Activation of Casein Kinase I1 by Sphingosine*

activates casein kinase I1 in the presence of endogenous substrates as well as a synthetic peptide substrate. The activation response occurred between 12 and 25 wg/ml sphingosine and exhibited positive cooperativity with a Hill coefficient of 3.0. of casein the Km(app) for the peptide substrate from 0.5 0.08 mM. goid of the final enzyme preparations ranged from -30 t.o 70% and had a specific activity of 0.1-0.2 pmol/min/mg protein. The endogenous M, 42,000 substrate was partially purified from rat brain homogenates by extraction at pH 11, heat denat.uration at 70 "C, ammonium sulfate precipitation, and desalting on a Sephadex G-50 column (27), yielding a relatively crude substrate preparation which cont.ained multiple polypeptides hut. was devoid of detectable prot.ein kinase activity. Endogenous polypeptide phosphorylation was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (29) followed by autoradiography. Phosphate incorporation into the synthetic peptide was measured hy the 1'41 filter method (30). Assays were performed for 10 min at 30 "C in a 30-pl reaction mixture containing 25 mM Tris-HCI, pH 7.5, 5 mM MgCI,, 50 p~ ATP, 2 pCi of [y-'"P]ATP, 1 mM peptide suhstrate, and -20 ng of enzyme protein. Any changes in these parameters are specified where appropriate. Linear condi-tions for the assay of peptide phosphorylation were initially estab- lished prior to t,he'detailed examination of the regulation of enzyme activity. Assays were performed in triplicate, and each experiment was repeated a minimum of three times. Proteins were determined by the Bradford method (31).

over, the sphingosine effect could be abrogated by KC1 and NaC1, which alone are known to induce enzyme activation and dissociation of aggregated casein kinase I1 protein; LiCl and NH&l also inhibited the sphingosine effect. Polyamines, known activators of casein kinase 11, partially mimicked the effect of sphingosine on endogenous polypeptide phosphorylation but failed to do so with the peptide substrate. These observations demonstrate that sphingosine is a potent activator of casein kinase 11. The potential pharmacological and physiological modulation of casein kinase I1 by sphingoid bases is discussed.
Sphingolipids are membrane lipids with important roles in cell growth, differentiation, and oncogenesis (1,2). The discovery that sphingosine and related lysosphingolipids inhibit protein kinase C activity in uitro (3,4) and in cell systems (5)(6)(7) established important links between sphingolipids and signal transduction mechanisms. Sphingosine also modulates the activity of other biochemical targets in uitro; it inhibits phosphatidate phosphohydrolase (8), calmodulin-dependent enzymes (9), binding of factor VI1 to tissue factor (lo), and * This work was supported in part by National Institutes of Health Grant GM-43825. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ll Pew scholar. **To whom correspondence should be addressed Div. of Cell Biology, Burroughs Wellcome Co., Research Triangle Park, NC 27709. Tel.: 919-248-4184. binding of thyrotropin releasing hormone to its receptor (11). Sphingosine has also been shown to activate the epidermal growth factor receptor kinase (12) and to activate phospholipase D (13,14). At the cellular level, sphingosine appears to inhibit most, if not all, protein kinase C-dependent processes (1,15). It also, however, appears to influence activities not related to protein kinase C. Most notably, sphingosine appears to be mitogenic to Swiss 3T3 cell fibroblasts at low concentrations (16) and to inhibit tissue factor activity in mononuclear cells (10).
Casein kinase I1 (CK-11)' is a tetrameric protein kinase (17, 18) found in cytoplasm and nucleus (19). It is distinguished by its specificity for serine residues within clusters of acidic amino acids (20), its ability to use either GTP or ATP as nucleotide substrates (17), and its inhibition by low concentrations of heparin (21). Although CK-I1 is activated in response to growth factors (22, 23) and during different phases of the cell cycle (24,25), the actual biochemical steps involved in the regulation of CK-I1 activity are poorly understood. The enzyme is also activated by high concentrations of polyamines (26) although the significance of this regulation is unknown.
During the characterization of a rat brain sphingosineactivated protein kinase (27), it was noted that this enzyme had chromatographic and catalytic properties, suggesting that it may be identical to CK-11. This observation prompted the present study aimed at investigating the effects of sphingosine on purified CK-11.

MATERIALS AND METHODS
[y-"P]ATP (-30 Ci/mm) was from Du Pont-New England Nuclear. Spermine, spermidine, D-erythrosphingosine, D-erthyrosphinganine, octadecylamine, psychosine, sphingosylphosphorylcholine, and all the phospholipids were from Sigma. C2-ceramide was synthesized by acetylation of sphingosine as described (28). Biosynthesis Inc. supplied the synthetic peptide substrate RRREEETEEE. Phospholipid suspensions were prepared by drying the organic solvent under N2 followed by sonication in 25 mM Tris-HC1, pH 7.7. The other lipids were dissolved as concentrated (5-10 mg/ml) solutions in dimethyl sulfoxide or ethanol/H20. CK-I1 was purified from adult male Sprague-Dawley rat brains. Fifteen brains were homogenized in 6 volumes of a buffer containing 25 mM Tris-HC1, pH 7.7, 1 mM EDTA, 2 mM dithiothreitol, 100 pg/ ml phenylmethylsulfonyl fluoride, and 10 gg/ml leupeptin. The 250,000 X g supernatant was successively chromatographed on DEAE-cellulose, octyl-Sepharose, heparin-Sepharose, and casein-Sepharose columns; the identity of the purified CK-I1 was ascertained by its alp-subunit composition, &subunit autophosphorylation, utilization of GTP as a substrate, inhibition by low concentrations of heparin, and immunoreactivity with antisera against calf thymus and hog spleen CK-I1 (gifts from Drs. M. E. Dahmus 42,000 substrate was partially purified from rat brain homogenates by extraction at pH 11, heat denat.uration a t 70 "C, ammonium sulfate precipitation, and desalting on a Sephadex G-50 column (27), yielding a relatively crude substrate preparation which cont.ained multiple polypeptides hut. was devoid of detectable prot.ein kinase activity. Endogenous polypeptide phosphorylation was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (29) followed by autoradiography. Phosphate incorporation into the synthetic peptide was measured hy the 1'41 filter method (30). Assays were performed for 10 min a t 30 "C in a 30-pl reaction mixture containing 25 mM Tris-HCI, pH 7.5, 5 mM MgCI,, 50 p~ ATP, 2 pCi of [y-'"P]ATP, 1 mM peptide suhstrate, and -20 ng of enzyme protein. Any changes in these parameters are specified where appropriate. Linear conditions for the assay of peptide phosphorylation were initially established prior to t,he'detailed examination of the regulation of enzyme activity. Assays were performed in triplicate, and each experiment was repeated a minimum of three times. Proteins were determined by the Bradford method (31).

RESULTS AND DISCUSSION
Sphingosine stimulated the phosphorylation of selected endogenous brain polypeptides by CK-I1 (Fig. 1) including a M , 42,000 component whose phosphorylation was also stimulated by higher concentrations of polyamines (Fig. 1). These observations are consistent with the ability of sphingosine to stimulate the phosphorylation of endogenous synaptosomal polypeptides (27). Whether any of these polypeptides is a physiologic substrate for CK-I1 has not been verified.
In order to characterize quantitatively the activation of CK-I1 by sphingosine, we resorted to the use of the synthetic peptide RRREEETEEE, previously designed as a selective CK-I1 substrate (32). Sphingosine activation was evident a t concentrations above 12 pg/ml, reaching a maximum a t 25 pg/ml (Fig. 2). This relatively narrow concentration range reflects the sigmoidal nature of the dependence of CK-I1 activation on the concentration of sphingosine. In fact, a Hill  plot of the data (Fig. 2, inset) yielded a Hill coefficient of 3.0, confirming the positive cooperativity of the sphingosine response. Whether this cooperative response stems solely from the lipid-enzyme interaction or from lipid-lipid association has not been determined. A possible interaction between sphingosine and CK-I1 substrates has not been ruled out either. In contrast to sphingosine, polyamines failed to activate CK-I1 phosphorylation of t.he synthetic substrate (data not shown). Combined, Figs. 1 and 2 show that sphingosine stimulates CK-I1 phosphorylation of endogenous pol-ypeptides as well as synthetic peptides.
The sphingosine effect was further investigated by varying individually the concentrations of ATP, peptide substrate, or MgCI, a t fixed levels of the other two variables. Sphingosine increased the V,,, of CK-I1 2.3-fold when the enzyme was assayed a t different ATP concentrations (Fig. 3A ); however, the Km,nppl for ATP (7.5 p~) remained constant. Sphingosine also increased the VmnX values, which were obtained by changing the concentration of peptide substrate or MgCI, at. fixed ATP levels. The Km,npp, for MgCI, was increased from 0.12 to 0.7 mM (Fig. 3 R ) . More importantly, sphingosine caused a significant decrease in the Kmlnp,,, value for the peptide from 0.5 to 0.08 mM (Fig. X ) . Thus a major effect of sphingosine on CK-I1 appears to be an enhancement of affinity to peptide/ protein substrates. This mechanism appears to be shared by several second messengers which activate protein kinases by relieving the inhibitory effect of intrinsic pseudosubstrate domains (33). The effects of sphingosine on CK-I1 may result from a direct enzyme-sphingosine association which exerts an allosteric effect on the interaction of the CK-I1 with its substrate and with MgC12.
Activation of CK-11 by sphingosine displayed significant structural selectivity. Octadecylamine and psychosine were less effective activators of the kinase, behaving as partial agonists. Ceramide and sphingosylphosphorylcholine were without significant effect ( Table I). Sphinganine (dihydrosphingosine) could not be evaluated because of high filter background. These results indicate that a free NH? is required for activity and that a C1 hydroxyl may increase the potency. Substitutions at the C1 carbon significantly decrease (psvchosine) or eliminate (sphingosylphosphorvlcholine) activity, suggesting steric hindrance.
Although ceramide and sphingosylphosphorylcholine did not alter CK-I1 activity when used alone, they were able to modify the response of the kinase to sphingosine itself. Accordingly, these two compounds (at 30 pg/ml) were able to enhance the sphingosine effect by 60 and 90%, respectively (data not shown).

Effects of sphingosine-related lipids o n casein kinase II activity
The lipids were added at the designated concentrations; the increase in enzyme activity produced by 30 pg/ml sphingosine was considered to be 100%. The four major phospholipids, PC, PE, PI, and PS, were also examined for a possible direct effect on CK-I1 activity or on its response to sphingosine. When the phospholipids were added individually, they had no apparent influence on CK-I1 activity; however, they were able to modify the stimulation of the protein kinase by sphingosine (Fig. 4). PS abolished the sphingosine response fully at the lowest concentration tested  (5 pg/ml) (data not shown). PE and PI behaved similarly, although complete inhibition of the sphingosine effect required 20 pg/ml of either phospholipid. PI appeared to display a biphasic effect insofar as it potentiated the sphingosine response by 30% at 5 pg/ml phospholipid. This biphasic response was even more accentuated in the case of PC, which potentiated the sphingosine response by 50% at 10 pg/ml phospholipid while inhibiting the response by 40% at 80 pg/ ml phospholipid. The mechanism by which various lipids antagonize the activation of CK-I1 by sphingosine has not been addressed, although it may entail enzyme-phospholipid binding and/or phospholipid-sphingosine association. NaCl and KC1 are known to stimulate CK-I1 activity (34, 35) and appear to control the state of aggregation of casein kinase I1 by preventing the formation of elongated enzyme "polymers," which are observed in the absence of salt (36). These effects of KC1 and NaCl on the activity of CK-I1 and its molecular state prompted us to examine the influence of KC1 on phosphorylation of the synthetic substrate by CK-I1 and on the sphingosine response of the protein kinase. In the absence of sphingosine, KC1 exerted a modest biphasic effect on the phosphorylation of the synthetic peptide by casein kinase I1 with a maximum of about 40% activation at 100 mM KC1 (Fig. 5 ) . In the presence of sphingosine, a small stimu-II by Sphingosine latory increment was detected at low salt concentrations, but a purely inhibitory influence was seen at concentrations above 50 mM. At 150 m M KC1, enzyme activities measured in the presence or absence of sphingosine were nearly identical (Fig.  5). Other salts also had an inhibitory effect such that 15, 30, 38, and 48% inhibition of the sphingosine effect was produced by 100 mM LiCl, NaCI, NH&I, and KCl, respectively. The inhibitory salt effect is in fact reminiscent of the ability of KC1 and NaCl to interfere in uitro with protein kinase C activation by lipids (37). However, it should be noted that activation of protein kinase C by diacylglycerol is inhibited by 150 mM salt only in uitro, but not in cells (37). Consequently, the in uitro effect of salt on sphingosine-CK-I1 interaction may not be duplicated in intact cells.

CONCLUDING REMARKS
The present study demonstrates that sphingosine can serve as an activator as well as inhibitor of different protein kinases. Thus, sphingosine inhibits protein kinase C and calmodulindependent kinases but can activate CK-TI. The dual action of sphingosine is well illustrated by its effects on epidermal growth factor receptor phosphorylation (12,38). As expected, sphingosine inhibited the phorbol ester-induced phosphorylation of the receptor (12); however, sphingosine alone induced the phosphorylation of a threonine residue by a mechanism that appears to be independent of protein kinase C. Sphingosine is a potent activator of CK-I1 in uitro, and its effect is potentiated by low concentrations of phosphatidylcholine, suggesting that sphingosine may activate the enzyme in the presence of certain cellular membranes. Sphingosine may also enhance the binding of CK-I1 to membranes in analogy with the recruitment of protein kinase C to cell membranes by diacylglycerol. Alternatively sphingosine, which partitions to a significant level in cytosol (39), may activate CK-I1 in cytosol and not in membranes. Moreover, regulation of CK-I1 by sphingosine displays high structural specificity not found with other targets of sphingosine.
Whether sphingosine can be used as a useful reagent to study the consequences of CK-I1 stimulation in intact cells and the phosphorylation of putative endogenous substrates such as Myb (40), Myc (41), topoisomerases I and I1 (42,43), and GAP-43 (44) has not been evaluated. The ability of sphingosine to activate CK-I1 in intact cells would presumably depend on the accessibility of the lipid to its target enzyme, the presence of other factors which can modulate the response of casein kinase I1 to sphingosine, and the concurrent regulation of casein kinase I1 activity by specific growth factors such as epidermal growth factor and IGF 1 (29, 45). Further studies should aim at elucidating the effects of sphingosine on cellular CK-I1 and the ability of sphingosine to modulate the phosphorylation of cellular substrates of casein kinase 11. Determining of intracellular levels of sphingosine in response to growth factors, insulin, and other potential activators of CK-I1 should provide an important link in the physiologic role of sphingosine in activating CK-IT and in modulating other molecular targets. 14.