Lysophosphatidic Acid-induced Ca2 Mobilization Requires ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2000 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 275, No. 49, Issue of December 8, pp. 38532–38539, 2000 Printed in U.S.A.

Lysophosphatidic Acid-induced Ca2ⴙ Mobilization Requires Intracellular Sphingosine 1-Phosphate Production POTENTIAL INVOLVEMENT OF ENDOGENOUS EDG-4 RECEPTORS* Received for publication, July 25, 2000, and in revised form, August 22, 2000 Published, JBC Papers in Press, August 22, 2000, DOI 10.1074/jbc.M006631200

Kenneth W. Young‡§, Martin D. Bootman¶, Deborah R. Channing‡, Peter Lipp¶, Peter R. Maycox储, Jackie Meakin储, R. A. John Challiss‡, and Stefan R. Nahorski‡ From the ‡Department of Cell Physiology and Pharmacology, Medical Sciences Building, University of Leicester, University Road, Leicester, LE1 9HN United Kingdom, the ¶Laboratory of Molecular Signalling, The Babraham Institute, Babraham, Cambridge, CB2 4AT United Kingdom, and 储SmithKline Beecham Pharmaceuticals, Harlow, Essex, CM19 5AW United Kingdom

Lysophosphatidic acid (LPA)-mediated Ca2ⴙ mobilization in human SH-SY5Y neuroblastoma cells does not involve either inositol 1,4,5-trisphosphate (Ins(1,4,5)P3)or ryanodine-receptor pathways, but is sensitive to inhibitors of sphingosine kinase. This present study identifies Edg-4 as the receptor subtype involved and investigates the presence of a Ca2ⴙ signaling cascade based upon the lipid second messenger molecule, sphingosine 1-phosphate. Both LPA and direct G-protein activation increase [3H]sphingosine 1-phosphate levels in SH-SY5Y cells. Measurements of 45Ca2ⴙ release in premeabilized SH-SY5Y cells indicates that sphingosine 1-phosphate, sphingosine, and sphingosylphosphorylcholine, but not N-acetylsphingosine are capable of mobilizing intracellular Ca2ⴙ. Furthermore, the effect of sphingosine was attenuated by the sphingosine kinase inhibitor dimethylsphingosine, or removal of ATP. Confocal microscopy demonstrated that LPA stimulated intracellular Ca2ⴙ “puffs,” which resulted from an interaction between the sphingolipid Ca2ⴙ release pathway and Ins(1,4,5)P3 receptors. Down-regulation of Ins(1,4,5)P3 receptors uncovered a Ca2ⴙ response to LPA, which was manifest as a progressive increase in global cellular Ca2ⴙ with no discernible foci. We suggest that activation of an LPAsensitive Edg-4 receptor solely utilizes the production of intracellular sphingosine 1-phosphate to stimulate Ca2ⴙ mobilization in SH-SY5Y cells. Unlike traditional Ca2ⴙ release processes, this novel pathway does not require the progressive recruitment of elementary Ca2ⴙ events.

The lipid signaling molecule sphingosine 1-phosphate (SPP)1 is now recognized to be the extracellular ligand for a subset of the recently cloned Edg-receptor family (1). Edg-1, -3, -5, and

* This work was supported by Wellcome Trust Grant 0168895. 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. § To whom correspondence should be addressed: Tel.: 0116-252-5249; Fax: 0116-252-5045; E-mail: [email protected]. 1 The abbreviations used are: SPP, sphingosine 1-phosphate; GPCR, G protein-coupled receptor; Ins(1,4,5)P3, inositol 1,4,5-trisphosphate; Ins(1,4,5)P3R, inositol 1,4,5-trisphosphate receptor; KHB, KrebsHenseleit buffer; LPA, lysophosphatidic acid; MCH, methacholine; SPC, sphingosylphosphorylcholine; SPH, sphingosine; DMS, N,N-dimethylsphingosine; RYR, ryanodine receptor; fMLP, formylmethionylleucylphenylalanine; RT-PCR, reverse transcriptase-polymerase chain reaction; PTX, pertussis toxin; GTP␥S, guanosine 5⬘-3-O-(thio) triphosphate.

probably -8, are SPP-sensitive cell surface G protein-coupled receptors (GPCRs), which activate specific heterotrimeric Gproteins (2) to produce changes in cell shape (3), increases in Ins(1,4,5)P3 and Ca2⫹ (4), alterations in cAMP levels (5), and activation of mitogen-activated protein kinase (6). However, SPP is also capable of acting as an intracellular second messenger, where it can stimulate mitogenesis and also mobilize intracellular Ca2⫹ (7). A number of extracellular stimuli including platelet-derived growth factor, fMLP, and IgE produce increases in intracellular levels of SPP, thus supporting a second messenger role for this molecule (8 –10). Furthermore, sphingosine kinase, the enzyme responsible for SPP generation, has been recently cloned (11), and overexpression of this enzyme, which increases SPP levels, also promotes cell growth via an intracellular action (12). The ability of sphingolipids such as SPP to mobilize intracellular Ca2⫹ has been subject to sporadic interest over the last 10 years (13–15). Importantly, SPP and the related compound sphingosylphosphorylcholine (SPC) release Ca2⫹ from purified vesicles of the endoplasmic reticulum (16), indicating that sphingolipids can have effects other than at the cell surface. The ability of inhibitors of sphingosine kinase to block Ca2⫹ mobilization in response to such disparate agonists as IgE, epidermal growth factor, carbachol (via activation of M2- and M3-muscarinic receptors), as well as Fc␥-receptor aggregation, suggests that the involvement of SPP may be quite widespread (9, 15, 17, 18). However, investigating the role of sphingosine kinase in agonist-driven Ca2⫹ responses has been hindered by the additional ability of many of these agonists to stimulate Ins(1,4,5)P3-mediated Ca2⫹ mobilization. Thus, it is not clear whether SPP production is a Ca2⫹ signal in its own right, or whether it merely supports the better characterized Ins(1,4,5)P3 response. In a previous study in the human SHSY5Y neuroblastoma cell line we have demonstrated that lysophosphatidic acid (LPA) stimulates intracellular Ca2⫹ release via a mechanism that does not involve either Ins(1,4,5)P3 receptor (InsP3R) or ryanodine receptor (RYR) pathways, but is sensitive to inhibitors of sphingosine kinase (19). Therefore, LPA signaling in these cells appears to be an ideal model to investigate GPCR-mediated activation of the sphingosine kinase pathway independently of the generation and action of Ins(1,4,5)P3. In this present study we identify Edg-4 as the probable LPA-receptor involved, and confirm the presence of a sphingolipid-mediated Ca2⫹ release pathway in SH-SY5Y cells. Furthermore, we describe for the first time the subcellular Ca2⫹ responses associated with increases in intracellular SPP,

38532

This paper is available on line at http://www.jbc.org

LPA-mediated Ca2⫹ Signaling and uncover a functional interaction with InsP3R-mediated elementary Ca2⫹ events. EXPERIMENTAL PROCEDURES

Cell Culture—SH-SY5Y neuroblastoma cells were grown in minimal essential medium with Earle’s salts, 5% (v/v) fetal bovine serum, 5% (v/v) newborn bovine serum, 2 mM L-glutamine, 100 units ml⫺1 penicillin, 100 ␮g ml⫺1 streptomycin, and 2.5 ␮g ml⫺1 fungisone, in a humidified atmosphere of 5% CO2, 95% air. Imaging of [Ca2⫹]i—SH-SY5Y cells, grown on 22-mm diameter coverslips, were incubated in Krebs-Henseleit buffer (KHB, in mM: NaCl, 118; KCl, 4.7; MgSO4, 1.2; CaCl2, 1.3; KH2PO4, 1.2; NaHCO3, 4.2; Hepes, 10; pH 7.4) supplemented with 2 ␮M fura-2 AM for 60 min at 22 °C before being mounted on the stage of Nikon Diaphot inverted epifluorescene microscope and heated to 37 °C. Images at wavelengths above 510 nm were collected, after excitation at 340 and 380 nm, with an intensified charge-coupled device camera (Photonic Science). For experiments investigating the effects of DMS, cells were preincubated in KHB to which no Ca2⫹ had been added, plus DMS (where appropriate) for 10 min before agonist addition. Applications of LPA were supplemented with 10 ␮g/ml fatty acid-free bovine serum albumin (which acts as a low affinity carrier). Confocal measurements of [Ca2⫹]i were made, at room temperature, at a sample rate of 7.5 Hz using a Noran Oz confocal scanning microscope. SH-SY5Y cells were loaded with fluo-3 AM, and approximate [Ca2⫹]i was calculated using a self-ratio method (20). Immunoprecipitation of [35S]GTP␥S-labeled G␣ Subunits—SH-SY5Y cells membranes (2 mg/ml) were incubated in 10 mM Hepes, 100 mM NaCl, 10 mM MgCl2 (pH 7.2) supplemented with 10 ␮Ci of [35S]GTP␥S (12.5 mCi/ml) and various concentrations of GDP (10 ␮M for G␣i3, 1 ␮M for G␣q/11, or no GDP for experiments measuring G␣13 activation), for 2 min. Reactions were terminated by the addition of 1 ml of ice-cold buffer and centrifugation at 20,000 ⫻ g for 6 min. Nonspecific [35S]GTP␥S binding was determined by the addition of 10 ␮M GTP␥S. The pellet was resuspended in 100 ␮l of ice-cold solubilization buffer (100 mM Trizma (Tris base)-HCl, 200 mM NaCl, 0.1% SDS, 1 mM EDTA, 1.25% Igepal, pH 7.4), and precleared with rabbit serum (1:100 dilution) plus 30 ␮l of a 3% solution of protein A-Sepharose. After centrifugation at 20,000 ⫻ g, 100 ␮l of supernatant was incubated with antisera raised against specific G␣-subunits (1:100 dilution, Santa Cruz) for 2 h (G␣i3 and G␣q/11) or 16 h (G␣13) and the [35S]GTP␥S-labeled G␣ subunits immunoprecipitated by the addition of 70 ␮l of protein A-Sepharose. The samples were washed 3 times and 35S content was measured by liquid scintillation counting (21, 22). The antisera raised against G␣i3 is also reactive against G␣i1 and G␣i2. Quantitative RT-PCR Analysis of Edg-receptor Levels—Cell pellets were homogenized in Trizol reagent (Life Technologies) and total RNA was extracted. The RNA was resuspended in autoclaved, ultra-pure water and the concentration calculated by A260 nm measurement. RNA quality was assessed by electrophoresis on a 1% agarose gel. First-strand cDNA synthesis was carried out by oligo(dT) priming from 1 ␮g of each RNA (1 mM dithiothreitol, 0.5 mM each dNTP, 0.5 ␮g of oligo(dT) primer, 40 units of RNaseOUT ribonuclease inhibitor (Life Technologies), and 200 units of Superscript II reverse transcriptase (Life Technologies)). Triplicate reverse transcription (RT) reactions were performed along with an additional reaction in which the reverse transcriptase was omitted to allow for assessment of genomic DNA contamination of the RNA. The resulting cDNA products were divided into 20 aliquots for parallel Taqman PCR reactions using different primer and probe sets for quantification of multiple cDNA sequences. Quantitative RT-PCR was carried out using an ABI 7700 sequence detector (PerkinElmer Life Sciences) on the cDNA samples (2.5 mM MgCl2, 0.2 mM dATP, dCTP, dGTP, and dUTP, 0.1 ␮M each primer, 50 mM Taqman probe, 0.01 units of AmpErase uracil-N-glycosylase (PerkinElmer Life Sciences), 0.0125 units of Amplitaq Gold DNA polymerase (PerkinElmer Life Sciences); 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, 60 °C for 1 min). Additional reactions were performed on each 96-well plate using known dilutions of genomic DNA or plasmid containing the specific gene of interest as a PCR template to allow construction of a standard curve relating threshold cycle to template copy number. The initial template copy numbers for each sample were divided by the glyceraldehyde-3-phosphate dehydrogenase initial copy numbers as derived from the standard curves. The resulting values were expressed as arbitrary units. Measurement of [3H]SPP Levels—Intracellular levels of [3H]SPP were measured by a modified method described in Ref. 15. SH-SY5Y cells, grown on 6-well plates, were incubated in KHB for 15 min before

38533

replacing the buffer with 1 ml of KHB containing 0.1 mg/ml fatty acid free-bovine serum albumin and 30 nM [3H]sphingosine (20 Ci/mmol) for 1 min. This buffer was then removed and KHB ⫾ agonists added for the appropriate time. Reactions were terminated by removal of KHB and the addition of 0.75 ml of ice-cold methanol. For experiments requiring DMS pretreatment, DMS was added during the 15-min equilibration period. The [3H]sphingolipids were collected by scraping the cells and rinsing with another 0.25 ml of MeOH. Chloroform (0.5 ml) was then added to the samples and the cells mixed vigorously, before centrifugation at 20,000 ⫻ g for 5 min. A portion of the supernatant (1 ml) was then vacuum-dried in a Speedvac centrifuge. The samples were redissolved in 25 ␮l of ethanol and spotted onto Whatman 60 A silica gel TLC plates. Samples were run in a butanol/acetic acid/water (3:1:1) solvent system, and sphingosine and SPP identified with ninhydrin. Bands were scraped and measured via liquid scintillation counting. 45 Ca2⫹ Release Assay—SH-SY5Y cells were harvested in Hepes/ EDTA solution (10 mM Hepes, 0.9% NaCl, 0.02% EDTA, pH 7.4) and washed twice in cytosol-like buffer (in mM: KCl, 135; MgCl2, 2.5; EGTA, 0.05; CaCl2 0.02; Hepes, 20; pH 7.1. Free [Ca2⫹] approximately 100 nM) before being resuspended in cytosol-like buffer supplemented with 100 ␮g/ml ␤-escin, 10 mM creatine phosphate, 10 units/ml creatine kinase, 2 mM ATP, and 1.7 ␮Ci/ml 45Ca2⫹ for 10 min. Incubations were for 2 min at 37 °C, unless otherwise stated, and reactions were terminated by centrifugation. The pellet was isolated in a 1:1 mixture of Dow Corning 556:550 oil (BDH), which was then removed and 45Ca2⫹ content measured by vial liquid scintillation counting. Values were calculated as the mean 45Ca2⫹ release expressed as the percentage of the Ca2⫹ pool released by 20 ␮M ionomycin. Stock SPH and SPC, made up in methanol, were vacuum dried and resuspended in 1 mg/ml fatty acid freebovine serum albumin before being diluted 1:100 in the experiment. The activity of SPP if resuspended in this way was weak and erratic. Therefore stock SPP, made up in chloroform/methanol (1:1), was added directly to the incubations in a 1:100 dilution. The effect of chloroform/ methanol alone was always subtracted from the SPP response. RESULTS 2⫹

LPA-mediated Ca Mobilization in SH-SY5Y Cells Is Inhibited by Dimethylsphingosine but Not by Caffeine—Single SHSY5Y cells responded to an extracellular application of LPA with a transient increase in [Ca2⫹]i which was graded according to the strength of the stimulus (Fig. 1A). Pretreatment of SH-SY5Y cells with the sphingosine kinase inhibitor DMS (30 ␮M, 10 min) reduced the peak Ca2⫹ responses to 1 and 10 ␮M LPA by 88 ⫾ 5 and 46 ⫾ 19%, respectively (n ⫽ 3– 4). Although higher concentrations of LPA were not tested, it was estimated that this DMS treatment would produce a greater than 100fold increase in the EC50 for the LPA-mediated Ca2⫹ response (Fig. 1B). Previous studies in this laboratory have demonstrated that LPA-mediated Ca2⫹ responses occurred independently of the phosphoinositide-signaling pathway (19). This was further confirmed using caffeine to inhibit any potential Ins(1,4,5)P3-mediated Ca2⫹ release (23). Thus, a 1-min addition of 40 mM caffeine was without effect on Ca2⫹ increases stimulated by a submaximal concentration of LPA, but abolished the Ca2⫹ response to methacholine (MCH), a prototypic Ins(1,4,5)P3 mobilizing agonist in SH-SY5Y cells (Fig. 2). In addition to inhibiting InsP3R-mediated Ca2⫹ release, caffeine also stimulates RYRs. However, as caffeine itself did not produce a Ca2⫹ response in the majority of the cells examined, RYR-induced Ca2⫹ mobilization does not appear to be a major release pathway in SH-SY5Y cells. To investigate whether the DMS-sensitive LPA Ca2⫹ response utilized the same intracellular Ca2⫹ pool as the Ins(1,4,5)P3-mediated MCH response, SH-SY5Y cells were incubated in 0 Ca2⫹ KHB for 1 min, and then stimulated with various combinations of agonists (also in 0 Ca2⫹ KHB). An initial 2-min application of 10 ␮M LPA reduced a subsequent response to 1 ␮M MCH by 32 ⫾ 4%. However, a 3-min application of 1 ␮M MCH (the time required for the MCH response to return to basal levels) reduced the response to a subsequent application of 10 ␮M LPA by 91 ⫾ 3% (n ⫽ 3–5). Characterization of Edg-receptor Subtypes and G-protein Ac-

38534

LPA-mediated Ca2⫹ Signaling

FIG. 1. Effect of dimethylsphingosine on LPA-induced Ca2ⴙ mobilization. Fura-2-loaded SH-SY5Y cells were: A, incubated in KHB and treated with increasing concentrations of LPA; B, incubated in the absence of extracellular Ca2⫹ for 10 min ⫾ 30 ␮M DMS as appropriate, before being treated with increasing concentrations of LPA. Values represent the mean peak Ca2⫹ response from three or four experiments, and are expressed as a percentage of the response to 1 ␮M LPA.

tivation in SH-SY5Y Cells—As the LPA-mediated Ca2⫹ response in SH-SY5Y cells appeared distinct from recombinantreceptor signaling (24), the Edg-receptor profile was examined using quantitative reverse transcription-polymerase chain reaction analysis (RT-PCR). Levels of Edg mRNA for each of the known human Edg-receptors were normalized to the amount of glyceraldehyde-3-phosphate dehydrogenase mRNA found in the cells. As seen in Table I, Edg-4 mRNA was found in greatest abundance, with a small amount of Edg-7 mRNA also present. Interestingly, mRNA for the SPP receptors Edg-3 and Edg-5 was also present, although in relatively low amounts, which is consistent with our previously noted effect of extracellular SPP on SH-SY5Y cells (19). The relative levels of Edg-receptor mRNA suggest that LPAstimulated responses in SH-SY5Y cells are predominantly mediated via the Edg-4 receptor subtype. Recombinant receptor data suggests that this receptor couples to both Gi/o and Gq/11 subfamilies of G-proteins (24). The G-protein coupling of the Edg-receptors in SH-SY5Y cells was examined directly by immunoprecipitation of [35S]GTP␥S-labeled G␣ subunits in a membrane preparation. MCH (100 ␮M), acting on endogenous muscarinic M3-receptors, produced a robust 3989 ⫾ 83 cpm increase in [35S]GTP␥S bound to G␣q/11 (p ⬍ 0.001, n ⫽ 3). However, neither LPA (10 ␮M) nor the fatty acid free-bovine serum albumin control, caused increases in [35S]GTP␥S binding to G␣q/11 (n ⫽ 3) (Fig. 3A). In contrast, both LPA and MCH activated G␣i, such that basal [35S]GTP␥S levels increased from 273 ⫾ 12 cpm to 603 ⫾ 70 and 1173 ⫾ 96 cpm, respectively

FIG. 2. Effect of caffeine on LPA- and MCH-induced Ca2ⴙ mobilization. Fura-2-loaded SH-SY5Y cells were treated with 40 mM caffeine for 1 min before the addition of: A, 0.1 ␮M LPA or, B, 1 ␮M MCH. Traces are representative cells from three experiments for each condition. MCH was washed out after a 60-s application, caffeine was present throughout the experiment. TABLE I Expression of Edg-receptor mRNA in SH-SY5Y cells Edg-receptor mRNA in SH-SY5Y cells was examined using quantitative RT-PCR on the TaqMan system. Levels of Edg mRNA were normalized to the amount of glyceraldehyde-3-phosphate dehydrogenase mRNA found in the cell. Results are expressed in arbitrary units and represent the mean data from three experiments. Receptor

Relative mRNA

Edg-1 Edg-2 Edg-3 Edg-4 Edg-5 Edg-6 Edg-7

Not detected Not detected 0.006 ⫾ 0.001 0.104 ⫾ 0.010 0.003 (n ⫽ 1) Not detected 0.016 ⫾ 0.007

(p ⬍ 0.01 for LPA, and p ⬍ 0.001 for MCH, n ⫽ 4) (Fig. 3B). No activation of either G␣s or G␣12 was observed, however, LPA did produce a significant increase in [35S]GTP␥S bound to G␣13 subunits (basal, 69 ⫾ 94 cpm; LPA stimulated, 523 ⫾ 13 cpm; p ⬍ 0.001, n ⫽ 3) (Fig. 3C). Sphingosine 1-Phosphate Production and Ca2⫹ Release in SH-SY5Y Cells—Although the effects of the sphingosine kinase inhibitor, DMS, suggested the possible involvement of intracellular SPP production in LPA-mediated Ca2⫹ responses, it was important to establish the presence of a SPP/Ca2⫹ release pathway in SH-SY5Y cells. To determine whether LPA, and GPCR activation in general, could elevate intracellular SPP, SH-SY5Y cells were preloaded with [3H]sphingosine ([3H]SPH) for 1 min, and then stimulated with either LPA (10 ␮M) or AlF4⫺ (10 mM NaF, 10 ␮M AlCl3) using a method modified from that


38535

FIG. 4. Time course of [3H]SPP production. SH-SY5Y cells were prelabeled with [3H]SPH for 1 min before being incubated with LPA (10 ␮M) or AlF4⫺ (NaF/AlCl3, 10 mM/10 ␮M). Reactions were terminated by ice-cold methanol and [3H]sphingolipids extracted in chloroform/methanol. Values represent the mean ⫾ S.E. [3H]SPP from three to six experiments in which basal levels have been subtracted. Each point was conducted in duplicate. *, denotes significant difference from control values (p ⬍ 0.05).

FIG. 3. Immunoprecipitation of [35S]GTP␥S-labeled G␣-protein subunits. SH-SY5Y cell membranes were incubated with [35S]GTP␥S plus agonists for 2 min before the G-protein ␣-subunits were extracted with specific antisera and protein A-Sepharose. Values are the mean ⫾ S.E. Counts/min [35S]GTP␥S bound after subtraction of nonspecific binding for G␣q/11 (A), G␣i3 (B), and G␣13 (C) for three to four experiments with each point conducted in duplicate. FAF, fatty acid free.

from (15). Both receptor activation by LPA and direct G-protein stimulation by AlF4⫺, elevated [3H]SPP levels in SH-SY5Y cells (Fig. 4). Basal [3H]SPP after the 1-min loading period was 608 ⫾ 14 disintegrations/min (n ⫽ 4). Both the LPA and AlF4⫺ responses peaked between 30 s and 1 min before returning to prestimulated levels, such that at 1 min, basal [3H]SPP increased from 494 ⫾ 43 to 611 ⫾ 43 and 717 ⫾ 69 disintegrations/min with LPA and AlF4⫺, respectively (p ⬍ 0.05, n ⫽ 5–9). Increasing the preloading time with [3H]SPH to 5 or 15 min did not alter the observed responses (data not shown). Pretreatment of SH-SY5Y cells with pertussis toxin (PTX, 100 ng/ml

16 –20 h) did not alter the response to 10 ␮M LPA, but reduced the AlF4⫺ stimulated response by 32 ⫾ 18% (n ⫽ 2). In contrast, pretreatment of SH-SY5Y cells with DMS (30 ␮M) abolished the responses to both LPA and AlF4⫺ (n ⫽ 3). This abolition was most probably due to an inhibition of sphingosine kinase activity as indicated by an increase in the cellular [3H]SPH content of the cells at the end of the experiment. It was also important to demonstrate that intracellular SPP could mobilize Ca2⫹ stores in SH-SY5Y cells. In order to remove possible involvement of cell surface SPP receptors (see above), ␤-escin permeabilized SH-SY5Y cells were diluted to approximately 2.5 ⫻ 106 cells/ml. Under these conditions, muscarinic M3-receptor stimulation by MCH did not produce a measurable release of 45Ca2⫹ (Fig. 5A), thus making it unlikely that any of the results described below could be attributable to effects at cell surface SPP receptors. A 2-min incubation with Ins(1,4,5)P3 (20 ␮M) released 81 ⫾ 2% of the total ionomycin releasable store of 45Ca2⫹ (n ⫽ 7). SPP and SPH (both at 50 ␮M) released 18 ⫾ 5 and 29 ⫾ 3% of the 45Ca2⫹ store, respectively (n ⫽ 5–7), and the related compound SPC (50 ␮M) released 34 ⫾ 7% (n ⫽ 9) (Fig. 5A). In contrast, DMS and N-acetylsphingosine were without effect, indicating that 45Ca2⫹ release was not a general response to all sphingolipids. Furthermore, the Ca2⫹ response to SPH could be inhibited by either pretreatment with DMS or removal of ATP (using hexokinase ⫹ glucose), suggesting that SPH must first be converted to SPP by sphingosine kinase before causing 45Ca2⫹ release. Thus, a 10-min pretreatment with 30 ␮M DMS inhibited the response to 50 ␮M SPH by 47 ⫾ 8% (n ⫽ 6), while a 1-min pretreatment with hexokinase and glucose (5 units/ml and 4 mM) inhibited the SPH response by 81 ⫾ 10% (p ⬍ 0.01, n ⫽ 3) (Fig. 5B). It should be noted that although DMS pretreatment had no significant effect on Ins(1,4,5)P3-mediated 45Ca2⫹ release, the addition of hexokinase and glucose reduced the response by 43 ⫾ 8% (p ⬍ 0.001, n ⫽ 3). Concentration-response experiments suggested that sphingolipid concentrations of greater than 10 ␮M were required to produce measurable 45Ca2⫹ release responses, and time course data indicated that although 45Ca2⫹ release responses were present at the shortest time measured (10 s), maximal release occurred around 2 min (data not shown). This contrasted sharply with the effect of Ins(1,4,5)P3 (20 ␮M) which maximally released 45Ca2⫹ within 10 s. Elementary Ca2⫹ Events Associated with LPA-mediated

38536


FIG. 6. Subcellular localization of elementary Ca2ⴙ events. SHSY5Y cells, loaded with fluo-3, were imaged using a Noran Oz confocal microscope at a capture rate of 7.5 Hz. Cells were stimulated with 0.1 ␮M LPA, washed for 10 min, and then stimulated with 0.1 ␮M MCH. Data is from a representative cell, which responded to both LPA and to MCH. A, temporal profile of Ca2⫹ events arising from the same perinuclear area (highlighted in B) in response to LPA and MCH. B, pseudocolor image of a single SH-SY5Y cell illustrating the foci of the Ca2⫹ event marked with * in A.

FIG. 5. Sphingolipid-induced mobilization of 45Ca2ⴙ from intracellular stores. A, ␤-escin permeabilized SH-SY5Y cells were incubated with 45Ca2⫹ for 10 min before being treated with Ins(1,4,5)P3 (20 ␮M), MCH (1 mM), or various sphingolipids (50 ␮M) for a further 2 min. Values are the mean ⫾ S.E. 45Ca2⫹ release expressed as the percentage of the total ionomycin-releasable Ca2⫹ store for four to nine experiments, with each point conducted in duplicate. B, 45Ca2⫹-loaded SH-SY5Y cells were incubated with 30 ␮M DMS for 10 min or 5 units/ml hexokinase plus 4 mM glucose for 1 min, before being treated with Ins(1,4,5)P3 (20 ␮M) or SPH (50 ␮M). Values are the mean ⫾ S.E. data for three to six experiments.

Ca2⫹ Signaling—Previous experiments have demonstrated that LPA-induced Ca2⫹ mobilization does not involve either RYRs or InsP3Rs (19). The above data suggests that the signaling components, which allow LPA to utilize the production of intracellular SPP, are present in SH-SY5Y cells. Furthermore, as other GPCRs so far investigated also appear to stimulate Ins(1,4,5)P3 production (18), the action of LPA on SHSY5Y cells may prove useful to examine the subcellular Ca2⫹ events associated with the SPP pathway in isolation. Elementary Ca2⫹ events in response to LPA-receptor activation were compared with MCH-induced responses in fluo-3-loaded SHSY5Y cells using confocal microscopy. LPA at 0.1 ␮M, which appeared to be near the threshold for responses, stimulated intracellular Ca2⫹ puffs in 38 out of 66 cells investigated. After a 10-min washout period, a subsequent application of MCH (0.1 ␮M) produced Ca2⫹ puffs originating from the same foci as the LPA response (Fig. 6). These Ca2⫹ puffs were indistinguishable in magnitude (26.3 ⫾ 1.5 nM from 86 puffs in response to LPA, and 29.2 ⫾ 2.2 nM from 72 puffs in response to MCH), and were manifest as repetitive puffs from a steady baseline, puffs superimposed upon a raised baseline, or puffs which appeared to summate as part of a global Ca2⫹ response (Fig. 7). Spatially, these puffs were most often observed in the cell body close to the nucleus (Fig. 6B). Higher concentrations of LPA were also

capable of stimulating Ca2⫹ puffs, however, higher concentrations of MCH produced an extremely rapid global Ca2⫹ response. In view of previous data, the ability of LPA to stimulate intracellular Ca2⫹ puffs was surprising, as this suggested an involvement of InsP3Rs in the Ca2⫹ response. This question was addressed directly by pretreating with the muscarinic receptor agonist carbachol (1 mM) for 16 –24 h, in an established protocol to down-regulate the InsP3R population in SHSY5Y cells (19, 25). Such treatment abolished the response to low concentrations of MCH, and prevented LPA from stimulating Ca2⫹ puffs (Fig. 8). However, the global Ca2⫹ response to LPA remained intact. Thus, the LPA-mediated Ca2⫹ response in InsP3R down-regulated cells, which presumably results from increases in intracellular SPP, was manifest as a progressive increase in global cellular Ca2⫹ with no discernible foci. It should be noted that down-regulated cells remained capable of producing Ca2⫹ puffs, as demonstrated by increasing the concentration of MCH to 100 ␮M (Fig. 8C). DISCUSSION

In SH-SY5Y cells, SPP appears to satisfy the criteria necessary to support a role as the Ca2⫹-mobilizing second messenger utilized by LPA. Intracellular levels of SPP were transiently raised through stimulation of an LPA-sensitive GPCR, or by direct G-protein activation. Inhibiting production of SPP by blocking sphingosine kinase activity substantially reduced the ability of LPA to produce a Ca2⫹ response. Finally in permeabilized cells, under conditions where extracellular effects were minimized, SPP could itself release Ca2⫹ from intracellular stores. The observation that LPA does not activate either InsP3R- or RYR-mediated Ca2⫹ release pathways in SH-SY5Y cells (this study and Ref. 19) makes it ideal to investigate GPCR-mediated stimulation of the sphingosine kinase/Ca2⫹ signaling cascade directly. Analysis of Edg-receptor mRNA expression in SH-SY5Y cells indicates that the predominant transcript is for Edg-4. This strongly suggests that Edg-4 is the most abundant Edg family member present in these cells. Functional data obtained from


38537

FIG. 7. Comparison of LPA- and MCH-induced Ca2ⴙ puffs originating from the same foci. Fluo-3 loaded SHSY5Y cells were stimulated with 0.1 ␮M LPA (A-C), washed for 10 min, and then stimulated with 0.1 ␮M MCH (D-F). Traces are from the same foci in matched cells (i.e. A and D, B and E, C, and F), which responded to both LPA and MCH.

SH-SY5Y cells contrasts with the observation that recombinant Edg-4 receptors stimulate phosphoinositide turnover and produce a Ca2⫹ response which is partially PTX-sensitive, and is completely blocked by the phospholipase C inhibitor U73122 (25). The action of LPA in SH-SY5Y cells could also involve the Edg-7 receptor subtype, however, recombinant data again suggests this receptor activates phospholipase C (26). Stimulation of sphingosine kinase by recombinant Edg-receptors has not been investigated, but the activation of phospholipase C may reflect a higher level of receptor expression in recombinant systems, or cell type-specific receptor coupling. In SH-SY5Y cells, LPA activated G␣i and G␣13, but not G␣q/11, confirming that phospholipase C activation by G␣q/11 is not the preferred Ca2⫹ mobilizing pathway. This does not appear to be due to a lack of sensitivity to G␣-subunit activation as a robust stimulation was observed even at submaximal concentrations of MCH (620 ⫾ 50 disintegrations/min increase in response to 1 ␮M MCH, n ⫽ 3). However, which G-proteins are involved in the LPA-mediated Ca2⫹ response is as yet unclear. We have previously demonstrated the Ca2⫹ response to LPA to be PTXinsensitive (19), excluding Gi/o involvement. Investigation of the role of G13-proteins is more problematic. Activated G␣13 subunits have been shown to stimulate the small molecular weight G-protein Rho through a specific interaction with a guanine nucleotide exchange factor (27, 28), and an LPA-mediated activation of Rho via G13 has been demonstrated using specific antibodies (29). Although activation of RhoA has been linked to intracellular Ca2⫹ responses involving H2O2 generation (30), we could not detect any such relationship in this study. Thus, although pretreatment of SH-SY5Y cells with the RhoA inhibitor Clostridium difficile Toxin B produced significant changes in cell shape, no inhibition of LPA-mediated Ca2⫹ signals was observed (data not shown). Similarly, the Rhokinase inhibitor Y-27632, and H2O2 scavengers were without

effect. Therefore a link between G13 activation and LPA-mediated Ca2⫹ signaling in SH-SY5Y cells remains to be established. Both the LPA-induced stimulation of [3H]SPP production and Ca2⫹ mobilization were transient in nature. This may be representative of sphingosine kinase-mediated Ca2⫹ responses in general. For example, in differentiated monocytes, SPPinduced Ca2⫹ signals in response to Fc␥RI activation are similarly transient (17). It should be noted that the method employed to detect increases in intracellular SPP levels in this present study is not a direct measure of changes in sphingosine kinase activity. However, work by Spiegel and colleagues (8, 31), has demonstrated that growth factors such as plateletderived growth factor and nerve growth factor increase sphingosine kinase activity via an action on the Vmax of the enzyme. It seems probable, therefore, that GPCR-mediated increases in SPP involve a similar process. The actual increase in [3H]SPP levels observed in this study and others (10, 15) is small (less than 100% over basal). This probably reflects the fact that the levels measured are determined not only by the rate of [3H]SPP production but also by the rate of metabolism. This may be further exacerbated by the lack of equilibrium loading of [3H]SPH. It should be noted that the increase in intracellular Ins(1,4,5)P3 required to maximally activate Ca2⫹ mobilization is similarly small (32, 33), and so the lack of a robust increase in [3H]SPP certainly does not rule out an involvement in intracellular signaling. The ability of sphingolipids such as SPC and SPP to release Ca2⫹ from intracellular stores has been noted for over 10 years (13), although the identification of Ca2⫹-mobilizing cell surface receptors for SPP means that some caution must be taken in interpreting the various data. Nevertheless, SPP and SPC can release Ca2⫹ through a direct action on the endoplasmic reticulum (16), and microinjection of SPP into cells in which extra-

38538


FIG. 8. Effect of InsP3R down-regulation on agonist induced intracellular Ca2ⴙ responses. SH-SY5Y cells were pretreated with 1 mM CCH for 16 –24 h before being stimulated LPA or MCH as indicated. A, threshold concentrations of MCH did not stimulate Ca2⫹ puffs in down-regulated cells. B, down-regulated cells responded to LPA by producing a Ca2⫹ rise with no discernable foci, and without the additional presence of Ca2⫹ puffs. Trace is representative of 20 cells measured during three separate experiments. C, higher concentrations of MCH produced intracellular Ca2⫹ puffs in down-regulated cells.

cellular effects of SPP have been blocked by PTX pretreatment, also produces a Ca2⫹-mobilizing response (15). In permeabilized SH-SY5Y cells, the recognized Ca2⫹-mobilizing sphingolipids (SPP, SPH, and SPC) produced a modest Ca2⫹ release response, whereas other sphingolipids were inactive (DMS and N-acetylsphingosine). Furthermore, SPH required the activity of sphingosine kinase and the presence of ATP to release Ca2⫹, suggesting that conversion to SPP is an obligate step in the SPH response. A similar process has previously been described in endoplasmic reticulum microsomes, and so it may be that SPP generation occurs in a localized environment in close proximity to its Ca2⫹ release channel (16). Such a localized production may explain the modest increase in [3H]SPP production observed in this study. The above results indicate that LPA-induced Ca2⫹ mobilization in SH-SY5Y cells should prove to be a useful model in which to investigate the subcellular Ca2⫹ events associated with GPCR-mediated activation of the SPP/Ca2⫹ signaling pathway. Confocal imaging technology has uncovered the presence of elementary Ca2⫹ events, which are the basic signaling units of Ca2⫹ release (34). These result from small “packets” of Ca2⫹ being released through functionally distinct groupings of InsP3Rs or RYRs. In mammalian cells, elementary Ca2⫹ puffs arising from InsP3R activation can summate to produce global

cellular Ca2⫹ responses (35), however, SPP-mediated elementary events have not yet been investigated. Our results indicate that, in the absence of InsP3R activation, LPA produces a progressive increase in intracellular Ca2⫹, which does not appear to arise from any specific foci. This distinguishes the LPA-mediated response from those involving InsP3Rs or RYRs, as these require recruitment of elementary events. Hence, LPA appears to stimulate a novel mechanism of raising intracellular Ca2⫹, and it appears to utilize the production of intracellular SPP to do this. The ability of LPA to also recruit InsP3Rmediated Ca2⫹ puffs indistinguishable from the effect of the Ins(1,4,5)P3-mobilizing agonist MCH is perhaps not surprising as Ca2⫹ is a co-activator of InsP3Rs (see Ref. 36). Although there appears to be a functional interaction between the two Ca2⫹ release mechanisms, it should be noted that LPA does not have an absolute requirement for InsP3Rs (19). In conclusion, we have identified a novel signaling pathway associated with endogenous Edg-4 receptor activation in SHSY5Y cells. The selective stimulation of intracellular SPP production by LPA has enabled subcellular Ca2⫹ responses to SPP to be observed for the first time. An involvement of intracellular SPP production may go some way to explain the apparent complex relationship between GPCR activation, Ins(1,4,5)P3 production, and Ca2⫹ mobilization. For example, in SH-SY5Y cells the observation that concentrations of bradykinin and MCH, which produce identical increases in Ins(1,4,5)P3 stimulate different peak intracellular Ca2⫹ release responses (37) might result from differential activation of the sphingosine kinase pathway. Sphingosine kinase activity is also sensitive to increases in intracellular Ca2⫹ (38), and so stimulation of Ins(1,4,5)P3 production may lead indirectly to SPP production through cross-talk between the signaling pathways. However, the ability of AlF4⫺ (which does not increase [Ca2⫹]i in SH-SY5Y cells) to increase intracellular SPP also indicates a direct role for G-protein activation. Therefore GPCRs may be able to use to varying extents both SPP and Ins(1,4,5)P3 production to stimulate intracellular Ca2⫹ responses. Variations in the levels of activation of these two pathways may be important for the resultant intracellular Ca2⫹ signal. REFERENCES 1. Lynch, K. R., and Im, D-S. (1999) Trends Pharmacol. Sci. 20, 473– 475 2. Windh, R. T., Lee, M-J., Hla, T., An, S., Barr, A. J., and Manning, D. R. (1999) J. Biol. Chem. 274, 27351–27358 3. Lee, M-J., Van Brocklyn, J. R., Thangada, S., Liu, C. H., Hand, A. R., Menzeleev, R., Spiegel, S., and Hla, T. (1998) Science 279, 1552–1555 4. An, S., Bleu, T., and Zheng, Y. (1999) Mol. Pharmacol. 55, 787–794 5. Kon, J., Sato, K., Watanabe, T., Tomura, H., Kuwabara, A., Kimura, T., Tamama, K., Ishizuka, T., Murata, N., Kanda, T., Kobayashi, I., Ohta, H., Ui, M., and Okajima, F. (1999) J. Biol. Chem. 274, 23940 –23947 6. Rakhit, S., Conway, A-M., Tate, R., Bower, T., Pyne, N. J., and Pyne, S. (1999) Biochem. J. 338, 643– 649 7. Spiegel, S., Foster, D., and Kolesnick, R. (1996) Curr. Opin. Cell Biol. 8, 159 –167 8. Olivera, A., and Spiegel, S. (1993) Nature 365, 557–560 9. Choi, O. K., Kim, J-H., and Kinet, J-P. (1996) Nature 380, 634 – 636 10. Alemany, R., Meyer zu Heringdorf, D., Van Koppen, C. J., and Jakobs, K. H. (1999) J. Biol. Chem. 274, 3994 –3999 11. Kohama, T., Olivera, A., Edsall, L., Nagiec, M. M., Dickson, R., and Spiegel, S. (1998) J. Biol. Chem. 273, 23722–23728 12. Olivera, A., Kohama, T., Edsall, L., Nava, V., Cuvillier, O., Poulton, S., and Spiegel, S. (1999) J. Cell Biol. 147, 545–557 13. Ghosh, T. K., Bian, J., and Gill, D. L. (1990) Science 248, 1653–1666 14. Kim, S., Lakhan, V., Costa, D. J., Sharara, A. I., Fitz, J. G., Huang, L-W., Peters, K. G., and Kindman, L. A. (1995) J. Biol. Chem. 270, 5266 –5269 15. Meyer zu Heringdorf, D., Lass, H., Alemany, R., Laser, K. T., Neumann, E., Zhang, C., Schmidt, M., Rauen, U., Jakobs, K. H., and van Koppen, C. J. (1998) EMBO J. 17, 2830 –2837 16. Ghosh, T. K., Bian, J., and Gill, D. L. (1994) J. Biol. Chem. 269, 22628 –22635 17. Melendez, A., Floto, R. A., Cameron, A. J., Gillooly, D. J., Harnett, M. M., and Allen, J. M. (1998) Curr. Biol. 8, 210 –221 18. Meyer zu Heringdorf, D., Lass, H., Kuchar, I., Alemany, R., Gio, Y., Schmidt, M., and Jakobs, K. H. (1999) FEBS Lett. 461, 217–222 19. Young, K. W., Challiss, R. A. J., Nahorski, S. R., and Mackrill, J. J. (1999) Biochem. J. 343, 45–52 20. Lipp, P., Thomas, D., Berridge, M. J., and Bootman, M. D. (1997) EMBO J. 16, 7166 –7173

LPA-mediated Ca2⫹ Signaling 21. Barr, A. J., Brass, L. F., and Manning, D. R. (1997) J. Biol. Chem. 272, 2223–2229 22. Burford, N. T., Tolbert, L. M., and Sadee, W. (1998) Eur. J. Pharmacol. 342, 123–126 23. Ehlrich, B. E., Kaftan, E., Bezprozvannaya, S., and Bezprozvanny, I. (1994) Trends. Pharmacol. Sci. 15, 145–149 24. An, S., Bleu, T., Zheng, Y., and Goetzl, E. J. (1998) Mol. Pharmacol. 54, 881– 888 25. Wojcikiewicz, R. J. H., Furuichi, T., Nakade, S., Mikoshiba, K., and Nahorski, S. R. (1994) J. Biol. Chem. 269, 7963–7969 26. Bandoh, K., Aoki, J., Honsono, H., Kobayashi, S., Kobayashi, T., MurakamiMurofushi, K., Tsujimoto, M., Arai, H., and Inoue, K. (1999) J. Biol. Chem. 274, 27776 –27785 27. Kozasa, T., Jiang, X., Hart, M. J., Sternweis, P. M., Singer, W. D., Gilman, A. G., Bollag, G., and Sternweis, P. C. (1998) Science 280, 2109 –2111 28. Hart, M. J., Jiang, X., Kozasa, T., Roscoe, W., Singer, W. D., Gilman, A. G., Sternweis, P. C., and Bollag, G. (1998) Science 280, 2112–2114

38539

29. Gohla, A., Harhammer, R., and Schultz, G. (1998) J. Biol. Chem. 273, 4653– 4659 30. Lee, Z-W., Kweon, S-M., Kim, B-C., Leem, S-H., Shin, I., Kim, J-H., and Ha, K-H. (1998) J. Biol. Chem. 273, 12710 –12715 31. Rius, R. A., Edsall, L. C., and Spiegel, S. (1997) FEBS Letts. 417, 173–176 32. Willars, G. B., and Nahorski, S. R. (1995) Br. J. Pharmacol. 114, 1133–1142 33. Young, K. W., Pinnock, R. D., and Nahorski, S. R. (1998) Cell Calcium 24, 59 –70 34. Berridge, M. J. (1997) J. Physiol. 499, 291–306 35. Bootman, M., Niggli, E., Berridge, M., and Lipp, P. (1997) J. Physiol. 499, 307–314 36. Berridge, M. J. (1993) Nature 361, 315–325 37. Martin, A. K., Nahorski, S. R., and Willars, G. B. (1999) Br. J. Pharmacol. 126, 1559 –1566 38. Olivera, A., Edsall, L., Poulton, S., Kazlauskas, A., and Spiegel, S. (1999) FASEB J. 13, 1593–1600