Eur. J. Biochem. 266, 855±864 (1999) q FEBS 1999

Anti-proliferative effect of two novel palmitoyl-carnitine analogs, selective inhibitors of protein kinase C conventional isoenzymes

Tea Garcia-Huidobro, Enrique Valenzuela, Andrea V. Leisewitz, Jaime Valderrama and Miguel Bronfman Departamento de BiologõÂa Celular y Molecular, Facultad de Ciencias BioloÂgicas y Facultad de QuõÂmica, P. Universidad CatoÂlica de Chile, Santiago, Chile

Previous studies have shown that palmitoyl-carnitine is an anti-proliferative agent and a protein kinase C inhibitor. Two new palmitoyl-carnitine analogs were synthesized by replacing the ester bond with a metabolically more stable ether bond. An LD50 value in the nm range was found in anti-proliferative assays using HL-60 cells and was dependent on the alkyl-chain length. The inhibitory action of these water-soluble compounds on protein kinase C in vitro was greatly increased with respect to palmitoyl-carnitine and was dependent on the length of the alkyl chain. Its effect was mediated by an increase in the enzyme's requirement for phosphatidylserine. Inhibition of the in situ phosphorylation of a physiological platelet protein kinase C substrate and of phorbol ester-induced differentiation of HL-60 cells was also observed. Finally, to test for isoenzyme selectivity, several human recombinant protein kinase C isoforms were used. Only the Ca2+-dependent classic protein kinase Cs (a, bI, bII and g) were inhibited by these compounds, yet the activities of casein kinase I, Ca2+/calmodulin-dependent kinase and cAMP-dependent protein kinase were unaffected. Thus, these novel inhibitors appear to be both protein kinase C and isozyme selective. They may be useful in assessing the individual roles of protein kinase C isoforms in cell proliferation and tumor development and may be rational candidates for anti-neoplasic drug design. Keywords: anti-proliferative drugs; palmitoyl-carnitine; protein kinase C isoenzymes.

Protein kinase C (pkC) isoenzymes play a major role in signal isoforms and are not selective for pkC because other transduction pathways affecting proliferation, differentiation intracellular receptors for phorbol esters have been identified and tumor development [1±4]. In this context, the design of [12]. Similarly, staurosporine [13], the most potent pkC selective pkC inhibitors is of interest because of their potential inhibitor described to date (IC50 = 6±10 nm), is also the most role as anti-proliferative and anti-neoplasic agents. The unspecific, as it inhibits other kinases with similar efficiencies classical isoforms of pkC (cpkC-a, cpkC-bI, cpkC-bII and [14,15]. The search for more selective pkC inhibitors has cpkC-g) are activated by Ca2+ and diacylglycerol or phorbol yielded many structurally distinct compounds directed against ester, and phosphatidylserine (PtdSer), whereas the novel either the catalytic or the regulatory domains of the enzyme [7], isozymes (npkC-d, npkC-:, npkC-u, npkC-h/L and npkC-m) although few of these compounds are truly isoenzyme selective. 2 respond to diacylglycerol and phorbol esters but not to Ca ‡. In early work on pkC inhibitors, alkyl-lysophospholipids [16] Finally, the atypical isoforms (apkC-z and apkC-zl/i) are were shown to possess selective anti-neoplasic activity in many activated only by anionic phospholipids such as PtdSer [2]. The cell cultures and in vivo [16±18]. Much of this work placed various pkC isoforms exhibit marked differences in tissue and special emphasis on ether analogs because of their higher subcellular distribution [5], compartmentalization [6], substrate metabolic stability and longer half-life [19]. Ether lysophos- specificity [7,8], protein±protein interactions [9,10] and pholipids were also found to be more toxic in cell cultures susceptibility to translocation and down-regulation [11]. than their respective ester analogs [19]. In the present study, These differences are suggestive of functional divergence we wished to combine this precedent with previous work on between individual pkC isotypes, although the exact roles of l-palmitoyl-carnitine (palm-car) in order to generate a new particular isoforms, particularly in tumor promotion, are as yet pkC inhibitor. Palm-car is an intermediate with long-chain unclear, in part because of the lack of pkC-isozyme-selective fatty acid metabolism and is capable of inhibiting cellular activators and inhibitors. Although phorbol esters are potent proliferation, differentiation and tumor promotion induced by pkC activators, they do not distinguish between individual phorbol esters, as well as pkC activity in vitro [20±23]. However, the in vivo inhibitor potential of palm-car is limited Correspondence to M. Bronfman, Departamento de BiologõÂa Celular by its rapid metabolic turnover, which is largely facilitated by y Molecular, Facultad de Ciencias BioloÂgicas, P. Universidad CatoÂlica its ester linkage and the breakdown of the acyl chain. Here we de Chile, Casilla 114-D, Santiago, Chile. Tel.: +56 2 686 2833, synthesized and evaluated two new palm-car analogs in which Fax: +56 2 635 2499, E-mail: [email protected] the ester linkages were replaced by an ether bond, to augment Abbreviations: CaMKII, Ca2+/calmodulin-dependent kinase II; CHO, the compounds' metabolic stability. We also verified the Chinese hamster ovary cells; CKI, casein kinase I; DO, diolein; importance of the acyl moiety by testing ether carnitines of palm-car, palmitoyl-carnitine; palm-CoA, palmitoyl-CoA; PtdIns(3,4,5)P3, varying alkyl-chain lengths, as well as nafenopin and phosphatidylinositol-3,4,5-triphosphate; pkA, cyclic AMP-dependent ciprofibrate acyl carnitines, two carboxylic acid-containing protein kinase; pkC, protein kinase C; PMA, 4b-phorbol 12-myristate carcinogenic peroxisome proliferators. The two novel palm-car 13-acetate; PtdSer, phosphatidylserine. analogs, POC-16 and AM-44, showed increased inhibitory (Received 7 September 1999, accepted 4 October 1999) capacity with respect to palm-car, both in proliferation studies 856 T. Garcia-Huidobro et al. (Eur. J. Biochem. 266) q FEBS 1999 using human leukemia cell cultures (HL-60), in which AM-44 penicillin and 100 mg´mL21 streptomycin. For proliferation showed greater potency, and pkC activity assays in vitro,in inhibition experiments, cells were plated in fresh medium at a which several pkC isoforms were tested. Similarly, the carnitine density of 1 105 cells´mL21 per well in 24-well plates for analogs inhibited the thrombin-induced phosphorylation of a HL-60, and at 5 105 cells´mL21 in 35 mm plates for CHO, physiological pkC substrate in intact human platelets and to which the indicated drug concentrations were applied in phorbol ester-induced differentiation of HL-60 cells. Both duplicate. Cells were then incubated at 37 8C in a humidified derivatives were found to be both pkC selective and isoenzyme 5% CO2 atmosphere for 24, 48 or 72 h, and cell viability selective. Only the Ca2+-dependent cpkCs (a, bI, bII and g) was calculated using the Trypan Blue exclusion method. The were inhibited by these compounds, while other second- HL-60 differentiation assay was performed as described messenger kinases were not. previously [25].

EXPERIMENTAL PROCEDURES Incorporation of 32P orthophosphate into platelet proteins Materials Platelets were obtained from freshly EDTA-anticoagulated Human recombinant pkC isozymes were obtained from Pan human blood, as described in Orellana et al. [24], and diluted to Vera Corp. (Madison, WI, USA) and the epsilon substrate 2 109 platelets´mL21 in 15 mm Tris/HCl pH 7.4 containing peptide from QCB Inc. (Hopkinton, MA, USA). Histone V-S 140Â mm NaCl, 5.5 mm glucose and 1 mg´mL21 of bovine (calf thymus), l-a-phosphatidylserine (P5660), diolein (DO), serum albumin. For 32P protein-labeling experiments, platelets ATP, 4b-phorbol 12-myristate 13-acetate (PMA) and cipro- 32 32 were incubated in the same medium as above plus P fibrate were purchased from Sigma Chemicals and [g- P]ATP 21 21 32 ortophosphate (50 mCi´mL ) for 1 h at 25 8C with gentle (10.0 mCi´mL ) and P phosphate from NEN Research agitation. Platelets were then incubated for 3 min at 37 8C and Products (Boston, MA, USA). Chemicals for drug synthesis a further 2 min in the presence of the acyl-carnitine derivatives were purchased from Aldrich (Milwaukee, WI, USA) and at the desired concentrations. Stimulation with thrombin nafenopin was provided by CIBA GEIGY (Basel, Switzerland). (0.25 U´mL21 final concentration) was carried out for 1 min All other reagents were obtained from commercial sources in a total volume of 60 mL. The incubation was terminated unless stated otherwise. The bisindolylmaleimide selective pkC by the addition of 40 mL SDS buffer [0.25 m Tris/HCl inhibitor Ro 31-8220 was donated by Roche Pharmaceuticals pH 6.8 containing 40% glycerol (v/v), 20% 2-mercaptoetha- (Herstfordshire, UK). nol (v/v), 10% SDS, and 0.2 mg´mL21 of bromophenol blue] and heating at 100 8C for 2 min. SDS/PAGE was then performed according to Laemmli [26] using 11% (w/v) Cell cultures acrylamide gels, that were stained with Coomassie Brilliant Human leukemia (HL-60) cells were maintained in RPMI- Blue R, dried and exposed to Kodak X-Omat film for 21 1640 medium supplemented with 2 mg´mL NaHCO3, autoradiography. Densitometric analysis of the autoradiograph 100 U´mL21 penicillin, 100 mg´mL21 streptomycin and either was performed using a CS-9000 scanning densitometer 2% or 5% heat-inactivated fetal bovine serum. Chinese hamster (Shimadzu, Kyoto, Japan), and the relative content of phos- ovary (CHO) cells were maintained in HamF12 medium phorylated proteins was expressed in arbitrary densitometric supplemented with 5% fetal bovine serum, 100 U´mL21 units.

Fig. 1. Structure of acyl-carnitines and alkyl-carnitines. Nafenopin-carnitine (I) and ciprofibroyl-carnitine (II) were synthesized as described previously [27]. Alkyl-carnitine ethers were synthesized from 1±2 epoxides of varying chain lengths and l-carnitine (see Experimental procedures) yielding a mixture with proportions of 65±70% compound III and 30±35% compound IV. n = 13: hexadecyl-O-ether-carnitine (POC-16); n =9: dodecanoyl-O-ether-carnitine (LOC-12); n =3: hexyl-O-ether-carnitine(HOC-6). The POC-16 derivative (V): 2-hexadecanoxy-propyl trimethyl ammonium chloride (AM-44), was synthesized to remove the carboxyl and hydroxyl groups from the carnitine and palmitoyl moieties, respectively. All structures, compound ratios and concentrations were determined by NMR. q FEBS 1999 Conventional protein kinase C isoenzyme inhibitors (Eur. J. Biochem. 266) 857

cpkC (a, bI, bII and g), the reaction medium contained 20 mm Synthesis of acyl-carnitine esters and alkyl-O-ether-carnitines 21 Hepes (pH 7.4), 100 mm CaCl2,10mmMgCl2, 200 mg´mL The carnitine esters of the peroxisome proliferators nafenopin histone V-S, 100 mm ATP, 100 mg´mL21 PtdSer, 20 mg´mL21 and ciprofibrate (Fig. 1, structures I and II) were synthesized DO, 0.03% Triton X-100 and trace [g-32P]ATP. For npkC (d according to Cervenka and Osmundsen [27] and their structures and :), the reaction medium was the same except that CaCl2 assessed by proton NMR. Alkyl-ether-carnitines of different was omitted and histone was substituted for 50 mg´mL21 chain lengths were synthesized by dissolving 0.25 g epsilon substrate peptide. The same reaction medium as npkC (1.26 mmol) of l-carnitine (inner salt) in 1 mL of methanol was used for apkC (z), except that DO was absent from the lipid together with a trace of 3H-carnitine, after which they were mix. The lipid mix, PtdSer/DO or PtdSer alone, was freshly dried under a N2 stream and lyophilized. To this, 2 mL of a prepared in CHCl3/methanol (9 : 1), dried with a N2 stream mixture of 1-2 epoxyhexadecane, 1-2 epoxydodecane or 1-2 and then resuspended in 10 mm Hepes (pH 7.4), 0.3% Triton epoxyhexane (10.5 mmol) and glacial acetic (0.5 mL) was X-100. added and the solution was maintained at 45 8C for 20 h. The reaction mixture was then loaded onto a silica gel column that had been equilibrated previously and washed (4 vol.) with Other kinase assays chloroform/methanol (8 : 1) and eluted using methanol/chloro- The effect of POC-16 on the activity of the catalytic fragment form/water/ammonium/formic acid (55 : 50 : 10 : 7.5 : 2.5). of pkA (cAMP-dependent protein kinase) was assessed using Fractions were then collected to obtain two radioactivity the method of Kikkawa et al. [29]. Recombinant casein peaks, the first corresponding to a mixture of the two expected kinase I (CKIa) from Xenopus laevis was a kind donation related compounds, alkyl-ether carnitines (Fig. 1, structures III from C. Connelly (Faculty of Medicine, Universidad de Chile). and IV), and the second to excess l-carnitine. The solvent of CKI was assayed under the same conditions as pkC isoforms, each ether peak was evaporated under a N2 stream and the with a few changes in the reaction mixture, 50 mm Hepes product dissolved in acetone and finally crystallized from (pH 7.5), 100 mm CaCl2,8mmMgCl2, 0.5 mm dithio- acetone/water. From this, a mixture of the two compounds was threitol, 5 mg´mL21 casein, 100 mm ATP and trace [g-32P]ATP. obtained, with a yield ranging from 20 to 50%. The chemical Ca2+/Calmodulin-dependent kinase II (CaMKII) activity was structures, concentrations and proportion of the two compounds assayed in fresh rat brain supernatants using the GIBCO BRL present (roughly 70% of III and 30% IV in all cases) were assay kit, according to the manufacturer's instructions assessed by proton NMR. Another alkyl-O-ether analog of palm-car, 2 hexadecanoxy-propyl trimetyl ammonium chloride (AM-44), which does not contain the carboxylic acid moiety RESULTS of carnitine and the hydroxyl group of the palmitoyl chain We have previously shown that palmitoyl-CoA (palm-CoA), (Fig. 1, structure V), was synthesized using Williamson which in the cell is in equilibrium with palm-carn, and the CoA synthesis through the reaction of sodium 1-dimethylamino- esters of ciprofibrate and nafenopin, two carcinogenic per- 2-propanoxide over hexadecyl bromide. The ether formed oxisome proliferators, are able to activate pkC [24,31,32], was permethylated in the presence of methyl iodide and AM-44 whereas palm-car acts as an inhibitor of the enzyme. Based on was obtained by pouring a solution of the resulting 2-hexa- this premise, the ability of the carnitine derivatives of decanoxy-propyl trimetylammonium iodide through an anionic ciprofibrate and nafenopin (Fig. 1, structures I and II) to inhibit exchange column in the chloride form. The chemical structure pkC was first evaluated in vitro using rat brain purified enzyme. of AM-44 was established by proton and 13C NMR. For each Both ciprofibroyl-carnitine and nafenopin-carnitine were found experiment, appropriate dilutions of the acyl-carnitine deriva- to have very low inhibitory activity compared with palm-car tives (1/10±1/40) were prepared in water and maintained at (Fig. 2), suggesting that an aliphatic acyl chain was necessary 220 8C. For cell culture experiments, stock solutions were within acyl-carnitine to inhibit pkC. Consequently, ether- filtered before preparing dilutions under sterile conditions. carnitine compounds of varying alkyl-chain lengths were synthesized in order to further investigate the importance of chain length and to optimize their inhibitor potential by Purification of rat brain pkC and activity assay introducing an ether linkage. The inhibitory capacity of these Rat brain pkC was purified to homogeneity using the method novel ether compounds was assessed on pkC in vitro and on the described by Woodgett and Hunter [28] and was assayed under proliferation of CHO cells. In both assays, the efficiency of the general conditions given by Kikkawa et al. [29], with the changes already described in Orellana et al. [24]. PtdSer and diacylglycerol (diolein; DO) were added to the reaction as mixed micelles with Triton X-100, prepared according to Hannun et al. [30].

Recombinant pkC isoforms activity assay Individual recombinant pkC isoforms were assayed according to the manufacturer's instructions with several modifications. All assays were carried out in triplicate, with a total volume of 30 mL. Assays were started by the addition of the enzyme (final Fig. 2. Inhibition of pkC activity by palmitoyl-carnitine and xenobiotic- 21 concentration 40 ng´mL ), incubated at 30 8C for 10 min and carnitine esters. The inhibitory actions of nafenopin-carnitine (A), stopped by spotting the entire volume on Whatman P-81 phos- ciprofibroyl-carnitine (X) and palm-car (W) were tested on pkC purified phocellulose strips. The membranes were then washed twice for from rat brain at the indicated concentrations (5, 10, 20, 50 and 100 mm). 10 min in 300 mL of 0.5% phosphoric acid under agitation, Only palm-car was seen to inhibit pkC in a concentration-dependent placed to dry and counted in 3 mL of scintillation fluid. For manner. 858 T. Garcia-Huidobro et al. (Eur. J. Biochem. 266) q FEBS 1999

synthesized, in which both the primary alcohol of the acyl chain and the carboxyl group of the carnitine moiety were removed. We therefore avoided the two-compound formation and assessed the importance of the positive charge, with the expectation that the selected configuration would prove more active. The structure of this molecule (AM-44; Fig. 1, structure V) resembles that of the predominant molecule in POC-16, but consists of a positively charged rather than a zwitterionic molecule. The anti-proliferative effects of POC-16 and AM-44 were compared using human promyelocytic leukemia cells HL-60. Cells maintained in either 2% or 5% fetal bovine serum were Fig. 3. Inhibition of pkC activity and CHO cell proliferation by alkyl- incubated in the presence of each drug for 24, 48 and 72 h, after carnitine ethers of varying chain lengths. (A) pkC purified from rat brain which viable cells were counted by Trypan Blue exclusion. was assayed in the presence of cofactors and increasing concentrations of Both POC-16 (Fig. 5A) and AM-44 inhibited the proliferation hexyl-O-carnitine (HOC-6, O), dodecanoyl-O-carnitine (LOC-12, B) and of HL-60 cells in a concentration-dependent manner, although hexadecyl-O-carnitine (POC-16, X). Inhibition of pkC activity was seen to AM-44 proved to be more potent (Fig. 5B). The inhibitory increase with the increase in chain length in a concentration-dependent effect of either compound was also dependent on the manner. (B) CHO cells (5 105 cells) were seeded in 35 mm plates and  concentration of serum used. The most marked effects were incubated for 48 h at increasing concentrations of HOC-6 (O), LOC-12 (B) seen in cells grown in 2% fetal bovine serum, whereas cells and POC-16 (X). Viable cells were then counted by Trypan Blue exclusion grown in 5% fetal bovine serum were more resistant to higher and results were expressed as percentage of control values. Variability drug concentrations, regardless of the incubation period. Only within points was , 10%. Similar results were obtained in three the results at 48 h are shown here for POC-16 and AM-44 independent experiments. Inhibition of proliferation resulted in a con- 3 (Fig. 5), although similar results were obtained for all three comitant reduction of H-thymidine incorporation into DNA (not shown). incubation periods, with only a slight increase in variability observed at 72 h. This was probably due to an increment in the instability of the drugs, although both compounds proved to be these compounds was seen to increase with the length of their sufficiently stable for 72 h. Hence, the LD50 for POC-16 ranged alkyl chains: hexadecyl (POC-16) . dodecanoyl (LOC-12) from 3.5 to 4.5 mm for cells maintained in 2% fetal bovine . hexyl (HOC-6). At the highest concentration tested serum and from 8 to 12 mm for cells maintained in 5% fetal (100 mm), pkC activity was inhibited 95% by POC-16, 50% bovine serum. However, in the case of AM-44, LD50 values for by LOC-12 and 10% by HOC-6 (Fig. 3A). Similarly, incu- cells grown in 2% fetal bovine serum varied according to the bation of CHO cells with these derivatives for 48 h inhibited incubation periods with values ranging from 0.1 to 0.9 mm, proliferation by 95% (POC-16), 80% (LOC-12) and 40% and an average of 0.62 mm was calculated at 48 h. In (HOC-6; Fig. 3B), thus correlating with their effect on pkC contrast, AM-44-exposed cells grown in 5% fetal bovine serum inhibition. Under similar experimental conditions, POC-16 was produced an equivalent response for all three incubation seen to produce half-maximal inhibitory activity on pkC at periods, with proliferation measured at 65±70% of control 12 mm, while palm-car required 35 mm (data not shown). values in all cases. The cytotoxicity of the drugs was estimated Overall, these results suggest a greater efficiency of POC-16 as by counting viable and dead cells in each experiment. The an anti-proliferative agent and pkC inhibitor, with respect to inserts in Fig. 5 show the number of dead cells (expressed as other acyl-carnitines and alkyl-carnitines. percentage of total cells) counted in each 48 h experiment. To elucidate the inhibitory mechanisms of POC-16 and palm- Although cells exposed to the highest concentrations of car on pkC activity, the cofactor dependence of the enzyme was POC-16 underwent an increase in cell death, a strong decrease studied in the presence of each of these compounds by assaying in proliferation was nonetheless observed at concentrations that pkC under increasing concentrations of either Ca2+, PtdSer, DO caused little or no change in the number of dead cells (Fig. 5A, or ATP. No changes in the enzyme's requirement for Ca2+ or in insert). This was also the case for AM-44 using 2% fetal bovine its apparent Km for ATP were observed, and only a moderate effect on its affinity for DO was noted in the presence of POC-16 and palm-car (data not shown). In contrast, the requirement of pkC for PtdSer increased four to five times in the presence of 20 mm POC-16 (Fig. 4). Using the same concentration of palm-car, only a slight increase in PtdSer dependence was seen, and considerably higher concentrations were needed to observe a comparable effect (data not shown). The increase in POC-16 potency compared with palm-car led us to investigate whether other changes in the POC-16 molecular structure could improve its anti-proliferative and pkC inhibitory activity. During the synthesis of POC-16, two related compounds are generated by the ether bond formation (Experimental procedures; Fig. 1, structures III and IV). These compounds are difficult to separate but can be distinguished by Fig. 4. POC-16 augments the PtdSer-dependence of pkC. PkC activity NMR. The compound presenting a secondary alcohol structure was measured in the presence of cofactors using the mixed micellar assay in the hexadecyl moiety is the most abundant (70%), yet the [24,29] and increasing concentrations of PtdSer, both in the absence (X) inhibitory activity of POC-16 cannot be attributed solely to this and presence of POC-16: 5 mm (A) and 20 mm (K). This compound clearly molecule. With this in mind, a second ether derivative was increased the phospholipid requirement of the enzyme. q FEBS 1999 Conventional protein kinase C isoenzyme inhibitors (Eur. J. Biochem. 266) 859

Fig. 5. Inhibition of HL-60 cell proliferation by the alkyl-O-ether carnitines: POC-16 and AM-44. HL-60 cells were seeded (105 cell´mL21) in medium supplemented with either 2% (X) or 5% fetal bovine serum (W), and incubated for 48 h in the presence of POC-16 (A) and AM-44 (B) at the indicated concentrations. Both viable and dead cells (insets) were counted by Trypan Blue exclusion, and results were expressed as a percentage of control cells (incubated in the absence of drug). Values represent the mean ^ SD of 3±4 experiments carried out in duplicate. Insets: numbers of dead cells are given as percentage of total cells, grown in either 2% (black bars) or 5% fetal bovine serum (white bars), for 3±5 experiments carried out in duplicate. * Statistically different from the control (P . 0.05 or less, paired Student's t-test). Under similar experimental conditions in 5% fetal bovine serum, Ro 31-8220 0.5 mm and 1 mm inhibited HL-60 cell proliferation by 60 ^ 10% and 85 ^ 8%, respectively (not shown).

serum cells (Fig. 5B, insert), whereby the increase in cell death did not account for the sharp decrease in proliferation observed at 1 mm. No significant increase in the number of dead cells was noted for cells maintained in 5% fetal bovine serum even at the highest AM-44 concentration. Moreover, in reversion experiments, cells maintained in 5% fetal bovine serum for 48 h in the presence of AM-44 recuperated approximately three times as much as control cells, regardless of the drug concentration used, when placed in drug-free medium for a further 36 h. For 2% fetal bovine serum cells, low drug concentrations allowed the cells to recuperate fully (four times) but only 1.5-times recovery was observed at 0.5 mm. Under similar experimental conditions, in 5% fetal bovine serum, the bisindolylmaleimide-selective pkC inhibitor Ro 31-8220 at 0.5 mm and 1 mm inhibited HL-60 cell proliferation by 60%

Fig. 7. POC-16, AM-44 and Ro 31-8220 inhibit the thrombin-induced phosphorylation of a 47-kDa pkC-specific substrate protein in human platelets. Platelets were isolated as described in Experimental procedures, activated using 0.25 U´mL21 thrombin and incubated in the presence and absence of the pkC inhibitors at increasing concentrations. Samples were then submitted to SDS/PAGE and phosphorylation of the 47-kDa band was observed by autoradiography. (A) Autoradiogram showing the effect of POC-16 on a-thrombin induced phosphorylation. Lane 1, untreated cells; lane 2, a-thrombin alone; lanes 3±5, a-thrombin in the presence of 50, 100 Fig. 6. Inhibition of PMA-induced HL-60 cell differentiation by POC-16, and 200 mm POC-16. (B) Autoradiogram showing the effect of AM-44 AM-44 and Ro 318220. HL-60 cells were treated with 2 nm PMA for 48 h on a-thrombin induced-phosphorylation. Lane 1, untreated cells; lane 2, in the absence and presence of increasing concentrations of the pkC a-thrombin alone; lanes 3±5, a-thrombin in the presence of 10, 25 and inhibitors. Differentiation was assessed by measuring the number of 50 mm AM-44. (C) Autoradiogram showing the effect of Ro 31-8220 on adherent cells [25] after PMA treatment, which was 100% in the absence of a-thrombin-induced phosphorylation. Lane 1, untreated cells; lane 2, pkC inhibitors. (A) POC-16, (B) AM-44 (W) and Ro 31-8220 (X). Values a-thrombin alone; lanes 3±5, a-thrombin in the presence of 1, 5 and 10 mm represent the mean ^ SD of three measurements of one out of three Ro 31-8220. (D) The labeling of P47 was measured by densitometric independent experiments with similar results. Except for 5 mm POC-16, all analysis of the autoradiograms and expressed as a percentage of the values are statistically different (P , 0.01 or less, paired Student's t-test) phosphorylation in the presence of only a-thrombin (100%). (X) POC-16; from the control in the absence of inhibitor (100%). IC50, calculated by (B) AM-44; (W) Ro 31-8220. IC50 values, calculated by direct fitting to the direct fitting to the curves was 10.4 ^ 3.2, 0.19 ^ 0.07 and curves, were 87.4 ^ 33.6; 20.5 ^ 10.5 and 3.2 ^ 1.3 mm for POC-16, 0.07 ^ 0.01 mm for POC-16, AM-44 and Ro 31-8220, respectively. AM-44 and Ro 31-8220, respectively. 860 T. Garcia-Huidobro et al. (Eur. J. Biochem. 266) q FEBS 1999 and 85%, respectively, with , 10% increase in death cells In platelets, physiological ativators such as a-thrombin (not shown). Overall, these results indicated that the anti- initiate a series of early events including phosphoinositide proliferative effects of the ether drugs were due to a reversible turnover, Ca2+ mobilization and pkC activation [39±41]. In decrease in cell growth rather than a cytotoxic increase in cell this cellular system, activation of pkC by this agonist leads death. In order to further investigate the involvement of pkC to the marked phosphorylation of a protein of apparent Mr inhibition in the in situ effects of the carnitine ether derivatives, of 40 000±47 000 or plekstrin [42,43]. We therefore decided to two model systems were used: inhibition of HL-60 cell compare the effect of the synthesized pkC inhibitors with that differentiation induced by a phorbol ester, and inhibition of a of the bisindolylmaleimide-selective pkC inhibitor Ro 31-8220 physiological pkC susbtrate phosphorylation in intact platelets. [37,38] in this straightforward cellular model. Phorbol esters induce growth arrest and differentiation of Human platelets labeled with 32P were incubated for 2 min HL-60 cells to macrophage-like cells, and it has been shown with different concentrations of the synthesized inhibitors and that activation of the b-isozyme of pkC (pkC-b) is both RO 31-8425, followed by 1 min incubation with a-thrombin. necessary and sufficient for phorbol ester-induced differ- The pattern of protein phosphorylation was subsequently entiation [33,36] of these cells. In the presence of PMA, the analyzed by SDS/PAGE and autoradiography. Figure 7 shows HL-60 cells became adherent and presented a macrophage-like that POC-16, AM-44 and RO 31-8425 inhibited a-thrombin- morphology [34±36]. We have previously found that treatment stimulated P47 phosphorylation in a dose-dependent manner. of HL-60 cells in 5% fetal bovine serum with 2 nm PMA for The labeling of P47 was abolished completely by at 200, 50 and 48 h induced growth arrest and complete differentiation, as 10 mm of POC-16, AM-44 and RO 31-8425, respectively, with measured by the percentage of adherent cells, by the expression half-maximal inhibition of 87.4, 20.5 and 3.2 mm. Overall, of macrophage specific markers or morphologically [25]. experiments with these two in situ cell model systems strongly POC-16 25 mm or 2 mm AM-44 inhibited the PMA-induced suggest that pkC inhibition is involved in the anti-proliferative HL-60 cell differentiation almost completely, with IC50 values effects of the carnitine derivatives and further support the of 10.4 mm and 0.18 mm, respectively (Fig. 6A,B). The highest activity of AM-44 compared with POC-16. bisindolylmaleimide-selective pkC inhibitor Ro 31-8220 Increasing emphasis is currently being placed on elucidating [37,38] used as a positive control completely inhibited the individual roles of the various pkC isoforms. Hence, we PMA-induced HL-60 cell differentiation at 1 mm (IC50 = wanted to determine whether the ether drugs can discriminate 0.072 mm; Fig. 6B). between pkC isoenzymes in vitro. To do this, we individually

Fig. 8. Only the Ca2+-dependent isoforms of pkC are inhibited by POC-16 and AM-44. Recombinant human pkC isoforms were assayed under identical conditions (with the exception of varying cofactor requirements) using increasing concentrations of POC-16 (A) and AM-44 (B), as described in Experimental procedures. Results are expressed as percentage of control activity assayed without drugs, and values represent the mean ^ SE of 2±3 experiments carried out in triplicate. Both drugs clearly inhibited the cpkC isozymes [inhibition for 10 and 100 mm of drugs statistically different from the control (P , 0.05 or less, paired Student's t-test)]. Significant inhibition of pkC-z was also observed for 10 and 100 mm of the drugs (P , 0.05 or less, paired

Student's t-test), although the effect was much less than for the cpkCs. No effect was observed for pkC-d, and pkC-:. The calculated IC50 values for POC-16 were 36.8 ^ 19.8, 16.6 ^ 2.0, 33.5 ^ 5.6 and 19.3 ^ 1.3 mm for cpkC-a, cpkC-bI, cpkC-bII and cpkC-g, respectively, whereas those for AM-44 were 36.5 ^ 7.4, 24.8 ^ 11.2, 7.7 ^ 0.95 and 12.1 ^ 0.5 mm, respectively, for the same isozymes. q FEBS 1999 Conventional protein kinase C isoenzyme inhibitors (Eur. J. Biochem. 266) 861

that this compound activated, rather than inhibited, CaMKII substrate phosphorylation, whereas AM-44 had no effect (Fig. 9A). Neither compound was seen to affect CaMKII autophosphorylation activity (data not shown). CKI is also a Ca2+-dependent enzyme and shares many substrate targets with pkC. Activity assays, carried out using casein as a substrate and recombinant CKIa purified from yeast, showed no inhibitory effect on this kinase for either of the ether compounds (Fig. 9B). Similarly, POC-16 (up to 100 mm) did not affect the activity of the catalytic fragment of pkA. Therefore, the acyl-ether-carnitine derivatives were found to be not only pkC selective within the range of kinases tested, but also isozyme selective.

DISCUSSION In our current work we have synthesized two novel water- soluble ether lipids derived from palm-car, presenting anti- Fig. 9. POC-16 and AM-44 effect on CaMKII and CKI-a activity. proliferative properties and pkC isoenzyme-selective inhibitory CaMKII (A) and CKI-a (B) activities were assayed, using a crude rat brain action. The carnitine esters of nafenopin and ciprofibrate were extract and recombinant purified enzyme, respectively, in the presence of found to have no inhibitory effect on pkC, in the concentration either POC-16 (W) or AM-44 (X). Results are expressed as a percentage of range where inhibition is observed for palm-car, thus high- control activity assayed without drugs, and values represent the mean ^ SD lighting the importance of the aliphatic acyl chain. These of two experiments carried out in triplicate. A significant increase in results are in contrast to previous findings using the CoA-esters activity was observed at 100 mm of POC-16 for CaMKII (P , 0.05, paired Student's t-test). of these same compounds, which were seen to strongly magnify pkC activity in much the same way as palm-CoA [24,31,32]. Therefore, inhibition of pkC using carnitine derivatives seemed assayed seven pkC human recombinant isozymes under to have stronger structural requirements for an acyl chain than identical conditions (changing only reaction mixture cofactors) for its activation with CoA esters. The ether variant of palm-car in the presence of varying concentrations of either POC-16 or (POC-16), believed to be more stable than its parent molecule, AM-44. As shown in Fig. 8, the Ca2+-dependent cpkC proved to be three times more potent than palm-car toward isozymes were inhibited by these ether drugs, with complete purified pkC. inhibition at 100 mm in both cases. POC-16 appeared to have a In early work with POC-16, mice were inoculated with a small degree of selectivity toward cpkC-bI and cpkC-g, and suspension of tumor cells and treated 15 days after tumoral AM-44 was most selective toward cpkC-bII and cpkC-g. With induction with daily intratumoral doses of POC-16 (100 mg per AM-44, a large degree of variability was observed for kg body mass) for 5 days. The average diameter and volume of cpkC-bII, although 7.7 mm was the average IC50 value the tumors excised after 28 days were found to decrease determined from four experiments performed in triplicate. In significantly in treated mice without affecting other systemic the case of the npkCs (d and :), no significant inhibition was or local toxicity parameters, whereas palm-car had no effect observed even at the highest drug concentration, and for (S. Rojas and M. Bronfman, unpublished results). These apkC-z, only a minimal, although significant, decrease in preliminary results indicated that POC-16 possessed anti- activity (< 20%) was noted (Fig. 8). To test whether a higher proliferative activity and could potentially be used for drug threshold was required for these pkC isoforms, assays prolonged treatment in vivo. Moreover, the ether bond may were repeated using concentrations of 500 mm and 1 mm.At also be of particular relevance considering the finding that 500 mm, the activities of pkC-d and pkC-z were reduced malignant cells displayed greater sensitivity than normal cells slightly, whereas that of pkC-: was unaffected by either to the destructive effect of ether lysophospholipids [16±18]. compound (data not shown). In contrast, 1 mm POC-16 was This selective antineoplastic activity was found to correlate found to inhibit all three isoforms, although this effect could be with a lack of 1-O-alkyl cleavage enzyme in various neoplastic reversed completely by doubling the concentration of PtdSer tissues [44], which in part accounted for the relative inability of (data not shown). These results clearly indicated that the new tumor cells to metabolize these ether-linked compounds. It is ether compounds POC-16 and AM-44 were cpkC isoenzyme possible, therefore, that the selective accumulation of ether selective and pointed to a cpkC-mediated anti-proliferative lipids could interfere with vital lipid metabolism pathways [17] response in HL-60 cells. and may serve to accentuate the in vivo anti-proliferative However, the possibility that these compounds also affected activity of compounds such as POC-16. other kinases could not be discarded, particularly when The inhibitory efficiency of various POC-16 analogs (in vitro comparing in vitro (IC50, Fig. 8) and cell culture (LD50) and in cell cultures) was seen to decrease with a decrease in results. In order to study this possibility, inhibition assays were their alkyl-chain length, with hexadecyl (C16) forming the most carried out using three serine/threonine kinases known to be active substitution and hexyl (C6) the least. This is also the case functionally linked or related to pkC: CaMKII, CKI and the for acyl-CoAs, which activate pkC in a chain length-dependent catalytic fragment of pkA. CaMKII is regulated both by manner [31]. Moreover, the results of our experiments with calmodulin binding in the presence of Ca2+ and by autophos- CHO cells paralleled exactly those observed with pkC in vitro, phorylation, which makes the enzyme Ca2+ independent and indicating that inhibition of this enzyme was likely to underlie calmodulin independent. Both types of CaMKII activity were the anti-proliferative response. Furthermore, as already assayed using a crude rat brain extract and autocamtide-3, a described for the bisindolylmaleimide-selective pkC inhibitors highly specific CaMKII substrate. Assays with POC-16 showed [45], a concentration-dependent inhibition of a physiological 862 T. Garcia-Huidobro et al. (Eur. J. Biochem. 266) q FEBS 1999 pkC substrate phosphorylation in intact platelets was isozyme has been reported to constitute a strong growth observed for both alkyl-O-carnitine derivatives. Similarly, the promoter [56]. PMA-induced differentiation of HL-60 cells, which has been Structurally, the isoforms of pKC are defined according to shown to depend on pkC-b activation [33,36], was strongly four conserved domains, C1±C4 [2,12] that determine their inhibited by both compounds, as well as by the bisindolyl- capacity to bind diacylglycerol/phorbol esters/Zn2+ (C1), 2+ maleimide Ro 31-8220. Although lower IC50 values were found PtdSer/Ca (C2), ATP (C3) and execute phosphoryl transfer in this case than for P47 phosphorylation, the strength of the (C4). All isotypes share a common functional requirement inhibitors was the same and also correlates with their in vitro for PtdSer or anionic phospholipids. The classical isoforms effect on pkC. Overall, these experiments support the view that (cpkC-a, cpkC-bI, cpkC-bII and cpkC-g) also require the anti-proliferative action of the synthesized alkyl-O-carnitine synergistic activation by Ca2+ and diacylglycerol or phorbol ethers relies on cpkC inhibition. esters, whereas the novel isozymes npkC-d, npkC-:, npkC-u, The second ether derivative (AM-44) synthesized, in which npkC-h/L and npkC-m, respond to diacylglycerol and phorbol both a carboxyl and hydroxyl group of POC-16 were removed, esters, but lack a Ca2+ binding region within their C2 domain was more active than the parent compound. When tested in and are insensitive to this cation. Finally, the atypical isoforms HL-60 cells, this compound proved more effective than its apkC-z and apkC-l/i lack part of the C1 and C2 domains and counterpart, with LD50 values (for cells maintained in 2% are activated only by anionic phospholipids such as PtdSer, fetal bovine serum) in the range 0.1±0.4 mm, compared with phosphatidylinositol-3,4,5-triphosphate [PtdIns(3,4,5)P3] [57] 7±12 mm for POC-16. Similar conclusions can be reached from and ceramide [58]. the experiments on platelets and on PMA-induced HL-60 cell The finding that POC-16 acts competitively with respect to differentiation. Thus, the synthesis and evaluation of AM-44 PtdSer suggests that the ether compounds modulate pkC allowed us to determine that the carboxyl group of the carnitine activity by interacting directly with the C2 domain of the moiety is not a structural requisite and that the dominant isomer enzyme, whereas the majority of cpkC-selective inhibitors, of POC-16, in which the hydroxyl group is present as a such as GoÈ 6976 [59], UCN-01 [60], suramin [61] and rottlerin secondary alcohol and that most closely resembles AM-44, was [62], are directed at the catalytic domain of the enzyme. The C2 likely to constitute the most active species. domain, which is partially deleted in the npkC and apkC POC-16 and AM-44 were employed in cellular assays at isoforms, is known to contain the PtdSer/Ca2+-binding region concentrations below the normal cytosolic concentrations of and has recently been identified as a protein±protein interaction acyl-carnitines (25 mm) as measured in rat hearts [46]. domain [12]. It is therefore possible that the deletion that Furthermore, acyl-carnitines and the long chain alkyl-carnitines confers Ca2+-independence in these isoforms may also underlie will probably bind to proteins and membranes and may be their insensitivity to the ether drugs, which could interact at a actively taken up in the cells. It is probable that their site that overlaps both the PtdSer-binding and Ca2+-binding intracellular free concentration is much less than 25 mm. regions or that is intercalated between them. Indeed, phorbol Therefore, membrane perturbation, which may be caused by esters and diacylglycerol are known to bind the C1 region of these ether derivatives [47] was unlikely to sustain their action pkC and the presence of two binding sites within C1 with in cell culture. The reversible nature of the inhibition observed opposing affinities for these compounds has been demonstrated with POC-16 in HL-60 cells, the ability of both compounds to [63]. It is also interesting to note that pkC activation by palm- inhibit proliferation without causing cell death and the selective CoA results from a decrease in the PtdSer-dependence of the inhibition of Ca2+-dependent cPKC isoforms observed in vitro enzyme [24], thus opposing the effect of palm-car and further support this view. suggesting that both molecules are likely to be physiological The inhibitory activity of both POC-16 and AM-44 proved to modulators of pkC. be cpkC selective when tested in vitro using recombinant pkC To verify the kinase specificity of the ether drugs, these isozymes. The IC50 values for both compounds varied among compounds were tested on second messenger-dependent the cpkC isozymes in the range 10±30 mm, with the lowest kinases known to be important in mitogenic responses. No affinity observed for pkC-a and the highest for pkC-g. significant effect for either POC-16 or AM-16 was observed Interestingly, IC50 values were in close proximity to the basal with any of these kinases, with the exception of CaMKII, which cytosolic concentrations of acyl-carnitines measured in rat heart was activated to a small extent by POC-16. This may have (25 mm; [47]), indicating that a small increase in intracellular resulted from either a direct effect upon the enzyme or a levels of acyl-carnitines could have functional implications on secondary effect produced by pkC inhibition, considering that the activity of cpkC isoforms. Although the precise role of each coupling between CaMKII-mediated and pkC-mediated path- isotype remains unclear, there is evidence to suggest that pkC-b ways occurs in many cell types [3]. In fact, CaMKII has been and pkC-g are involved in proliferation control [47,48], reported to be both inhibited and unaffected by palm-car [22]. It whereas pkC-a is closely linked to cell differentiation [50,51] is also interesting to note that the ether lipid drugs did not affect and has been described as a growth inhibitor [11]. For instance, CKI, a kinase whose activity and association to integral pkC-bII was seen to translocate to the nucleus in response to membrane proteins is known to be negatively regulated by proliferative, but not differentiation, stimuli in rat fibroblasts PtdIns(3,4,5)P3 [64]. Similar phospholipids have been shown to [52] and has been purported to play a role in the suppression of be potent activators of pkC-d, pkC-: and pkC-h, and to a lesser apoptosis [53]. However, pkC-b has also been implicated in extent pkC-z [64]. Therefore, the regulation of CKI by lipid HL-60 differentiation [54,55], as well as in the induction of second messengers made this enzyme a potential target for the apoptosis [11], making it difficult to discern its cell-specific lipid inhibitors of pkC. functions. Nevertheless, it would seem that the simultaneous Given the experimental evidence provided here, we consider inhibition of all cpkC isoforms causes proliferation inhibition, both POC-16 and AM-44 to be excellent candidates for rational despite the fact that not all cpkC isoforms are required for this drug design. The selectivity of these compounds, their water function. For instance, the action of pkC-:, which was solubility and their apparent low toxicity make them especially unaffected by the ether inhibitors in vitro, was not sufficient suited for this purpose. The use of molecular dynamics, to maintain the proliferative cycle even though this mechanics and docking techniques to evaluate small structural q FEBS 1999 Conventional protein kinase C isoenzyme inhibitors (Eur. J. Biochem. 266) 863 modifications of these lead compounds, could generate an array alkyl-lysophosphoplipid derivatives and low-alkyl-cleavage enzyme of pharmacophores with greater affinities for the cpkC isotypes activities in rat brain tumor cells. Cancer Res. 43, 541±545. or even selective isozymes within this group. 20. Wise, B.C. & Kuo, J.F. (1983) Modes of inhibition by acylcarnitines, adryamycin and trifluoroperazine of cardiac phos- pholipid-sensitive calcium-dependent proten kinase. Biochem. ACKNOWLEDGMENTS Pharmacol. 32, 1259±1265. We are indebted to Dr Nibaldo C. Inestrosa and Alfonso GonzaÂlez for 21. Katoh, N., Wren, R.W., Wise, B.C., Shoji, M. & Kuo, J.F. (1981) Substrate proteins for calmodulin-sensitive and phospholipid- critical reading of the manuscript. This work was supported by the Chilean 2+ National Fund for Scientific and Technological Development (FONDECYT sensitive Ca -dependent protein kinases in heart, and inhibition grant 1971232 and 1990155) and by a Chilean Presidential Chair in of their phosphorylation by palmitoylcarnitine. Proc. Natl Acad. Sci. Sciences to M. Bronfman. USA 78, 4813±4817. 22. Nakaki, T., Mita, S., Yamamoto, S., Nakadate, T. & Kato, R. (1984) Inhibition by palmitoylcarnitine of adhesion and morphological REFERENCES changes in HL-60 cells induced by 12-O-tetradecanoylphorbol-13- acetate. Cancer Res. 44, 1908±1912. 1. Nishizuka, Y. (1996) Studies and perspectives of protein kinase C. 23. Nakadate, T., Yamamoto, S., Aizu, E. & Kato, R. (1986) Inhibition of Science 233, 305±312. 12-O-tetradecanoyl phorbol-13-acetate-induced tumor promotion and 2. Hug, H. & Sarre, T.F. (1993) Protein kinase C isoenzymes: divergence epidermal ornithine decarboxylase activity in mouse skin by in signal transduction? Biochem. J. 291, 329±343. palmitoylcarnitine. Cancer Res. 46, 1589±1593. 3. Nishizuka, Y. (1995) Protein kinase C and lipid signaling for sustained 24. Orellana, A., Hidalgo, P., Morales, M.N., Mezzano, D. & Bronfman, M. cellular responses. FASEB J. 9, 484±496. (1990) Palmitoyl-CoA and the acyl-CoA thioester of the 4. Lavin, M.F., Watters, D. & Song, Q. (1996) Role of protein kinase carcinigenic peroxisome-proliferator ciprofibrate potentiate diacyl- activity in apoptosis. Experientia 52, 979±994. glycerol-activated protein kinase C by decreasing the phosphatidyl- 5. Nishizuka, Y. (1998) The molecular heterogeneity of protein kinase C serine requirement of the enzyme. Eur. J. Biochem. 190, 57±61. and its implications for cellular regulation. Nature 334, 661±665. 25. Bronfman, M., Ponce, C., Rojas, S., Roth, A., Loyola, G., Vollrath, V. 6. Mochly-Rosen, D. (1995) Localization of protein kinases by anchoring & Chianale, J. (1998) Enhanced differentiation of HL-60 leukemia proteins: a theme in signal transduction. Science 268, 247±251. cells to macrophages induced by ciprofibrate. Eur. J. Cell. Biol. 77, 7. Hofmann, J. (1997) The potential for isoenzyme-selective modulation 214±219. of protein kinase C. FASEB J. 11, 649±669. 26. Laemmli, U.K. (1970) Cleavage of structural proteins during the 8. Decker, L.V. & Parker, P.J. (1994) Protein kinase C ± a question of assembly of the head of bacteriophage T4. Nature 227, 680±685. specificity. Trends Biol. Sci. 19, 73±77. 27. Cervenka, J. & Osmundsen, H. (1982) Synthesis of unsaturated 9. Chapline, C., Ramsay, K., Klauck, T. & Jaken, S. (1993) Interaction carnitine esters with N-acyl imidazoles. J. Lipid Res. 23, 1243±1246. cloning of protein kinase C substrates. J. Biol. Chem. 268, 28. Woodgett, J.R. & Hunter, T. (1987) Isolation and characterization of 6858±6861. two distinct forms of protein kinase C. J. Biol. Chem. 262, 10. Liao, L., Hyatt, S.L., Chapline, C. & Jaken, S. (1994) Protein kinase C 4836±4843. domains involved in interactions with other proteins. Biochemistry 29. Kikkawa, U., Minakuchi, R., Takai, Y. & Nishizuka, Y. (1983) 33, 1229±1233. Calcium-activated, phospholipid-dependent protein kinase (protein 11. Borner, C., Ueffing, M., Jaken, S., Parker, P. & Weinstein, J.B. (1995) kinase C) from rat brain. Methods Enzymol. 99, 288±298. Two closely related isoforms of protein kinase C produce reciprocal 30. Hannun, Y.A., Loomis, C.R. & Bell, R.M. (1985) Activation of protein effects on the growth of rat fibroblasts. Possible molecular kinase C by Triton X-100 mixed micelles containing diacylglycerol mechanisms. J. Biol. Chem. 270, 78±86. and phosphatidylserine. J. Biol. Chem. 260, 10039±10043. 12. Mellor, H. & Parker, P.J. (1998) The extended protein kinase C 31. Bronfman, M., Morales, M.N. & Orellana, A. (1988) Diacylglycerol superfamily. Biochem. J. 332, 281±292. activation of protein kinase C is modulated by long-chain acyl-CoA. 13. Tamaoki, T., Nomoto, H., Takahashi, I., Kato, Y., Morimoto, M. & Biochem. Biophys. Res. Commun. 152, 987±992. Tomita, F. (1986) Staurosporine, a potent inhibitor of phospholipid/ 32. Bronfman, M., Orellana, A., Morales, M.N., Bieri, F., Waechter, F., Ca++ dependent protein kinase. Biochem. Biophys. Res. Commun. StaÈubli, W. & Bentley, P. (1989) Potentiation of diacylglycerol- 135, 397±402. activated protein kinase C by acyl-coenzyme A thioesters of 14. RuÈegg, U.T. & Burgess, G.M. (1989) Staurosporine, K-252 and hypolipidaemic drugs. Biochem. Biophys. Res. Commun. 159, UCN-01: potent but nonspecific inhibitors of protein kinases. Trends 1026±1031. Pharmacol. Sci. 10, 218±220. 33. MacFarlane, D. & Manzel, L. (1994) Activation of the b-isoenzyme of 15. Meyer, T., Renenass, U., Fabbro, D., Alteri, E., RoÈsel, J., MuÈller, M., protein kinase C (PKCb) is necessary and sufficient for phorbol Caravatti, G. & Matter, A. (1989) A derivative of staurosporine (CGP ester-induced differentiation of HL-60 promyelocytes. J. Biol. Chem. 41,251) shows selectivity for protein kinase C inhibition and in vitro 269, 4327±4331. anti-proliferative as well as in vivo anti-tumor activity. Int. J. Cancer 34. Collins, S. (1987) The HL-60 promyelocytic leukemia cell line: 43, 851±856. proliferation, differentiation and cellular oncogene expression. Blood 16. Munder, P.G., Weltzein, H.U. & Modolell, M. (1976) Lysolecithin 70, 1233±1244. analogs, a new class of immunopotentiators. In Seventh International 35. Tanaka, H., Abe, T., Miyaura, C., Shiina, Y. & Suda, T. (1983) 1-alpha, Symposium on Immunopathology (Germany) (Miescher, P.A., ed.), 25-dihydroxyvitamin D3 induces differentiation of human promyelo- pp. 411±424. Schwabe Publishers, Basel, Switzerland. cytic leukemia cells (HL-60) into monocyte-macrophages, but not 17. Modolell, M., Andreesen, R., Pahlke, W., Brugger, U. & Munder, P.G. into granulocytes. Biochem. Biophys. Res. Commun. 117, 86±92. (1979). Disturbance of phospholipid metabolism during the selective 36. Tonetti, D.A., Henning-Chubb, C., Yamanishi, D.T. & Huberman, E. destruction of tumor cells induced by alkyl-lysophospholipids. (1994) Protein kinase C-b is required for macrophage differentiation Cancer Res. 39, 4681±4686. of human HL-60 leukemia cells. J. Biol. Chem. 269, 23230±23235. 18. Berdel, W.E., Fink, U., Egger, B., Reichert, A., Munder, P.G. & 37. Wilkinson, S.E., Parker, P.J. & Dixon, J.S. (1993) Isoenzyme Rastetter, J. (1981) Inhibition by alkyl-lysophospholipids of specificity of bisindolylmaleimides, selective inhibitors of protein tritiated thymidine uptake in cells of human malignant urologic kinase C. Biochem. J. 294, 335±337. tumors. J. Natl Cancer Inst. 66, 813±817. 38. Davis, P.D., Hill, C.H., Keech, E., Lawton, G., Nixon, J.S., Sedwick, 19. Berdel, W.E., Greiner, E., Fink, U., Stavrou, D., Reichert, A., A.D., Wadsworth, J., Wesmacott, D. & Wilkinson, S.E. (1989) Potent Rastetter, J., Hoffman, D.R. & Snyder, F. (1983) Cytotoxicity of selective inhibitors of protein kinase C. FEBS Lett. 259, 61±63. 864 T. Garcia-Huidobro et al. (Eur. J. Biochem. 266) q FEBS 1999

39. Kawahara, Y., Minakuchi, Y., Sano, K. & Nishizuka, Y. (1980) Phos- 53. Evans, C.A., Lord, J.M., Owen-Lybch, P.J., Johson, G., Dive, C. & pholipid turnover as a possible transmembrane signal for protein Whetton, A.D. (1995) Suppression of apoptosis by v-ABL protein phosphorylation during human platelet activation by thrombin. tyrosine kinase is associated with nuclear translocation and activation Biochim. Biophys. Res. Commun. 97, 309±317. of protein kinase C in an interleukin-3-dependent hemopoietic cell 40. Bell, R.L. & Majerus, P.W. (1980) Thrombin-induced hydrolysis of line. J. Cell. Sci. 108, 2591±2599. phosphatidylinositol in human platelets. J. Biol. Chem. 255, 54. MacFarlane, D.E. & Manzel, L. (1994) Activation of beta-isozyme of 1790±1792. protein kinase C (PKC beta) is necessary and sufficient for phorbol 41. Karniguaian, A., Grelac, F., Levy-Toledano, S., Legrand, Y.J. & Rendu, ester-induced differentiation of HL-60 promyelocytes. Studies with F. (1990) Collagen-induced platelet activation mainly involves the PKC. J. Biol. Chem. 269, 4327±4331. protein kinase C pathway. Biochem. J. 268, 325±331. 55. Nishikawa, M., Komada, F., Uemura, Y., Hidaka, H. & Shirakawa, S. 42. Imaoka, T., Lynham, J.A. & Haslam, R.J. (1983) Purification and (1990) Decreased expression of type II protein kinase C in HL-60 characterization of the 47,000-dalton protein phosphorylated during variant cells resistant to induction of cell differentiation by phorbol degranulation of human platelets. J. Biol. Chem. 258, 11404±11414. diester. Cancer Res. 50, 621±626. 43. Tyers, M., Haslam, T.M., Rachubinski, R.A. & Harley, C.B. (1989) 56. Mischak, H., Goodnight, J., Kolch, W., Martiny-Baron, G., Molecular analysis of pleckstrin: the major protein kinase C substrate Schaechtle, C., Kazanietz, M.G., Blumberg, P.M., Pierce, J.H. of platelets. J. Cell. Biochem. 40, 133±145. & Mushinski, J.F. (1993) Overexpression of protein kinase C- 44. Soodsma, J.F., Piantadosi, C. & Snyder, F. (1970) The biocleavage of delta and -epsilon in NIH 3T3 cells induces opposite effects on alkyl-glyceryl ethers in Morris hepatomas and other transplantable growth, morphology, anchorage dependence, and tumorigenicity. neoplasms. Cancer Res. 30, 309±311. J. Biol. Chem. 268, 6090±6096. 45. Toullec, D., Planett, P., Coste, H., Bellevergue, P., Grand-Perret, T., 57. Nakanishi, H., Brewer, K.A. & Exton, J.H. (1993) Activation of Ajakane, M., Baudet, V., Boissin, P., Boursier, E., Loriolle, F., the zeta isozyme of protein kinase C by phosphatidylinositol Duhamel, L., Charon, D. & Kirilovsky, J. (1991) The bisindolyl- 3,4,5-trisphosphate. J. Biol. Chem. 268, 13±16. maleimide GF 109203X is a potent and selective inhibitor of protein 58. Lozano, J., Berra, E., Municio, M.M., Diaz-Meco, M.T., Dominguez, kinase C. J. Biol. Chem. 266, 15771±15981. I., Sanz, L. & Moscat, J. (1994) Protein kinase C zeta isoform is 46. GonÄi, F.M., Requero, M.A. & Alonso, A. (1996) Palmitoylcarnitine, a critical for kappa B-dependent promoter activation by sphingo- surface-active metabolite. FEBS Lett. 390,1±5. myelinase. J. Biol. Chem. 269, 19200±19202. 47. Fñrgeman, N.J. & Knudsen, J. (1997) Role of long-chain fatty acyl- 59. Martiny-Baron, M., Kazanietz, M.G., Mischak, H., Blumberg, P.M., CoA esters in the regulation of metabolism and in cell signaling. Kochs, G., Hug, H., MarmeÂ, D. & SchaÈchtele, C. (1993) Selective Biochem. J. 323, 1±12. inhibition of protein kinase C isozymes by the indolocarbazole Go 48. Robinson, P.J. (1992) Potencies of protein kinase C inhibitors are 6976. J. Biol. Chem. 268, 9194±9197. dependent on the activators used to stimulate the enzyme. Biochem. 60. Mizuno, K., Noda, K., Ueda, Y., Hanaki, H., Saido, T.C., Ikuta, T., Pharmacol. 44, 1325±1334. Kuroki, T., Tamaoki, T., Hirai, S., Osada, S. & Ohno, S. (1995) 49. Housey, G.M., Johnson, M.D., Hsiao, W.L., O'Brien, C.A., Murphy, UCN-01, an anti-tumor drug, is a selective inhibitor of the J.P., Kirschmeirer, P. & Weinstein, I.B. (1988) Overproduction of conventional PKC subfamily. FEBS Lett. 359, 259±261. protein kinase C causes disordered growth control in rat fibroblasts. 61. Gshwendt, M., Muller, H.-J., Kielbassa, K., Zang, R., Kittstein, W., Cell 52, 343±354. Rincke, G. & Marks, F. (1994) Rottlerin, a novel protein kinase 50. Mischak, H., Pierce, J.H., Goodnight, J., Kazanietz, M.G., Blumberg, inhibitor. Biochem. Biophys. Res. Commun. 199, 93±98. P.M. & Mushinski, J.F. (1993) Phorbol ester-induced myeloid 62. Gshwendt, M., Kittstein, W. & Johannes, F.J. (1998) Differential effects differentiation is mediated by protein kinase C-alpha and -delta and of suramin on protein kinase C isoenzymes. A novel tool for not by protein kinase C-beta II-epsilon-zeta, and -eta. J. Biol. Chem. discriminating protein kinase C activities. FEBS Lett. 421, 165±168. 268, 20110±20115. 63. Bazenet, C.E., Brockman, J.L., Lewis, D., Chan, C. & Anderson, R.A. 51. Parrow, V., FaerstroÈm, G., Meyerson, G., Nanberg, E. & Pahlman, S. (1990) Erythroid membrane-bound protein kinase binds to a (1995) Protein kinase C-alpha and -epsilon are enriched in membrane component and is regulated by phosphatidylinositol growth cones of differentiating SH-SY5Y human neuroblastoma 4,5-bisphosphate. J. Biol. Chem. 265, 7369±7376. cells. J. Neurosci. Res. 41, 782±791. 64. Auger, K.R., Serunian, L.A., Soltoff, S.P., Libby, P. & Cantley, 52. Hocevar, B.A. & Fields, A.P. (1991) Selective translocation of beta II- L.C. (1989) PDGF-dependent tyrosine phosphorylation stimulates protein kinase C to the nucleus of human promyelocytic (HL60) production of novel polyphosphoinositides in intact cells. Cell leukemia cells. J. Biol. Chem. 266, 28±33. 57, 167±175.