Calcium Release-activated Calcium Current (I CRAC) Is a Direct Target for Sphingosine*

Whole cell patch-clamp recordings were made to study the regulation of the store-operated calcium release-activated calcium current(I CRAC) by metabolites involved in the sphingomyelin pathway in RBL-2H3 cells. Sphingosine, a regulator of cell growth, inhibits I CRAC completely within 200 s and independently from conversion to either sphingosine 1-phosphate or ceramide. Structural analogs of sphingosine, includingN,N-dimethylsphingosine,dl-threo-dihydrosphingosine, andN-acetylsphingosine (C2-ceramide) also blockI CRAC. This effect is always accompanied by an elevation of whole cell membrane capacitance. These sphingolipids appear, therefore, to accumulate in the plasma membrane and directly block I CRAC channels. Sphingosylphosphorylcholine also increases capacitance but does not inhibit I CRAC, demonstrating structural specificity and that the elevation of capacitance is necessary but not sufficient for block. Nerve growth factor, which is known to break down sphingomyelin, inhibits I CRAC, and this inhibition can be antagonized by reducing sphingosine production withl-cycloserine, suggesting thatI CRAC is a physiologically relevant and direct target of sphingosine. We propose that sphingosine directly blocksI CRAC, suggesting that the sphingomyelin pathway is involved in I CRAC regulation.

Agonists that stimulate phospholipase C (PLC) 1 and elevate InsP 3 levels activate Ca 2ϩ entry that is important for refilling depleted stores (for review see Refs.[1][2][3].In mast cells and T-lymphocytes, this store-operated calcium (SOC) current is highly selective for Ca 2ϩ ions and has been termed calcium release-activated-calcium current (I CRAC ) (4,5).I CRAC has also been measured in the mast cell line RBL-2H3, in hepatocytes, in thyrocytes, 3T3 fibroblasts, and HL-60 cells (6).Both protein kinase C (PKC) and a small guanosine triphosphate-binding protein regulate I CRAC (7,8), although the exact mechanism for activation of I CRAC remains unknown.
One of the difficulties in pin-pointing a definite mechanism for regulation of I CRAC may relate to the fact that PLC activation functions in parallel with other intracellular signaling processes.The sphingomyelin (SM) pathway, for example, "cross-talks" with the PLC pathway (9).Activation of the SM pathway produces important second messengers such as sphingosine, sphingosine 1-phosphate (S1P), and ceramide (10).Generally, sphingosine and S1P mediate cell growth and proliferation, whereas ceramide elevations cause programmed cell death (apoptosis) or cell cycle arrest (9).Although several direct targets have been characterized for ceramide (10), the picture is less clear for sphingosine.PKC has been shown to be a target for sphingosine (11), but the action of sphingosine on other proteins and intracellular processes may be mediated by conversion to S1P and/or ceramide (10,12,13).
Intracellular Ca 2ϩ elevations may be an important connection for cross-talk between the PLC and SM pathways.Indeed, it has been demonstrated that elevations in [Ca 2ϩ ] i can modulate cell growth (14 -16), and sphingosine has been shown to alter Ca 2ϩ homeostasis in various preparations (17)(18)(19)(20)(21).We have therefore investigated the role of the SM pathway in regulating intracellular calcium concentration, focusing on the major Ca 2ϩ influx pathway provided by the SOC current I CRAC in the mast cell line RBL-2H3.We surveyed the sphingomyelin pathway for possible regulators of I CRAC and found that sphingosine and structurally related compounds (but not S1P nor ceramides) are inhibitors.None of the agents tested activate I CRAC .A requirement for inhibition of I CRAC is incorporation of the lipid molecule into the plasma membrane.Sphingolipid incorporation was measured as an elevation of the whole cell capacitance, which represents a new approach for monitoring lipid or possibly drug accumulation in the plasma membrane.
To provide a link between cellular sphingosine and inhibition of I CRAC , we tested several growth factors, which are known to activate sphingomyelinase and/or to elevate sphingosine levels (9,(22)(23)(24)(25)(26)(27).We observed that nerve growth factor (NGF 2.5 S subunit; 1 g/ml) inhibits I CRAC .NGF-induced inhibition of I CRAC was reduced by treatment with L-cycloserine (LCS; 2 mM), which lowers SM levels (26) and hinders sphingosine production by NGF.Implications for sphingosine-dependent block of I CRAC are discussed.

EXPERIMENTAL PROCEDURES
Cell Culture-Rat basophilic leukemia cells (RBL-2H3) were plated on glass coverslips at low density and incubated at 37 °C with 10% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 45 mM sodium bicarbonate, 5 mM glucose, 0.12 mg/ml streptomycin, and 0.60 mg/ml penicillin.For the recordings, coverslips were transferred into the recording chamber containing the desired external solution.RBL-2H3 cells stably transfected with m1 muscarinic receptors (RBL-m1 (28,29)) were a kind gift from Dr. O. H. Choi (National Institutes of Health, Bethesda).RBL-m1 cells were cultured identically to RBL-2H3 cells, except 400 g/ml G418 was added to the RBL-m1 growth medium to select for transfected cells.The muscarinic receptor agonist carbamylcholine (CCh; 100 M) activated calcium release (as measured by fura-2) or I CRAC (in patch-clamp experiments) in every RBL-m1 cell tested.
Whole cell break-in without activation of I CRAC was achieved by excluding InsP 3 , clamping intracellular Ca 2ϩ to ϳ80 nM with 2 to 1 ratio of Cs⅐EGTA/Ca⅐EGTA, and including nucleotides (4 mM Mg⅐ATP and 300 M Na 2 ⅐GTP) to maintain intracellular integrity and to prevent depletion of internal Ca 2ϩ stores.This internal solution was used when testing for activation of I CRAC by various sphingomyelin pathway intermediates.
High resolution current recordings were acquired by a computerbased patch-clamp amplifier system (EPC-9).Currents due to cellular membrane capacitance were recorded and canceled before each voltage ramp using the automatic capacitance compensation of the EPC-9.In this manner, whole cell capacitance was measured throughout the experiment.Series resistance was between 2 and 10 megohms, and inhibition of I CRAC did not correlate with series resistance.Currents were filtered at 3.3 kHz and digitized at 100 s.Currents elicited by voltage ramps in Fig. 1 were filtered at 1 kHz off-line.Ramps were given every 2 s (Ϫ100 to ϩ100 mV in 50 ms) from a holding potential of 0 mV, and the Ca 2ϩ current was analyzed at the ramp segment corresponding to Ϫ80 mV.This protocol minimized Ca 2ϩ entry because of the delay between ramps and because of the relatively small driving force for Ca 2ϩ at 0 mV.Minimizing Ca 2ϩ entry is important to limit Ca 2ϩdependent inactivation of I CRAC by local accumulation of Ca 2ϩ (30,31).All currents were leak-subtracted by averaging the first two to six ramps after breaking in and then subtracting this from all subsequent traces.For display purposes, averaged time courses of I CRAC were graphed at slightly lower temporal resolution (i.e.0.25-0.125Hz).Extracellular solution changes were made by local pressure application from a wide-tipped micropipette placed within 20 m of the cell.A liquid-junction potential of Ϫ10 mV was corrected.
We have eliminated all experiments from our analysis in which nonspecific membrane breakdown by sphingosine occurred within the typical observation window of 5-10 min.During patch-clamp recordings, breakdown was characterized by sudden increases in linear ramp currents, and in fluorescence measurements, increases in [Ca 2ϩ ] as well as dye leakage occurred.Such breakdown was seen in approximately 50% of the cells, both in patch-clamp and fluorescence experiments.
Intracellular Calcium Concentration-Intracellular [Ca 2ϩ ] was monitored with a photomultiplier-based system and calculated from the fluorescence ratio (360/390).For single cell Ca 2ϩ measurements, coverslips were incubated in a Ringer's solution containing the following (in mM): 140 NaCl, 2.8 KCl, 1 MgCl 2 , 2 CaCl 2 , 11 glucose, and 10 Na⅐HEPES (pH 7.2), to which 5 M fura-2/acetoxymethyl ester had been added.Following incubation at 37°C for 15-20 min the coverslips were washed five times in the Ringer's solution.Experiments were done at room temperature.
Ionic Currents Other Than I CRAC -Inward rectifier potassium currents were measured in RBL-2H3 cells.The external saline solution contained the following (in mM): 140 NaCl, 20 KCl, 2 CaCl 2 , 2 MgCl 2 , 11 glucose, and 10 HEPES⅐NaOH (pH 7.2).The internal saline solution contained the following (in mM): 145 K⅐Glu, 8 NaCl, 1 MgCl 2 , 2 Mg⅐ATP, 0.3 Na 2 ⅐GTP, 10 K⅐EGTA, and 10 HEPES⅐KOH (pH 7.2).Ramps were given every 2 s (Ϫ100 to ϩ100 mV in 50 ms), and cells were held at 0 mV between ramps.Currents were not leak-subtracted.Under these conditions, the K ϩ current (measured at Ϫ80 mV) did not run down, but in fact the amplitude steadily increased in control experiments over a period of approximately 200 s.
Delayed rectifier potassium currents were measured in rat skeletal myoballs isolated from newborn rats (2-5 days old) as described (32).The external solution contained the following (in mM): 140 NaCl, 2. Current records obtained from all experiments in which voltageactivated currents were studied were leak-subtracted by a standard P/n procedure, where 4 leak pulses of Ϫ0.1ϫ the actual test pulse amplitude were averaged, scaled, and subtracted from the current evoked by the test pulse.
Sphingosine and palmitoyl coenzyme A were dissolved in dimethyl sulfoxide (Me 2 SO) at 10 mM.DL-threo-Dihydrosphingosine, sphingosylphosphorylcholine, sphingosine 1-phosphate, and sphingomyelin were dissolved in methanol at 10 mM.C 2 -ceramide, C 8 -ceramide, and ceramide 1-phosphate were dissolved in either Me 2 SO or methanol at 10 mM.In most experiments, the vehicles Me 2 SO or methanol were diluted in external saline by a ratio of 1:1000 and neither vehicles significantly altered membrane currents when applied alone (up to 0.5%).Sphingomyelinase, LCS, and phospholipase A 2 were dissolved directly in external saline.Phospholipase D was dissolved in special buffer containing 10 mM Tris⅐HCl (pH 8.0), 0.05% bovine serum albumin, and 0.1% Triton X-100 to a concentration of 500 units/ml (pH 8.0) and was further diluted in external solution by a factor of 1:1000.Growth factors, including NGF, were dissolved in external saline containing 0.1% essentially fatty acid-free BSA (Sigma).The external solution used for experiments with sphingomyelinase and palmitoyl coenzyme A contained 2 mM Ca 2ϩ (instead of the usual 10 mM) to increase the solubility of these molecules.

RESULTS
Block of I CRAC by Exogenous Sphingosine-In RBL cells, InsP 3 indirectly activates I CRAC by depleting internal Ca 2ϩ stores.Fig. 1A shows the time course of I CRAC following whole cell break-in with 20 M InsP 3 in the pipette solution (left panel).Voltage ramps (Ϫ100 to ϩ100 mV in 50 ms; right panel) were applied every 2 s, and I CRAC was measured from the region on the ramp trace corresponding to Ϫ80 mV.Averaged values from 22 cells are shown (left panel).The Ca 2ϩ current activates and peaks in approximately 50 s.Then I CRAC is sustained, decaying only 9.1 Ϯ 2.5% (mean Ϯ S.E.; n ϭ 21) 300 s after break-in.The experiments were done without ATP in the pipette solution to minimize endogenous kinase-dependent inactivation of I CRAC (8).
Externally applied sphingosine (5 M) blocks InsP 3 -evoked I CRAC (Fig. 1B).Sphingosine was applied 50 s after whole cell break-in, and the onset of block follows a delay of 18 Ϯ 3 s (n ϭ 6).The delay probably relates to the time taken for sphingosine to accumulate in the plasma membrane (see "Elevation of Membrane Capacitance," below).Complete inhibition of the current occurs within 200 s after addition of sphingosine, and the effect is not reversible within measurement time (5-10 min).Sphingosine also inhibits activation of I CRAC when applied at least 1 min before break-in with InsP 3 (n ϭ 2, not shown).Furthermore, sphingosine was also able to suppress I CRAC when activated by thapsigargin, which represents an alternative pool-emptying protocol (Fig. 1C).Sphingosine when co-applied with thapsigargin essentially abolished thapsigargin-induced I CRAC (n ϭ 3), whereas control cells that were only stimulated with thapsigargin consistently activated regularsized I CRAC after a delay of about 60 s (n ϭ 3).In these experiments, cells were perfused with the standard internal solution (buffered to resting [Ca 2ϩ ] i and without InsP 3 ), and either thapsigargin alone (1 M) or in combination with sphingosine (5 M) were applied externally.
This block of I CRAC by sphingosine seems to be PKC-inde-pendent for the following reasons: 1) the PKC inhibitor staurosporine (5 M; co-applied externally) has no effect on sphingosine block of I CRAC (n ϭ 3, not shown), and 2) exogenous sphingosine inhibits PKC (11).Because PKC inhibitors sustain I CRAC ( 8), an explanation dependent on PKC inhibition by sphingosine would be opposite to the blocking effect illustrated in Fig. 1B.Another possible explanation for inhibition might be that sphingosine maximally activates I CRAC , in a manner similar to ionomycin (4).Maximal activation could, in principle, cause intracellular Ca 2ϩ accumulation during long applications and subsequent inhibition of I CRAC channels by calciumdependent inactivation (30,31).This possibility was ruled out by the following experiment: whole cell break-in without activation of I CRAC was achieved by excluding InsP 3 from the pipette solution, including ATP, and buffering internal Ca 2ϩ to ϳ80 nM.With these conditions, applying sphingosine 50 s after break-in never activated I CRAC (n ϭ 6, not shown), indicating that sphingosine block is not mediated by "over activation."It could still be possible, however, that sphingosine interacts with the [Ca 2ϩ ] i inactivation site of CRAC channels thereby mimicking Ca 2ϩ -induced inactivation of I CRAC .Interestingly, sphingosine fails to inhibit I CRAC when applied from the inside (Fig. 1D).Whole cell break-in was conducted with or without sphingosine in the pipette solution (omitting In control cells, extracellular application of thapsigargin (1 M) consistently activated I CRAC after a short delay (n ϭ 3, filled symbols).In both types of experiments, cells were perfused with the standard intracellular solution containing 10 mM EGTA and no added ATP.D, when applied intracellularly, sphingosine (10 M) does not inhibit I CRAC .Ionomycin (2 M) was used to activate I CRAC with and without sphingosine in the pipette solution.The graph in the left panel shows that no significant difference (p ϭ 0.2; t test) was observed between controls (n ϭ 3; time constant ϭ 1731 Ϯ 459 s) and cells treated with internal sphingosine (n ϭ 4; time constant ϭ 1062 Ϯ 79 s).Currents were normalized to whole cell capacitance.
InsP 3 and including 10 mM EGTA), and a short application of ionomycin (2 M) was used to activate I CRAC .The time course of I CRAC in control cells (n ϭ 3) versus the time course in cells dialyzed internally with 10 M sphingosine (n ϭ 4) is similar (Fig. 1D).
Block of I CRAC by external sphingosine is dose-dependent (Fig. 2).Concentrations between 10 and 30 M block ϳ100% of the Ca 2ϩ current within 100 s.Lower concentrations inhibit more slowly.For example, 1 M sphingosine blocks approximately 70% of I CRAC in 400 s.Average I CRAC time courses during application of 300 nM (n ϭ 5), 1 M (n ϭ 5), and 10 M (n ϭ 4) sphingosine are plotted in Fig. 2A.The average I CRAC time course during treatment with 300 nM sphingosine is indistinguishable from the control (cf.Fig. 1A).The large variability observed at 1 M is due to variations of endogenous decay of I CRAC .Concentrations of sphingosine between 100 nM and 30 M were tested, and a dose-response curve was constructed by measuring the percentage of inhibition at 50 s after application (Fig. 2B).The corresponding apparent half-maximal inhibitory concentration (IC 50 ) for sphingosine is at 6 M.
Because full block occurs at lower concentrations, but over a much longer time, we also constructed a dose-response curve using the time course of decay as the dependent variable (Fig. 2C).To measure the blocking time, single exponentials were fit to the I CRAC time course during inhibition by sphingosine.The time constants of inhibition ranged from 24 Ϯ 4 s with 30 M sphingosine (n ϭ 3) to 1,529 Ϯ 400 s with 100 nM sphingosine (n ϭ 4).The apparent IC 50 using this method is at 1 M.
To determine whether sphingosine blocks I CRAC activated by agonist, we used muscarinic type 1 receptors stably transfected in RBL-2H3 cells (RBL-m1).Muscarinic receptor activation with carbamylcholine (CCh) elevates InsP 3 levels which release Ca 2ϩ from internal stores (Fig. 3B (28,29)).Following store depletion, I CRAC activates and provides a pathway for Ca 2ϩ entry.In whole cell recordings, CCh (100 M) activated I CRAC in 9 of 9 cells tested (Fig. 3A).Co-application of sphingosine (5 M) and CCh decreased the amplitude of I CRAC by greater than 50%, and the current decayed to base line with a time constant of 105 Ϯ 53 s (n ϭ 5).This rate of inhibition is similar to the rate measured for sphingosine-mediated block of InsP 3 -evoked I CRAC at the same concentration (76 Ϯ 20 s; n ϭ 6).Therefore, sphingosine can block I CRAC activated by a physiologically relevant stimulus.
These results were confirmed by measuring intracellular calcium concentration ([Ca 2ϩ ] i ) changes in single RBL-m1 cells loaded with fura-2/acetoxymethyl ester.Fig. 3B shows averaged muscarinic receptor-mediated Ca 2ϩ signals in the presence of 10 mM extracellular Ca 2ϩ (control), 0 Ca 2ϩ , or 10 mM Ca 2ϩ ϩ 5 M sphingosine.In the control cells, CCh elevates Ca 2ϩ levels to 1.3 Ϯ 0.1 M (n ϭ 17) following a short delay.The [Ca 2ϩ ] i remained elevated for longer than 100 s before declining slightly.I CRAC contributes to this sustained elevation of Ca 2ϩ .Following removal of CCh the [Ca 2ϩ ] i returns toward base line (not shown).The Ca 2ϩ dependence of the sustained phase is demonstrated by similar experiments in nominally Ca 2ϩ -free external saline where the [Ca 2ϩ ] i returns quickly  where IC 50 represents the half-maximal inhibitory concentration, and n is the Hill coefficient.The IC 50 is at 6 M and n ϭ 1.2.C, the dose-response was also measured as the time constant of decay caused by sphingosine block as derived from single exponential fits.These data were then fit by a by a similar Boltzmann equation, giving an IC 50 value of 1.1 M and n ϭ 1.6.
toward base-line values following the peak elevation (1.3 Ϯ 0.1 M; n ϭ 5).When sphingosine was co-applied with CCh, the sustained phase was dramatically reduced (n ϭ 5).Clearly, the time course of sphingosine inhibition is similar to the time course of I CRAC block shown in Fig. 3A.Assessment of the mechanism by which sphingosine blocks I CRAC depends on whether this sphingolipid alters calcium homeostasis; however, the reported effects of sphingosine on intracellular calcium homeostasis are highly variable.In some preparations, sphingosine elevates intracellular InsP 3 (34,35) and [Ca 2ϩ ] i (17)(18)(19)(20).In other cell types, exogenous sphingosine has no effect on resting Ca 2ϩ levels (21,36).We observed variability within the population of RBL cells tested.In 15 of 21 cells tested, sphingosine (5 M) by itself elevated Ca 2ϩ levels following a delay of 81 Ϯ 11 s (data not shown).In nominally Ca 2ϩ -free external solution, sphingosine elevated Ca 2ϩ levels in 8 of 16 cells after a delay of 49 Ϯ 10 s.However, including 0.1% fatty acid-free bovine serum albumin (BSA) to the perfusion solution along with sphingosine prevented [Ca 2ϩ ] i changes in Ca 2ϩ -containing (n ϭ 7) and Ca 2ϩ -free (n ϭ 8) solutions.BSA had no effect on sphingosine-mediated block of I CRAC (n ϭ 2).When co-applied with sphingosine, BSA prevents cell permeabilization and subsequent dye leakage (37), which would appear as an increase in Ca 2ϩ levels.The approximately 50% occurrence of a sphingosine effect on [Ca 2ϩ ] i correlated well with the percentage of patch-clamped cells that showed evidence of membrane breakdown (i.e.seal loss or large nonspecific ramp currents) in the presence of sphingosine (see "Experimental Procedures").This supports a concept where exogenous sphingosine by itself has no specific effect on Ca 2ϩ levels in RBL cells.
Sphingosine Analogs and Related Agents-We investigated the mechanism of block by testing sphingosine analogs and intermediates from the sphingomyelin pathway.I CRAC was evoked with InsP 3 in the patch pipette solution, and various sphingolipids or related agents were applied externally 50 s after break-in.Fig. 4A shows that dimethylsphingosine (DMS; 5 M) inhibits I CRAC with a time constant of 42 Ϯ 4 s (n ϭ 6) and by 96 Ϯ 3% after 250 s.The sphingosine kinase inhibitor, DL-threo-dihydrosphingosine (38), also inhibits I CRAC at 5 M, although with slightly less potency (see Table I for inhibition profile).D-erythro-Dihydrosphingosine (sphinganine; 5 M) is another sphingosine analog that inhibits I CRAC with similar potency to sphingosine (n ϭ 3, not shown).However, sphingosylphosphorylcholine (lysosphingomyelin), a sphingolipid with less structural similarity to sphingosine, has no effect on the calcium current (see also Table I).
Spingosine 1-phosphate (S1P) is an important intermediate because it is formed from sphingosine by endogenous sphingosine kinase.Interestingly, S1P at 10 M had no effect on I CRAC (Fig. 4A; n ϭ 5), and it failed to do so even at a higher concentration of 50 M (n ϭ 3; not shown).Moreover, the sphingosine kinase inhibitor DL-threo-dihydrosphingosine blocks I CRAC in a manner similar to that of sphingosine (Fig. 4A).Since DMS inhibits both I CRAC and sphingosine kinase (39), and since S1P fails to block I CRAC , while inhibitors of S1P formation readily block, we conclude that sphingosine inhibition does not seem to be mediated by conversion to S1P.
Calcium measurements were conducted to supplement the results obtained from direct patch-clamp recordings of I CRAC inhibition by sphingosine analogs.Calcium signals were activated by CCh (100 M) in RBL-m1 cells.When co-applied with CCh, none of these sphingolipids alter peak amplitudes of the intracellular calcium signals (ϳ1.5 M) due to release from internal stores (Fig. 4B).However, DMS and DL-threo-dihy- drosphingosine dramatically inhibit the plateau phase, in the same order of potency observed with patch-clamp.On the other hand, neither sphingosylphosphorylcholine (n ϭ 9; Fig. 4B) nor sphingosine 1-phosphate (n ϭ 5; not shown) significantly affect the sustained phase of the Ca 2ϩ signal, consistent with the observation that both compounds do not alter I CRAC (Fig. 4A).
We also tested short chain ceramide analogs, which possess a sphingosine backbone thus sharing structural features with sphingosine.One of these analogs, C 2 -ceramide (N-acetylsphingosine), is similar to sphingosine, yet its biological effects can differ from those of sphingosine.For example, it has been reported that exogenous C 2 -ceramide does not inhibit PKC (40).Nevertheless, we observed that 10 M C 2 -ceramide inhibits I CRAC slowly, with a time constant of 408 Ϯ 114 s (n ϭ 6; Fig. 5A).A 250-s treatment of C 2 -ceramide inhibits 47 Ϯ 10% of I CRAC (Table I).Since there is no evidence that exogenous C 2 -ceramide converts to sphingosine in cells (41), it seems that C 2 -ceramide blocks I CRAC independently from sphingosine formation.
A longer chain ceramide analog (C 8 -ceramide) has no effect on the calcium current even at a higher concentration of 50 M (n ϭ 6; Fig. 5B).Interestingly, however, the phosphorylated form of C 8 -ceramide (C1P) inhibits I CRAC at 50 M (n ϭ 6; Fig. 5B).The inhibition by C1P is slow and has a similar profile to that of C 2 -ceramide (Table I).Accumulation of charged lipids at the plasma membrane and the subsequent alteration of the voltage field across the bilayer can modulate ion channels (42,43).If charge-screening mediated block of I CRAC , we would expect opposite effects of positively and negatively charged sphingolipids.This was not the case.Both positively charged C 2 -ceramide and sphingosine as well as negatively charged C1P inhibit I CRAC .Moreover, negatively charged palmitoyl coenzyme A accumulates in the plasma membrane, as measured by an increase in capacitance, but it does not inhibit nor activate (n ϭ 3) I CRAC (Table I).
To complete the profile of inhibition, we tested agents related to sphingosine or the SM pathway for their ability to block I CRAC (Table I).The sphingosine precursors sphingomyelin and palmitoyl coenzyme A have no effect on I CRAC , suggesting that breakdown of sphingosine or conversion to sphingomyelin cannot account for block.In support of this conclusion is our observation that sphingomyelinase (SMase; 0.1 units/ml) does not significantly alter I CRAC (n ϭ 3).
Since exogenous SMase preferentially elevates ceramide levels (44,45), the fact that SMase does not block I CRAC also supports the assertion that sphingosine blocks I CRAC independently from ceramide formation.To confirm further that sphingosine acts separately from ceramide, we used fumonisin B 1 , an inhibitor of ceramide synthase (46).Fumonisin B 1 has no effect on sphingosine block of I CRAC when applied from the patch pipette (i.e.intracellularly) or together with sphingosine from the outside (n ϭ 3; not shown).
Sphingosine could potentially block I CRAC through pathways involving arachidonic acid (AA) and/or phosphatidic acid (PA).Therefore, we tested phospholipase A 2 (0.5 units/ml) and phospholipase D (0.5 units/ml) which elevate AA and phosphatidic acid, respectively.These enzymes neither block (Table I) nor activate I CRAC , indicating that sphingosine does not mediate its effect via accumulation of AA nor PA, although in other preparations, elevation of AA and PA levels follow sphingosine or S1P treatment (9).
Elevation of Membrane Capacitance-Sphingosine and all of the analogs tested in this study that inhibit I CRAC also elevate membrane capacitance by approximately 10% (in 150 -250 s).We interpret this effect as accumulation of the lipid in the plasma membrane.Membrane capacitance does not change in control cells, even after long, 10-min experiments.In control cells, the average membrane capacitance at the beginning of the experiment (just following whole cell break-in) is 16.9 Ϯ 0.4 pF (n ϭ 21) and does not change significantly after 250 s (17.0 Ϯ 0.4 pF).On the contrary, sphingosine (5 M) elevates membrane capacitance by 11.0 Ϯ 2.0% after 250 s, which is significantly different from controls (0.8 Ϯ 0.6%; p Ͻ Ͻ 0.001; t test).Fig. 6A shows an example from a representative experiment.Higher concentrations of sphingosine lead to larger increases in capacitance, and the lowest concentration of 100 nM has no significant effect (Fig. 6B).Sphingosine also elevates capacitance when co-applied with CCh to RBL-m1 cells (by 7.2 Ϯ 0.9% after 250 s; n ϭ 3).
Nonspecific effects are unlikely to account for the increase in capacitance, since two of the lipids that do not block I CRAC (sphingosylphosphorylcholine and palmitoyl Co-enzyme A) also elevate membrane capacitance (Table I).Fig. 6C shows a representative experiment in which sphingosylphosphorylcholine increases membrane capacitance without altering I CRAC .Furthermore, sphingosine did not lead to blebbing of RBL cells, which would indicate cytotoxic effects.Taken together, these results suggest that sphingosine accumulates in the plasma membrane, a finding consistent with the notion that exogenous sphingosine directly blocks I CRAC .
Interestingly, S1P fails to inhibit I CRAC or elevate membrane capacitance (Table I).This could be due to the fact that S1P is highly insoluble in the absence of BSA, preventing incorporation into the plasma membrane.Thus we applied S1P in the presence of 0.1% fatty acid-free BSA.This, however, neither increased the whole cell capacitance nor enabled S1P to block I CRAC (n ϭ 3).A lack of accumulation in the plasma membrane could indicate rapid metabolism of S1P and/or possible transport to different intracellular compartments.
Sphingosine and Other Ionic Currents-We further investigated possible effects of sphingosine on K ϩ , Ca 2ϩ , or Na ϩ currents in various preparations (Table II).Sphingosine (5 M) inhibits endogenous inward rectifier potassium currents in RBL-2H3 cells, with a potency similar to that for block of I CRAC .In skeletal myoballs, however, sphingosine has no effect on voltage-gated, delayed rectifier K ϩ currents.Sphingosine also differentially affected voltage-gated Ca 2ϩ channels in two preparations.In chromaffin cells, sphingosine inhibits N-type Ca 2ϩ currents, with a potency greater than for block of I CRAC in RBL cells.With L-type Ca 2ϩ currents in skeletal myoballs, however, there was no effect.Finally, sphingosine strongly inhibits Na ϩ currents measured in chromaffin cells.Table II summarizes these results and compares them to block of I CRAC by sphingosine.
In all cases where block occurred, sphingosine also elevated membrane capacitance (Table II).Only in rat skeletal myoballs did sphingosine fail to block the two channel types tested or to A Link between Cellular Sphingosine and I CRAC -In order to determine whether a connection exists between cellular sphin-gosine and inhibition of I CRAC , we tested several growth factors known to activate SMase and/or to elevate sphingosine levels (9,(22)(23)(24)(25)(26)(27).Insulin growth factor (1 g/ml), platelet-derived growth factor (1 g/ml), stem cell factor (1 g/ml), or tumor necrosis factor (1 g/ml) had no significant effect on I CRAC in RBL-2H3 cells, when tested for over 10 min.Mast cells also express NGF receptors (47,48), and although we are unaware of a report with measurement of sphingosine levels following NGF treatment in mast cells, this growth factor is known to stimulate SM breakdown in other systems (22,23).
NGF could potentially act on I CRAC by elevating ceramide, sphingosine, and/or S1P levels.Because, however, S1P and ceramide do not mediate block of I CRAC by sphingosine (Table I), it seems that NGF inhibits I CRAC by elevating sphingosine levels, especially since other known effects of NGF could be ruled out through the following experimental observations.(i) Inhibition of I CRAC by NGF is not mediated by alterations in Ca 2ϩ store content, as NGF failed to release Ca 2ϩ from internal stores in fura-2/acetoxymethyl ester experiments with intact cells (n ϭ 4).(ii) Inhibition by NGF is not mediated by PKC, the only other inhibitory modulator of I CRAC currently known in these cells, as the selective PKC inhibitor bisindolylmaleimide (1 M; n ϭ 9) failed to significantly alter NGF-mediated inhibition of I CRAC .(iii) Inhibition by NGF is unlikely to result from phosphorylation events through other kinases either, since high concentrations of the less selective kinase inhibitor staurosporine (5 M; n ϭ 12) also failed to affect NGF-mediated inhibition of I CRAC .In addition, experiments in which ATP␥S (1 mM) was perfused into the cell revealed that the inhibition by ATP␥S was indistinguishable in control (n ϭ 11) and LCStreated cells (16 -24 h, 2 mM, see below; n ϭ 9; data not shown).Together, these results suggest that NGF most likely inhibits I CRAC via elevation of cellular sphingosine levels.
To test this notion further, we employed LCS which reduces SM levels (26) and should accordingly prevent NGF from maximally stimulating SMase and ceramidase (converting ceramide to sphingosine).Indeed, when treating cells for 16 -24 h with 2 mM LCS, the NGF-mediated inhibition of I CRAC was reduced by about half (Fig. 7).This reduction of the NGF effect by LCS is statistically significant (45 Ϯ 6% inhibition after 450 s; time constant ϭ 1552 Ϯ 285 s; n ϭ 10; p Ͻ 0.01; t test).The fact that LCS reduces the NGF effect on I CRAC by ϳ40% is consistent with the previously reported magnitude of reduction in sphingosine levels by about 50% in LCS-treated cells stimulated by platelet-derived growth factor (26).

DISCUSSION
The main conclusion of this paper is that sphingosine directly blocks I CRAC .The following lines of evidence support a direct block by sphingosine: neither ceramide, nor S1P, nor other SM pathway metabolites mediate the blocking effect of sphingosine.Furthermore, structural analogs of sphingosine (C 2 -ceramide, DMS, DL-threo-dihydrosphingosine), which can differentially modulate SM pathway intermediates, inhibit I CRAC .Also, our results strongly support the notion that sphingosine accumulates in the plasma membrane (which in and of itself does not cause block), and, last but not least, sphingosine blocks other ion channels.Therefore, it appears that both the sphingomyelin and phosphoinositide pathways can regulate I CRAC .The observation that NGF inhibits I CRAC by elevating sphingosine levels provides an important link between cellular sphingosine and I CRAC .Furthermore, we demonstrate that whole cell capacitance measurements provide an excellent indication of exogenously applied lipid accumulation in the plasma membrane.
Sphingosine block of I CRAC behaves in a dose-dependent and time-dependent manner.That is, lower concentrations of sphingosine inhibit more slowly, albeit completely, and assessing the potency of inhibition at a fixed point in time will underestimate the actual potency of sphingosine.To solve this problem we used both the amount of block at a fixed point in time (50 s of sphingosine application) and the time course of decay during application to assess the the value of half-maximal inhibition.Consideration of both methods of analyzing the dose-response relationship gives an IC 50 value of ϳ4 M.This value agrees well with estimates from previous reports of sphingosine-dependent inhibition of calcium signals in Jurkat T-lymphocytes (21), pancreatic duct adenocarcinoma cells (49), and sphingosine-mediated enhancement of epidermal growth factor-evoked Ca 2ϩ signals in human epidermoid carcinoma A431 cells (36).
Surprisingly, intracellularly applied sphingosine did not exhibit any inhibitory effect on I CRAC .However, since there was no change in the membrane capacitance, sphingosine applied in this manner presumably does not accumulate in the plasma membrane.Internally applied sphingosine may not accumulate in the plasma membrane because 1) it becomes quickly metabolized, 2) it traverses the plasma membrane too easily and escapes into the bath, or 3) it cannot assimilate into the inner leaflet of the membrane.The latter two explanations are unlikely, since sphingosine inhibits voltage-independent potassium channels when applied to either cell-attached, inside-out, or outside-out patches in smooth muscle cells (50).We suggest, therefore, that internally applied sphingosine may become metabolized before it can incorporate into the plasma membrane of RBL-2H3 cells.We cannot entirely discard the other explanations, since we also observed that membrane incorporation of sphingosine (as evidenced by capacitance changes) can be vari-FIG.5. Slow inhibition of I CRAC by a short chain ceramide analog.A, N-acetylsphingosine (C 2 -ceramide; 10 M) inhibits InsP 3 -evoked I CRAC , albeit slowly (n ϭ 6), compared with controls done in parallel (n ϭ 3).B, C1P, 50 M, significantly and slowly inhibits I CRAC activated by InsP 3 (n ϭ 8).However, N-octanoylsphingosine (C 8 -ceramide; 50 M) has no effect on I CRAC (n ϭ 6).

able in different cell types (see below).
Assessment of the mechanism by which sphingosine blocks I CRAC depends on whether this sphingolipid alters calcium homeostasis.The possibility that sphingosine inhibits I CRAC indirectly by promoting internal Ca 2ϩ store refilling seems to be unlikely.If sphingosine promoted refilling by enhancing microsomal Ca 2ϩ -ATPase activity then it should lower resting Ca 2ϩ levels, and we never observed this effect (n ϭ 37).Moreover, sphingosine has been shown to inhibit microsomal Ca 2ϩ -ATPase activity (18) and does not promote refilling in Jurkat T cells (21).
A second possibility is that sphingosine might block the InsP 3 receptor, thereby interfering with store depletion.Note, however, that in a previous report sphingosine failed to modify epidermal growth factor-stimulated calcium release in A431 cells (36).In our experiments with both clamped and unclamped (fura-2 experiments) cells, the Ca 2ϩ stores are rapidly depleted by InsP 3 , and sphingosine is acting after store depletion has occurred.We have previously demonstrated that under our experimental conditions (no internal ATP and 10 mM EGTA intracellularly), even a complete block of InsP 3 receptors (e.g. by heparin) would not turn off I CRAC (4).Furthermore, sphingosine remained an effective inhibitor of I CRAC when stores were depleted by thapsigargin.We therefore conclude that sphingosine is not blocking I CRAC indirectly via alteration of internal Ca 2ϩ stores (although such an effect remains a possibility when sphingosine is applied before depleting stores; we are currently investigating this possibility).If indeed sphingosine also affected the InsP 3 receptor, both effects could synergistically lead to an even stronger suppression of Ca 2ϩ influx.The present article, however, clearly demonstrates that sphingosine can potently and completely inhibit I CRAC directly after stores have been depleted.
To elucidate the mechanism of I CRAC inhibition induced by sphingosine, we tested sphingosine analogs and intermediates from the sphingomyelin pathway.Although structural analogs to sphingosine are similarly potent in inhibiting I CRAC , analogs derived from ceramide, and thus containing a sphingosine backbone, were either far less potent or had no effect at all (see Table I).These results obtained from patch-clamp measurements corroborate the Ca 2ϩ measurements from intact cells, and together the results demonstrate that only sphingolipids with structural similarity to sphingosine seem to be able to inhibit I CRAC (with the notable exception of S1P).
Interestingly, the phosphorylated form of C 8 -ceramide (C1P) does have an effect on I CRAC although with much less efficiency than C 2 -ceramide or even sphingosine itself (Table I).Why then does C 8 -ceramide fail to inhibit I CRAC ?It seems tempting to relate the inability of C 8 -ceramide to block I CRAC to its lack of incorporation in the plasma membrane, since application of both C1P and C 2 -ceramide elevate membrane capacitance, whereas C 8 -ceramide fails to do so (see "Elevation of Membrane Capacitance" under "Results").Still, C1P is five times less efficient in blocking I CRAC than C 2 -ceramide.Moreover, C1P is structurally less similar to sphingosine than C 2 -ceramide, and C 8 -ceramide has no effect.Therefore, we favor the view that structural similarity to sphingosine, and not ceramide itself, mediates the slow inhibition by the ceramide analogs C 2 -ceramide and C1P.
The fact that short-chained ceramide analogs (C 2 -ceramide and C 8 -ceramide 1-phosphate) inhibit I CRAC may have important implications for studies using short-chained ceramide analogs to mimic the actions of endogenous ceramides.For example, it was shown that C 8 -ceramide 1-phosphate stimulates DNA synthesis and cell division (41), an effect typically associated with sphingosine or S1P.However, our data support the idea that some effects of permeable ceramide analogs may occur because they mimic sphingosine actions.
In vivo, sphingosine is a secondary product derived from the membrane lipid sphingomyelin via ceramide formation.In principle, block of I CRAC could be mediated by sphingosine precursors, breakdown products, or enzymes involved in the SM cycle rather than by sphingosine itself.However, the experiments with sphingomyelin, palmitoyl coenzyme A, S1P, or fumonisin B 1 clearly show no effect of these substances on I CRAC .Even exogenous SMase does not affect calcium influx, and therefore, whereas some sphingosine may indeed become FIG.6. Sphingosine elevates whole cell membrane capacitance.I CRAC was activated by whole cell break-in with InsP 3 at time 0. I CRAC ramps and whole cell capacitance were measured simultaneously.A, a long sphingosine application (350 s) is shown from a representative cell.Inhibition of I CRAC by sphingosine followed a delay (below).The delay coincides with the time taken for a noticeable increase in capacitance (above).Throughout the sphingosine application, membrane capacitance increased slowly to a maximum of ϳ10% of the original value by the time sphingosine was removed.Note that I CRAC does not recover, but the membrane capacitance does slightly.B, a histogram plot of sphingosine concentration versus the average percent capacitance increase (ϮS.E.) after 150 s shows that the elevation of capacitance is concentration-dependent.Importantly, there was no significant increase (p ϭ 0.46; t test) in capacitance at the lowest concentration of 100 nM sphingosine (i.e. the same as in control experiments).The increase in capacitance is nearly 15% with 10 M sphingosine.The asterisk indicates statistical significance compared with the control mean (p Ͻ 0.01; t test).C, sphingosylphosphorylcholine (SPC) application (5 M) increases membrane capacitance without altering InsP 3 -evoked I CRAC .An experiment from a representative cell is shown.ceramide in the plasma membrane induced by SMase (10), ceramide does not appear to mediate block of I CRAC by exogenous sphingosine.The exception would be if ceramide took the form C 2 -ceramide, which inhibits I CRAC and can occur in vivo (51).
These results (summarized in Table I) thus demonstrate that sphingosine blocks I CRAC independently from ceramide, sphingosine 1-phosphate, or other sphingomyelin pathway intermediates.The only other reasonable alternative is that sphingosine itself blocks I CRAC directly or via a closely associated protein that interacts with I CRAC .This supports the established notion that fatty acids directly regulate ion channels (52).For example, it has been shown that AA modulates voltage-gated calcium channels (53,54).Moreover, fatty acids, lysophosphatidate, acyl coenzyme A, and sphingosine differentially regulate a voltage-independent potassium channel in smooth muscle cells (50).Concerning sphingosine, the effect is independent of soluble messengers, particularly in the case of excised-patch experiments.However, the report by Petrou et al. (50) did not rule out involvement of S1P or ceramide in the plasma membrane.We demonstrate that block by sphingosine is independent from ceramide and S1P, and the previous reports described above support our conclusion that sphingosine acts directly on I CRAC channels (or a closely associated protein).Our results differ, however, from those of Petrou et al. (50) for two reasons: 1) we did not observe opposite effects of positively and negatively charged lipids (i.e. in RBL cells we observe that palmitoyl coenzyme A, which is negatively charged, has no effect on I CRAC , despite the fact that it elevates membrane capacitance), and 2) unlike Petrou et al. (50) we observed that the inhibition of I CRAC by sphingosine is irreversible within tens of minutes.These differences suggest that sphingosine can differentially block ion channels in various preparations.
Effects on Membrane Capacitance-All the substances investigated in this study that inhibited calcium influx also significantly elevated membrane capacitance, indicating accumulation of these substances in the plasma membrane itself.Concomitant increase in capacitance does not seem to be a nonspecific effect since, as already stated under "Results," neither sphingosylphosphorylcholine nor palmitoyl coenzyme A block I CRAC although elevating capacitance.
Further evidence favoring a positive correlation between the ability of sphingosine to block calcium influx and its membrane incorporation emanates from our experiments assessing sphingosine effects on other ion channels in various tissues.Although all the channel types investigated in RBL cells or chromaffin cells showed significant block and capacitance increase upon exogenous sphingosine application, this sphingolipid failed both to block the channel types tested and to elevate membrane capacitance in rat skeletal myoballs (Table II).For the latter case, it seems that sphingosine either becomes rapidly metabolized or quickly transferred to a different intracellular compartment.Therefore, the extent of sphingosine accumulation in the plasma membrane may have a tremendous impact on block of ion channels.It would imply that sphingosine potentially could block all ion channels if allowed to accumulate in the plasma membrane, which, however, remains to be determined.A differential accumulation of sphingosine in different cell types might help explain the variability and differences in effects of sphingosine between preparations.The whole cell patch-clamp technique provides an excellent tool to measure whether or not sphingosine accumulates in the plasma membrane, important information regarding the site of action of this sphingolipid.
The ability of sphingosine to block many different ion channels may prove to have far-reaching physiological relevance.Along with preventing Ca 2ϩ influx, which would modulate cell growth, blocking other channels would decrease cellular excitability and could have various effects on membrane potential depending on the channel constituents in a particular cell type.Blocking all channels may actually be desirable at various times during the life of a cell.Whether or not a reduction in excitability (or an increase in membrane input resistance) is important for cellular growth and proliferation remains to be determined (55).
Physiological Significance-To establish a physiological relevance of I CRAC inhibition by sphingosine, we tested several growth factors known to activate SMase and/or to elevate sphingosine levels (9,(22)(23)(24)(25)(26)(27).However, we only found NGF to have significant inhibitory action regarding I CRAC , and this inhibition can be significantly prevented by LCS treatment.Our experiments show that neither PKC nor phosphorylation or alterations in Ca 2ϩ store content can account for the NGF effect on I CRAC , thus strongly favoring the notion that NGF inhibits I CRAC via elevation of cellular sphingosine levels.
NGF promotes mast cell survival (47,56), instead of proliferation which is the case for many growth factors (9).Mast cells produce NGF themselves or receive it from nervous, endocrine, or immune input (56).Inhibition of I CRAC by NGF may be an important signal in the process that leads to increased survival of mast cells.This places I CRAC in a favorable position as a potentially important ion channel for regulating mast cell growth and may also regulate growth in other cell types.Pharmacological tools aimed at inhibition of I CRAC may be useful, therefore, for enabling cell proliferation, and activators of I CRAC may mediate apoptosis.Such agents are still forthcoming.
Our observation that NGF inhibits I CRAC presumably via elevation of cellular sphingosine levels suggests that block of this SOC current contributes to cellular events attributed to sphingosine, such as proliferation and survival.Increases in intracellular calcium can inhibit cell growth (14 -16, 57), and therefore, the inhibition of I CRAC may be an important signal to prevent elevation of intracellular Ca 2ϩ levels by I CRAC in a situation where cell growth or division is necessary.
It is noteworthy that when the PLC and SM pathways act in parallel, the end result is inhibition of I CRAC .Whole cell break-in with InsP 3 followed by NGF application mimics coactivation of these pathways.In regard to regulation of I CRAC , the PLC and SM pathways may operate in parallel or at different times in the life of a cell.For example, agonists that stimulate only the PLC pathway can regulate I CRAC by activation of PKC (8).Activation of both the PLC and SM pathways by multiple agonists is necessary for sphingosine block of I CRAC (i.e. when sphingosine production becomes stimulated by external stimuli, such as NGF, following receptor-mediated Ca 2ϩ release and activation of I CRAC ).
Finally, we speculate that sphingosine might be involved in the activation mechanism of I CRAC .In the resting state, sphingosine could act as a blocker of I CRAC .Upon depletion of internal stores, metabolism of sphingosine, possibly by conversion to sphingosine 1-phosphate by sphingosine kinase, could lower the sphingosine levels and lead to the disinhibition of I CRAC .This would be compatible with the observed elevations in [Ca 2ϩ ] i associated with antigen-induced stimulation of sphingosine kinase and elevation of S1P (58).

FIG. 1 .
FIG.1.Exogenous sphingosine inhibits I CRAC .Whole cell patch-clamp experiments were conducted, and I CRAC was measured by applying voltage ramps from Ϫ100 to ϩ100 mV at a holding potential of 0 mV (see protocol, upper right).A, following break-in with InsP 3 (20 M) and 10 mM EGTA (see "Experimental Procedures") I CRAC develops and peaks within approximately 50 s (left panel).Currents were normalized to the amplitude at 50 s for comparison to experiments with sphingosine or other agents that were always applied at this time in patch-clamp experiments.The average from 22 cells is shown (ϮS.E.).Examples of ramp currents from a representative cell are shown to the right of the graph.There is virtually no change in the I CRAC amplitude, even after 250 -300 s following break-in.B, sphingosine (5 M) applied 50 s after break-in inhibits I CRAC following a delay of approximately 15 s.Illustrated in the graph is the average normalized time course from six cells (ϮS.E.).Ramp currents (right) from a representative cell show that I CRAC is almost completely inhibited after 200 s.C, sphingosine (5 M), when co-applied with the calcium pump inhibitor thapsigargin (1 M) prevents the thapsigargin-induced activation of I CRAC (n ϭ 3, open symbols).In control cells, extracellular application of thapsigargin (1 M) consistently activated I CRAC after a short delay (n ϭ 3, filled symbols).In both types of experiments, cells were perfused with the standard intracellular solution containing 10 mM EGTA and no added ATP.D, when applied intracellularly, sphingosine(10 M) does not inhibit I CRAC .Ionomycin (2 M) was used to activate I CRAC with and without sphingosine in the pipette solution.The graph in the left panel shows that no significant difference (p ϭ 0.2; t test) was observed between controls (n ϭ 3; time constant ϭ 1731 Ϯ 459 s) and cells treated with internal sphingosine (n ϭ 4; time constant ϭ 1062 Ϯ 79 s).Currents were normalized to whole cell capacitance.

FIG. 3 .
FIG. 3. Sphingosine blocks I CRAC activated by CCh in RBL-2H3 cells stably transfected with m1 muscarinic receptors.A, a saturating concentration of CCh (100 M) was applied to patch-clamped cells with and without sphingosine (5 M).In control cells, application of CCh activated I CRAC without much delay, to a peak amplitude of approximately Ϫ3 pA/pF (n ϭ 9).Sphingosine coapplication inhibited the amplitude by ϳ50% and caused the current to decay back toward base-line values.The time constant of decay during sphingosine treatment, 105 Ϯ 53 s (mean Ϯ S.E.; n ϭ 5), is significantly different from control values, 1278 Ϯ 204 s (n ϭ 9), as assessed by a t test (p Ͻ 0.01).B, averaged single cell calcium measurements (n ϭ 5-17) from intact cells confirm the patch-clamp results.In each case, 10 mM CsCl was included in the bath solution to mimic the solution used in patch experiments and to prevent changes in membrane potential caused by alterations of the inward rectifier potassium conductance.Sphingosine inhibits the component of the plateau phase of the CCh-activated Ca 2ϩ signal, in a manner similar to that of removal of external Ca 2ϩ (0 Ca 2ϩ ϭ nominally Ca 2ϩ -free, i.e. no Ca 2ϩ or EGTA added).

FIG. 4 .
FIG. 4. Sphingosine analogs inhibit I CRAC .A, DMS and DL-threo-dihydrosphingosine (DL-threo-DHS) inhibit InsP 3 -evoked I CRAC , whereas sphingosylphosphorylcholine (SPC) and S1P have no effect.DMS inhibits I CRAC by 96 Ϯ 3% (mean Ϯ S.E.; n ϭ 5) 250 s after addition, with a time constant of 42 Ϯ 4 s (n ϭ 6).DL-threo-DHS is slightly less potent, inhibiting I CRAC by 82 Ϯ 13% (n ϭ 5) after 250 s, with a time constant of 96 Ϯ 15 s (n ϭ 6).For a comparison of structures for these compounds see Table I.B, single cell Ca 2ϩ measurements with RBL-m1 cells confirmed the patch-clamp results.Averaged Ca 2ϩ signals are shown.SPC added together with 100 M CCh (n ϭ 9) had no significant effect on the Ca 2ϩ signal compared with controls (see Fig.3B).However, DMS (n ϭ 8) and DL-threo-DHS inhibited the plateau phase of the Ca 2ϩ signal with the same order of potency as in the patch-clamp experiments (i.e.DMS Ͼ DL-threo-DHS).None of these compounds significantly altered the peak elevation of Ca 2ϩ due to InsP 3 -mediated Ca 2ϩ release from internal stores.

c 18 ϮFIG. 7 .
FIG. 7. Nerve growth factor 2.5 S subunit (NGF) inhibits I CRAC .When applied 50 s after break-in with InsP 3 , NGF reduces I CRAC by 71 Ϯ 6% (n ϭ 10) in 450 s.Cells treated with LCS (2 mM) showed 24% less inhibition in response to NGF.LCS treatment involved incubation in growth medium for 16 -24 h and continued presence in the experimental and application pipette solutions.Application of the vehicle (control), 0.1% BSA, had no significant effect on I CRAC (7 Ϯ 4% inhibition after 450 s; n ϭ 5).Vehicle controls with LCS-treated cells (n ϭ 7) were not significantly different from untreated cells.

TABLE I
Profile of sphingosine and related compounds, block of I CRAC and elevation of membrane capacitance NS, not significant (p Ͼ 0.2) compared to controls (t test).*, statistically significant (p Ͻ 0.01) compared to controls (t test).Data are mean values Ϯ S.E.
a Endogenous lipid molecule.b Experiments done in 2 mM Ca 2ϩ external saline.

TABLE II
Differential block of ion channels by exogenous sphingosine (5 M) NA, not applicable, i.e. no block.Data are mean values Ϯ S.E.3% block of inward rectifier K ϩ current and a 9.4 Ϯ 2.5% increase in membrane capacitance occurred 150 s after adding sphingosine.b 17 Ϯ 3% block of delayed rectifier K ϩ current and a 0.2 Ϯ 4% increase in membrane capacitance occurred 150 s after adding sphingosine (n ϭ 7).