates (Lavie et al., 1990; Mullmann et al., 1991). Thus, it has been suggested that sphingosine has a dual action on the regulation of phosphatidic acid levels (Lavie et al., 1990).

We observed that sphingosine-1-phosphate induced a rapid increase in the activity of phospholipase D, as measured by the formation of phosphatidylpropanol. The time course was similar for the effect of sphingosine-1-phosphate on phospha- tidic acid formation. This indicates that a rapid activation of phospholipase D is responsible for the sphingosine-l-phos- phate-induced phosphatidic acid accumulation. It should be emphasized that sphingosine-1-phosphate stimulated phos- pholipase D activity at much lower concentrations than sphin- gosine and reached the maximum response more rapidly. Therefore, it appears that sphingosine-1-phosphate is more potent than sphingosine in stimulating phospholipase D ac- tivity, suggesting that it could mediate the effect of sphingo- sine on phospholipase D activation. In contrast to earlier reports that the increase in phosphatidic acid induced by sphingosine in other cell types was accompanied by a decrease in the levels of [3H]diacylglycerol (Lavie et al., 1990; Mull- mann et al., 1991), we found that sphingosine-1-phosphate slightly increased the levels of 3H-labeled diacylglycerol in Swiss 3T3 cells, which lagged behind the elevation in phos- phatidic acid levels. This result indicates that sphingosine-l- phosphate, in contrast to sphingosine, did not inhibit phos- phatidic acid phosphohydrolase. Indeed, in contrast to pre- vious reports that sphingosine inhibits phosphatidic acid phosphohydrolase activity in uitro, optimal mitogenic concen- trations of sphingosine-1-phosphate had no measurable ef- fects on the activity of this enzyme. We only found an effect in Swiss 3T3 lysates at very high concentrations of sphingo- sine-1-phosphate (>lo0 PM), where it stimulated the activity rather than inhibiting it. These concentrations are much higher than required for the other effects of sphingosine-l- phosphate in cells, such as increased DNA synthesis, calcium release, or increases in phosphatidic acid levels (Zhang et al., 1991). Taken together, these results indicate that phospha- tidic acid phosphohydrolase probably does not contribute to the increase in the levels of phosphatidic acid in Swiss 3T3 fibroblasts induced by sphingosine-1-phosphate.

Another possibility that we examined is that sphingosine- 1-phosphate stimulates DAG kinase-catalyzed phosphoryla- tion of DAG. Previous studies demonstrated that sphingosine activates the 80-kDa isoenzyme form of DAG kinase while inhibiting the 150-kDa form (Sakane et al., 1989). Therefore, the effect of sphingosine on DAG kinase could be different depending on the type of DAG kinase isozyme present in a particular cell type. Using high concentrations of R-59022, an

23128 Sphingosine-1 -phosphate and Phosphatidic Acid Levels

inhibitor of the 80-kDa isozyme (Sakane et al., 1989), we found that it inhibits the accumulation of phosphatidic acid and mitogenesis induced by sphingosine-1-phosphate. How- ever, at low concentrations, R-59022 stimulated both re- sponses. Although the nature of R-59022 effects are unclear, nonspecific actions of this compound have been reported (Nunn and Watson, 1987; Mahadevappa, 1988). Furthermore, it is still not clear which isoenzyme form of DAG kinase is present in Swiss 3T3 fibroblasts (Kanoh et al., 1990; Mac- Donald et al., 1988). Results using R-59022 should also be interpreted with caution since it may have an effect on phos- phatidic acid-dependent pathways.

Recently, phospholipase D has emerged as a key element of stimulus-response coupling in diverse cellular systems (re- viewed in Billah and Anthes, 1990; Dennis et al., 1991; Shukla and Halenda, 1991). In this study, we have shown that sphin- gosine-1-phosphate can increase the level of phosphatidic acid, most likely via activation of phospholipase D. This effect may be unrelated to its ability to mobilize calcium from internal stores (Ghosh et al., 1990; Zhang et al., 1991). Thus, sphingosine-1-phosphate could mediate the effect of sphin- gosine on phosphatidic acid accumulation in Swiss 3T3 fibro- blasts (Zhang et al., 1990b) and may regulate cellular prolif- eration by affecting multiple transmembrane signaling path- ways. It still remains to be established by further experiments whether mitogenic stimuli regulate intracellular levels of sphingosine-1-phosphate. If this is the case, sphingosine-l- phosphate could function not only as a novel intracellular second messenger inducing the release of calcium from inter- nal stores but also to amplify the cascade of events leading to mitogenic stimulation via its effect on phosphatidic acid lev- els.

REFERENCES

Arnold, R. S., and Newton, A. C. (1991) Biochemistry 30 , 7747-7754 Bajjalieh, S. M., Martin, T. F., and Floor, E. (1989) J. Biol. Chem. 264,14354-

Ben-Av, P., and Liscovitch, M. (1989) FEBS Lett. 259,64-66 Billah, M. M., and Anthes, J. C. (1990) Biochem. J . 269,281-291 Bocckino, S. B., Wilson, P. B., and Exton, J. H. (1991) Proc. Natl. Acad. Sci.

Buehrer, B. M., and Bell, R. M. (1992) J. Biol. Chem. 267,3154-3159 Davis, R. J., Girones, N., and Faucher, M. (1988) J. Biol. Chem. 2 6 3 , 5373-

Dennis, E. A,, Rhee, S. G., Billah, M. M., and Hannun, Y. A. (1991) FASEB J.

Dressler, K. A,, and Kolesnick, R. N. (1990) J. Biol. Chem. 265,14917-14921

14360

U. S. A. 88,6210-6213

5379

6,2068-2077

Exton, J. H. (1990) J. Biol. Chem. 265 , 1-4 Faucher. M.. Girones. N.. Hannun. Y. A,. Bell. R. M.. and Davis. R. J. (1988)

J. Bio~.Ckm:2-63; 5319-5327 ' , I

Ghosh, T. K., Bian, J., and Gill, D. L. (1990) Science 248,1653-1656 Hakomori. S. (1990) J. Biol. Chem. 265,18713-18716 Hannun, Y . A:, andBell, R. M. (1987) Science 236.670-674 Hannun, Y. A., and Bell, R. M. (1989) Science 243,500-506 Hannun, Y. A,, Loomis, C. R., Merrill, A. H., and Bell, R. M. (1986) J. Biol.

Hess, H., and Derr, J. E. (1975) Anal. Biochem. 63,607-613 Igarashi, Y., Hakomori, S., Toyokuni, T., Dean, B., Fujita, S., Sugimoto, M.,

3145 Ogawa, T., Ghendy, K. E., and Racker, E. (1989) Biochemistry 28. 3138-

Chem. 261,12604-12609

Jefferson, A. B., and Schulman, H. (1988) J. Biol. Chem. 263,15241-15244 Kanoh, H., Yamada, K., and Sakane, F. (1990) Trends Biochem. Sci. 15 , 47-

sn KG;, Z., and Anderson, W. B. (1990) J. Biol. Chern. 2 6 5 , 7345-7350 Krishnamurthi, S., Patel, Y., and Kakkar, V. V. (1989) Biochim. Biophys. Acta

Lavie, Y., and Liscovitch, M. (1990) J. Biol. Chem. 265,3868-3872 Lavie, Y., Piterman O., and Liscovitch, M. (1990) FEBS Lett. 2 7 7 , 7-10 MacDonald, M. L., Mack, K. F., Williams, B. W., King, W. C., and Glomset, J.

Mahadevappa, V. C. (1988) Biochem. Biophys. Res. Commun. 153, 1097-1104 Merrill, A. H. (1991) J. Bioenerg. Biornembr. 23,83-104 Merrill, A. H., and Stevens, V. L. (1989) Biochim. Biophys. Acta 1010 , 131-

1010,258-264

A. (1988) J. Biol. Chern. 263,1584-1592

1% Merril, A. H., Sereni, A. M., Stevens, V. L., Hannun, Y. A., Bell, R. M., and

Mohr, F. C'., and Fewtrell, C. (1990) Am. J. Physiol. 268 , C217-C226 Moolenaar, W. H., Kruijer, W., Tilly, B. C., Verlaan, I., Bierman, A. J., and

Morris, C., Rice, P., and Rozengurt, E. (1988) Biochem. Biophys. Res. Commun.

Mullmann. T. J.. Sieeel. M. I.. Eean. R. W.. and Billah. M. M. (1991) J. Biol.

"I

Kinkade J. M. (1986) J. Biol. Chem. 261,12610-12615

deLaat, S. W. (1986) Nature 323,171-173

156,561-568

them. 266 , 2o13-ioi6 ' '

. .

Nunn, D. L., and Watson, S. P. (1987) Biochem. J. 243,809-813 Pai, J. K., Siegel, M. I., Egan, R. W., and Billah, M. M. (1988) J. Biol. Chem.

Sakane, F. Yamada K., and Kanoh, H. (1989) FEBS Lett. 255,409-413 Shukla, S. D., and Halenda, S. P. (1991) Life Sci. 48,851-866 Spiegel, S. (1989) J. Biol. Chem. 264,16512-16517 Stoffel, W., Assmann, G., and Binczek, E. (1970) Hoppe-Seyler's 2. Physiol.

Chem. 361,635-642 Stoffel, W., Heimann, G., and Hellenbroich, B. (1973) Hoppe-Seyler's 2. Phys-

iol. Chem. 354,562-566 Sweeley, C. C. (1985) in Biochemisty of Lipids and yembranes (Vance, M. ,E.,

and Vance, J. E., eds) pp. 361-403, The Benjamln/Cummlngs Publishing

Van Veldhoven, P.P., and Mannaerts, G. P. (1991) J. Biol. Chem. 266,12502- Co., Menlo Park CA

Van Veldhoven, P. P., Foglesong, R. J., and Bell, R. M. (1989) J. Lipid Res. 12507

Wilson, E., Olcott, M. C., Bell, R. M., Merrill, A. H., and Lambeth, J. D. (1986) 30,611-616

Wilson, E., Wang, E., Mullins, R. E., Liotta, D. C., Merill, A. H., and Lambeth,

Winicov, I. and Gershengorn, M. C. (1988) J. Biol. Chem. 263,12179-12182 Yu, C.-L., b a i , M.-H., and Stacey, D. W. (1988) Cell 52,63-71 Zhang, H., Buckles, N. E., Gibson, K., and Spiegel, S. (1990a) J. Biol. Chem.

2 6 3 , 12472-12477

J. Biol. Chem. 2 6 1 , 12616-12623

J. D. (1988) J. Biol. Chem. 263,9304-9309

265,76-81

265,21309-21316

(1991) J. CellBiol. 114,155-167

Zhang, H., Desai, N. N., Murphey, J. M., and Spiegel, S. (199Ob) J. Biol. Chem.

Zhang, H., Desai, N. N., Olivera, A,, Seki, T., Brooker, G., and Spiegel, S.

~~