Jacob Bell
New Member
Enzymatic synthesis of anandamide, an endogenous ligand for the cannabinoid receptor, by brain membranes
(arachidonic acid/ethanolamine/hippaampus)
WILLIAM A. DEVANE AND JULIUS AXELROD
Laboratory of Cell Biology, National Institute of Mental Health, Building 36, Room 3A-17, Bethesda, MD 20892
Contributed by Julius Axelrod, April 14, 1994
ABSTRACT Anandamide, an endogenous eicosanoid de-rivative (arachidonoylethanolamide), binds to the cannabinoid receptor, a member of the G protein-coupled superfamily. It also bits both adenylate cycase and N-type calcium channel opening. The enzymatic synthesis of anandamide in bovine brain tissue was examined by incubating . brain membranes with [1°C]ethanolamine and aradddonic add. Fol owing incu¬bation and extraction into toluene, a radioactive product was Identified which had the same Rf value as authentic ananda¬mide in several ttiin-layer chromatographic systems. When structurally similar fatty acid substrates were compered, ar¬adtidonic acid exhibited the lowest ECso and the highest activity for enzymatic formation of the corresponding ethanol-amides. The concentration-response curve of arachidonic acid exhibited a steep slope, and at higher concentrations arachi¬donate inhibited enzymatic activity. When brain homogenates were separated into subcellubu tractions by sucrose density gradient centrifugation, anandamide synthase activity was highest in tractions enriched in synaptic vesides, myelin, and anicrosomal and synaptosomal membranes. When several ar¬eas of brain were examined, anandamide synthase activity was found to be highest in the hippocampus, followed by the thalamus, cortex, and striatum, and lowest in the cerebellum, pons, and medulla. The ability of brain tissue to enzymatically synthesize anandamide and the existence of specific receptors for this eicosanold suggest the presence of anandamide¬containing (anandaergic) neurons.
The psychoactive component of marijuana, A9-tetrahydro-cannabinol (1), has been shown to bind to a specific receptor in brain (2). Rat (3) and human (4) cannabinoid receptors have been cloned and found to be members of the G protein-coupled superfamily of proteins. Cannabinoid receptors functionally couple to the inhibition of adenylate cyclase (3-7) and N-type calcium channels (8, 9). Discovery of a functional cannabinoid receptor in mammalian brain sug¬gested the existence of a natural endogenous ligand. Such a ligand has recently been isolated from porcine brain and identified as the ethanolamide of arachidonic acid (ananda¬mide) (10). Based on its ability to inhibit 1,4-dihydropyridine binding to the L-type calcium channel, anandamide was also isolated from bovine brain (11).
Anandamide exhibits many of the properties of 09-tetra-hydrocannabinol and other cannabinoid monists in pharma¬cological and behavioral assays. Anandamide bound with relatively high affinity to an brain membranes (K; = 40-50 nM) (10) and the cloned an (12) and human (13) cannabinoid receptors. It inhibited forskolin-stimulated cAMP production (ICso = 160-200 nM) (12, 13) and the N-type calcium channel (ICso - 20 nM) (14). Anandamide also produced a non¬receptor-mediated liberation of arachidonic acid and mobilization of cytosolic calcium (13) similar to that exhibited by cannabinoid drugs (7). It produced many of the behavioral and physiological responses of cannabinoids, such as a decrease in locomotor activity (15, 16) and a reduction in body temperature (15, 16).
In this report we describe the enzymatic synthesis of anandamide in brain, its kinetics, and its localization.
MATERIALS AND METHODS
Materials. Arachidonic acid and other fatty acids were purchased from Nu-Chek Prep (Elysian, MN). Unlabeled fatty acid ethanolamides were synthesized as described (13). ['4C]Ethanolamine hydrochloride (55 mCi/mmol; 1 Ci = 37 GBq) and [3H]arachidonic acid (200 Ci/mmol) were pur¬chased from American Radiolabeled Chemicals (St. Louis). [3H]Anandamide was a gift from David Ahem (New England Nuclear). [14C]Anandamide was a gift from Christian Felder (National Institute of Mental Health, Bethesda, MD). Bovine brains from 18- to 24month-old Black Angus steers were obtained on ice from the slaughterhouse. Silica (silica G/UV, 250 µm) and reversed-phase (KC18F, 200 µm) thin-layer chromatography (TLC) plates were obtained from Whatman. All organic solvents were reagent grade. All other chemicals were from Sigma.
Membrane Preparations. Bovine brains were dissected and maintained at -70°C. P2 membrane fractions were prepared by homogenizing 2 g of brain tissue in 30 ml of 50 mM Tris-HCI, pH 9/1 mM EDTA. The homogenate was centri¬fuged for 10 min at 1800 rpm in a Beckman JS 4.2 swinging-bucket rotor. The supernatant was removed and centrifuged for 20 min at 18,000 rpm in a Sorvall SS-34 rotor. The P2 pellet was resuspended in 50 mM Tris-HCI (pH 9) and stored at -70°C until used.
Washed P2 membranes were prepared as above with the final pellet suspended in water. After 10 min on ice, the homogenate was centrifuged for 20 min at 18,000 rpm in a Sorvall SS-34 rotor. This step was repeated, and the pellet was suspended in 50 mM Tris-HCl (pH 9) and stored at -70°C until used. Subcellular fractionation by discontinous sucrose density gradient centrifugation was performed by the proto¬col of Whittaker et al. (17). To remove the sucrose, the various layers were suspended in 20 volumes of 50 mM Tris-HCI (pH 9) and centrifuged at 35,000 rpm for 60 min in a Beckman type 35 rotor. The pellets were resuspended in 50 mM Tris-HCl (pH 9) and stored at -70°C until used. Protein measurements were determined with the Pierce bicichommc acid (BCA) kit.
Measurement of the Enzymatic Formation of Anaadamide. A typical reaction mixture for measurement of the enzymatic formation of anandamide contained =70 µg of protein from brain tissue, [14C]ethanolamine hydrochloride (550 nCi), 20 mM ethanolamine, 100 µM arachidonic acid, and 125 mM Tris-HC1(pH 9) in a total volume of 2001d in a 13 mm x 100 mm glass tube. After 20-40 min of incubation at 37°C, 400 A of 0.15 M HCl was added, and the mixture was extracted with 3 ml of toluene. After vortexing and centrifugation for 5 min at 900 x g to facilitate phase separation, 2 ml of the toluene phase was transferred to a 12 mm x 75 mm glass tube. The solvent was removed under reduced pressure in a rotary evaporator. Twenty microliters of toluene was added to the dried residue and the product, along with an authentic ['4C]anandamide standard, was spotted onto a silica-gel TLC plate (10 x 20 cm) which was developed to 7 cm from the origin with solvent (chloroform/petroleum ether/methanol, 80
10 by volume). Under these conditions, anandamide had an Rf value ranging from 0.5 to 0.65, which varied with room temperature and tank saturation. For identification purposes, reverse-phase chromatography was also per-formed; KC18F TLC plates were developed 7 cm with solvent (acetonitrile/water, 92:8), and anandamide exhibited an Rf value of 0.4.
To quantify the amount of [14C]anandamide formed, stan-dards of several concentrations of [14C]ethanolamine were spotted on the plate after chromatography was complete. The radioactivity on the TLC plates was quantified by phosphor-imaging with the FUJIX BAS 2000 bioimaging analyzer (Fuji), which provides luminescence values proportional to the radiation dose (dpm) present. The data reported were obtained as follows. A curve of the luminescence values of the [14C]ethanolamine standards versus their dpm was used to transform the sample luminescence value into dpm. The sample dpm value was converted to nmol of anandamide formed by using the final specific activity of [14C]ethanol¬amine in the assay (usually 305 dpm/nmol). The overall recovery, through chromatography, of a test sample of [3H]-anandamide was about 60%; data were not corrected for recovery. Linear results were obtained with 25-150 µg of protein per tube. Experiments were performed in duplicate. As examples of the intraassay variability, the variance as percent of the mean of the duplicates in Fig. 2 was 5.3 ± 4.6% and in Fig. 3 was 7.3 ± 5.1% (mean ± SD).
When [3H]arachidonic acid was substituted for [14C]etha¬nolamine as the labeled compound incorporated into anan¬damide, P2 membranes (70 µg of protein) were incubated with [3H]arachidonic acid (1 MCi), 100 µM arachidonic acid, and 20 mM ethanolamine for 40 NaOH (400 A, 0.36 M) was added to the mixture instead of HC1 before toluene extrac¬tion. After chromatography, the radioactive products were visualized by autoradiography.
RESULTS
Enzymatic Formation of [14C]Anandamide by Brain Mem¬branes. To determine synthesis of anandamide, bovine cor¬tical membranes (P2 fraction) were incubated with arachi¬donic acid and [14C]ethanolamine at 37°C for 40 After incubation the reaction mixture was extracted with toluene and the solvent was evaporated. To determine the identity of the apparent [14C]anandamide formed enzymatically, the residue was subjected to TLC. A radioactive product having the same Rf value as authentic anandamide was detected after both normal-phase (silica) and reverse-phase (Co) chroma¬tography (Fig. 1 A and B). When the radioactive product identified after normal-phase chromatography was extracted and rechromatographed under reverse-phase conditions, the radioactivity migrated as one spot having the same Rf value as authentic anandamide (data not shown). When brain membranes were preheated at 100°C for 5 min and then incubated with substrates, extracted, and subjected to TLC, there was no radioactive product formed having the same Rf value as authentic anandamide (Fig. 1B). Further evidence for the enzymatic formation of anandamide was obtained when [3H]arachidonic acid was substituted for [14C]ethanol¬amine in the reaction mixture. The reaction mixture was extracted into toluene and subjected to silica chromatogra¬phy. The 3H-labeled product formed had the same Rfvalue as anandamide in both normal-phase (Fig. 1C) and reverse-phase (data not shown) chromatography.
PMSF, a serine protease inhibitor, has been shown to inhibit the enzymatic cleavage of anandamide which yields arachidonic acid and ethanolamine (18, 19). Including PMSF in our incubation mixture, however, greatly inhibited the production of anandamide (Fig. IA). The inclusion of PMSF also resulted in the formation of radioactive products which had different Rf values than synthetic anandamide (Fig. 1A).
The enzymatic formation of anandamide was linear for 50 min and then began to level off (Fig. 2). The pH optimum for enzymatic activity was between 9 and 10 (data not shown). Taken together, these observations indicate that incubation of brain membranes with ethanolamine and arachidonic acid results in the enzymatic formation of anandamide.
Substrate Specificity of Fatty Acid Ethanolamide Formation. The ability of a number of saturated and unsaturated fatty acids to form amides of ethanolamine was investigated. Enzyme activity was determined by incubation of a range of concentrations of fatty acids with [14C]ethanolamine and hippocampal membranes, followed by extraction into toluene and silica chromatography. Radiolabeled enzymatic products had the same Rf values as their respective synthetic fatty acid ethanolamides. The enzymatic activity of fatty acid ethanol-amide formation was quantitated with a phosphor-imaging plate scanner. Of all the fatty acids examined, arachidonic acid (C20:4) was the best substrate, having both the lowest EC50 (30-50 ) and the highest V. over the concentration range tested (Fig. 3). The closely related unsaturated fatty acids dihomo-y-linolenic acid (C20:3; cis-8,11,14-eicosa¬trienoic acid) and cis-11,14-eicosadienoic acid (C20:2) were the next best substrates, followed by palmitic acid (C16:0). The poorest fatty acid substrates were adrenic acid (C22:4; cis-7,10,13,16-docosatetraenoic) and docosahexaenoic acid (C22:6). Maximal activity was achieved with 100-300 µM arachidonic acid; with higher concentrations there was a marked inhibition in enzyme activity. None of the other substrates examined showed this inhibitory behavior. Arachidonyl-CoA was also examined as a substrate and was found to be almost 10-fold less potent than arachidonic acid (data not shown).
When 70 µg of hippocampal P2 membranes was incubated with 150 µM arachidonic acid and various concentrations of unlabeled ethanolamine for 20 min, the apparent K. for ethanolamine was 27 ± 4 mM and the apparent Vmax was 2230 ± 260 pmol/min per mg of protein (mean ± SD, determined by Lineweaver-Burk analysis).
Localization of Anandamide Synthase Activity in Brain Areas. To compare the enzymatic activity in various brain areas, washed P2 fractions of brain membranes were made from various regions of bovine brain. Tissue preparations from each brain area were incubated with [14C]ethanolamine and various concentrations of arachidonic acid (100-2400 AM), and [14C]anandamide synthesis was measured. A con¬centration of 100 µM arachidonic acid produced peak enzy¬matic activity in all brain areas. The amounts of anandamide produced enzymatically differed in the various brain regions examined (Fig. 4). The hippocampus had the highest enzy¬matic activity. Thalamus, striatum, and frontal cortex had a similar enzymatic activity, whereas pons, cerebellum, and medulla had the lowest activity.
To examine the subcellular distribution of the enzymatic activity, cortical P2 membranes were fractionated by density gradient centrifugation (step gradient of 0.4, 0.6, 0.8, and 1.2 M sucrose). Although enzymatic activity was observed in all fractions, enzymatic activity was highest in fractions of lower density than 0.8 M sucrose (370-530 pmol/min per mg of protein, for incubation with 1 mM arachidonic acid and 20 mM ethanolamine). These fractions contain synaptic vesi¬cles, microsomes, myelin fragments, and synaptosome ghosts (17). Lowest activity was observed in the pellet of 1.2 M sucrose (120 pmol/min per mg of protein), a fraction enriched in mitochondria (17). The cytosolic fraction exhib¬ited negligible activity.
DISCUSSION
The brain was shown to enzymatically synthesize ananda¬mide, a cannabimimetic eicosanoid, using arachidonic acid and ethanolamine as substrates. When incubated with either [14C]ethanolamine or [3H]arachidonic acid, bovine brain membranes synthesized a radioactive compound which had the same Rfvalue as anandamide following toluene extraction and chromatography. A previous report had described the enzymatic synthesis of palmitoylethanolamide from palmitic acid and ethanolamine in various tissues (20). The protocol described for the measurement of palmitoylethanolamide was not adequate to specifically measure the enzymatic synthesis of anandamide in brain, because the incorporation of [14C]ethanolamine into toluene-extractable products other than anandamide varied with both subcellular fraction and brain structure investigated, as well as with the length of incubation and the concentration of the substrates. Deter-mining whether the enzyme capable of synthesizing ananda¬mide in the brain is the same as that which synthesizes palmitoylethanolamide in other tissues must await their pu-rification.
The conjugation of ethanolamine with a fatty acid to form an amide may require a source of energy. The impurity of our preparation did not permit an extensive investigation into the possible cofactors involved. Arachidonyl-CoA was a poor substrate, and the addition of other potential energy-generating compounds such as acetyl-CoA or ATP had little effect on enzymatic activity. Enzymatic activity was retained after the membranes were washed several times in water, indicating that water-soluble cofactors are probably unnec¬essary. There are many competing reactions for the incor¬poration and metabolism of both ethanolamine and arachi¬donic acid in brain membrane fractions making further char¬acterization of the amide-forming enzyme difficult without greater purification.
Several unsaturated fatty acids have been shown to be substrates for the fatty acid ethanolamide-forming enzyme. Of the substrates examined, arachidonic acid had the highest
and the lowest ECso. In contrast to other fatty acid substrates, high concentrations of arachidonic acid were inhibitory. This inhibitory phenomena was found in the various subcellular fractions and in all areas of the brain examined. This type of kinetics suggests that anandamide synthase is regulated by the substrate arachidonic acid. It was previously found that the ethanolamides of dihomo-y-linolenic acid (C20:3) and adrenic acid (C22:4) had a potency similar to anandamide with respect to their ability to bind to the cloned human brain cannabinoid receptor and to inhibit cAMP production (13). However, dihomo-y-linolenyl etha-nolamide was less potent than anandamide at inhibiting the N-type calcium channel (13). The ethanolamides of dihomo-y-linolenic and adrenic acids have also recently been detected in porcine brain (21). These fatty acids, however, appear to be poorer substrates than arachidonic acid for the enzymatic synthesis of their respective ethanolamides.
The substrates of anandamide synthase, arachidonic acid and ethanolamine, can be generated by the activation of phospholipase A2 and phospholipase D, respectively, sug-gesting that they may arise from the hydrolysis of phospho-lipids. Preliminary evidence has indicated that the synthesis of anandamide in cultured cerebellar granule cells is evoked by a receptor-mediated influx of calcium, the activation of phospholipase A2, and the liberation of arachidonic acid
(C. C. Felder, E. M. Briley, G. -M. Yan, and J.A., unpub-lished data).
Anandamide synthase activity was present in all areas of bovine brain examined. The highest enzymatic activity was found in the hippocampus, and the lowest levels were found in the cerebellum and medulla. The cannabinoid receptor is also widely distributed with high levels in the striatum, cerebellum, and hippocampus, but the relative levels of cannabinoid receptor vary with species (22). A more detailed study of the distribution of anandamide synthesis in various areas of the brain is needed.
Anandamide, a highly lipophilic compound, could cross the synapse and be taken up presynaptically, thus serving as a retrograde transmitter. Lipids such as arachidonic acid (23) and platelet-activating factor (24) have been shown to be retrograde transmitters involved in regulating long-term po-tentiation. It is possible that anandamide may play a similar role. That the enzymatic activity is high in the hippocampus, a brain area considered to be important in the regulation of long-term potentiation, supports this suggestion.
Behavior studies in mice have shown that anandamide can affect locomotor activity (15, 16). The presence of ananda-mide synthase activity in striatum and cerebellum, areas of the brain involved in movement, suggests that anandaergic neurons may also be involved in the control of movement. Since cannabinoid drugs produce euphoria and paranoia, as well as alter the perception of time (25), anandaergic neurons may also be involved in these processes.
We thank Drs. D. C. Button, C. C. Felder, and J. K. Northup for helpful discussions and for critical reading of the manuscript.
1. Gaom, Y. & Mechoulam, R. (1964) J. Am. Chem. Soc. 86, 1646-1647.
2. Devane, W. A., Dysarz, F. A., III, Johnson, M. R., Melvin,
L. S. & Howlett, A. C. (1988) Mol. Pharmacol. 34, 605-613.
3. Matsuda, L. A., Lolait, S. J., Brownstein, M., Young, A. &
Bonner, T. I. (1990) Nature (London) 346, 561-564.
4. Gerard, C. M., Mollereau, C., Vassart, G. & Parmentier, M.
(1991) Biochem. J. 279, 129-134.
5. Howlett, A. C. & Fleming, R. M. (1984) Mol. Pharmacol. 27, 429-436.
6. Howlett, A. C., Bidaut-Russell, M., Devane, W. A., Melvin, L. S., Johnson, M. R. & Herkenham, M. (1990) Trends Phar¬macol. Sci. 13, 420-423.
7. Felder, C. C., Veluz, J. S., Williams, H. L., Briley, E. M. & Matsuda, L. A. (1992) Mol. Pharmacol. 42, 838-845.
8. Mackie, K. & Halle, B. (1992) Proc. Natl. Acad. Sci. USA 89, 3825-3829.
9. Caulfield, M. P. & Brown, D. A. (1992) Br. J. Pharmacol. 106, 231-232.
10. Devane, W. A., Hanus, L., Breuer, A., Pertwee, R. G.,
Stevenson, L. A., Griffin, G., Gibson, D., Mandelbaum, A.,
Etinger, A. & Mechoulam, R. (1992) Science 258, 1946-1949.
11. Johnson, D. E., Heald, S. L., Dally, R. D. & Janis, R. A.
(1993) Prostaglandins Leukotrienes Essent. Fatty Acids 48,
429-437.
12. Vogel, Z., Barg, J., Levy, R., Saya, D., Heldman, E. & Mechoulam, R. (1993) J. Neurochem. 61, 352-355.
13. Felder, C. C., Briley, E. M., Axelrod, J., Simpson, J. T., Mackie, K. & Devane, W. A. (1993) Proc. Natl. Acad. Sci. USA 90, 7656-7660.
14. Mackie, K., Devane, W. A. & Hille, B. (1993) Mol. Pharmacol. 44,498-503.
15. Fride, E. & Mechoulam, R. (1993) Eur. J. Pharmacol. 231, 313-314.
16. Crawley, J. N., Corwin, R. L., Robinson, J. K., Felder, C. C., Devane, W. A. & Axelrod, J. (1993) Pharmacol. Biochem. Behav. 46, 967-972.
17. Whittaker, V. P., Michaelson, I. A. & Kirkland, R. J. A. (1964) Biochem. J. 90, 293-303.
18. Deutsch, D. G. & Chin, S. A. (1993) Biochem. Pharmacol. 46, 791-7%.
19. Childers, S. R., Sexton, T. & Roy, M. B. (1994) Biochem. Pharmacol. 47, 711-715.
20. Bachur, N. R. & Udenfriend, S. (1966) J. Biol. Chem. 241, 1308-1313.
21. Hanus, L., Gopher, A., Almog, S. & Mechoulam, R. (1993) J. Med. Chem. 36, 3032-3034.
22. Herkenham, M., Lynn, A. B., Little, M. D., Johnson, M. R., Melvin, L. S., de Costa, B. R. & Rice, K. C. (1990) Proc. Natl. Acad. Sci. USA 87, 1932-1936.
23. Williams, J. H., Errington, M. L., Lynch, M. A. & Bliss, T. V. P. (1991) Nature (London) 341, 739-742.
24. to, K., Clark, G. D., Bazars, N. G. & Zorumski, C. F. (1994) Nature (London) 367, 175-179.
25. Hollister, L. E. (1986) Pharmacol. Rev. 38, 1-20.
Source: Enzymatic synthesis of anandamide, an endogenous ligand for the cannabinoid receptor, by brain membranes
(arachidonic acid/ethanolamine/hippaampus)
WILLIAM A. DEVANE AND JULIUS AXELROD
Laboratory of Cell Biology, National Institute of Mental Health, Building 36, Room 3A-17, Bethesda, MD 20892
Contributed by Julius Axelrod, April 14, 1994
ABSTRACT Anandamide, an endogenous eicosanoid de-rivative (arachidonoylethanolamide), binds to the cannabinoid receptor, a member of the G protein-coupled superfamily. It also bits both adenylate cycase and N-type calcium channel opening. The enzymatic synthesis of anandamide in bovine brain tissue was examined by incubating . brain membranes with [1°C]ethanolamine and aradddonic add. Fol owing incu¬bation and extraction into toluene, a radioactive product was Identified which had the same Rf value as authentic ananda¬mide in several ttiin-layer chromatographic systems. When structurally similar fatty acid substrates were compered, ar¬adtidonic acid exhibited the lowest ECso and the highest activity for enzymatic formation of the corresponding ethanol-amides. The concentration-response curve of arachidonic acid exhibited a steep slope, and at higher concentrations arachi¬donate inhibited enzymatic activity. When brain homogenates were separated into subcellubu tractions by sucrose density gradient centrifugation, anandamide synthase activity was highest in tractions enriched in synaptic vesides, myelin, and anicrosomal and synaptosomal membranes. When several ar¬eas of brain were examined, anandamide synthase activity was found to be highest in the hippocampus, followed by the thalamus, cortex, and striatum, and lowest in the cerebellum, pons, and medulla. The ability of brain tissue to enzymatically synthesize anandamide and the existence of specific receptors for this eicosanold suggest the presence of anandamide¬containing (anandaergic) neurons.
The psychoactive component of marijuana, A9-tetrahydro-cannabinol (1), has been shown to bind to a specific receptor in brain (2). Rat (3) and human (4) cannabinoid receptors have been cloned and found to be members of the G protein-coupled superfamily of proteins. Cannabinoid receptors functionally couple to the inhibition of adenylate cyclase (3-7) and N-type calcium channels (8, 9). Discovery of a functional cannabinoid receptor in mammalian brain sug¬gested the existence of a natural endogenous ligand. Such a ligand has recently been isolated from porcine brain and identified as the ethanolamide of arachidonic acid (ananda¬mide) (10). Based on its ability to inhibit 1,4-dihydropyridine binding to the L-type calcium channel, anandamide was also isolated from bovine brain (11).
Anandamide exhibits many of the properties of 09-tetra-hydrocannabinol and other cannabinoid monists in pharma¬cological and behavioral assays. Anandamide bound with relatively high affinity to an brain membranes (K; = 40-50 nM) (10) and the cloned an (12) and human (13) cannabinoid receptors. It inhibited forskolin-stimulated cAMP production (ICso = 160-200 nM) (12, 13) and the N-type calcium channel (ICso - 20 nM) (14). Anandamide also produced a non¬receptor-mediated liberation of arachidonic acid and mobilization of cytosolic calcium (13) similar to that exhibited by cannabinoid drugs (7). It produced many of the behavioral and physiological responses of cannabinoids, such as a decrease in locomotor activity (15, 16) and a reduction in body temperature (15, 16).
In this report we describe the enzymatic synthesis of anandamide in brain, its kinetics, and its localization.
MATERIALS AND METHODS
Materials. Arachidonic acid and other fatty acids were purchased from Nu-Chek Prep (Elysian, MN). Unlabeled fatty acid ethanolamides were synthesized as described (13). ['4C]Ethanolamine hydrochloride (55 mCi/mmol; 1 Ci = 37 GBq) and [3H]arachidonic acid (200 Ci/mmol) were pur¬chased from American Radiolabeled Chemicals (St. Louis). [3H]Anandamide was a gift from David Ahem (New England Nuclear). [14C]Anandamide was a gift from Christian Felder (National Institute of Mental Health, Bethesda, MD). Bovine brains from 18- to 24month-old Black Angus steers were obtained on ice from the slaughterhouse. Silica (silica G/UV, 250 µm) and reversed-phase (KC18F, 200 µm) thin-layer chromatography (TLC) plates were obtained from Whatman. All organic solvents were reagent grade. All other chemicals were from Sigma.
Membrane Preparations. Bovine brains were dissected and maintained at -70°C. P2 membrane fractions were prepared by homogenizing 2 g of brain tissue in 30 ml of 50 mM Tris-HCI, pH 9/1 mM EDTA. The homogenate was centri¬fuged for 10 min at 1800 rpm in a Beckman JS 4.2 swinging-bucket rotor. The supernatant was removed and centrifuged for 20 min at 18,000 rpm in a Sorvall SS-34 rotor. The P2 pellet was resuspended in 50 mM Tris-HCI (pH 9) and stored at -70°C until used.
Washed P2 membranes were prepared as above with the final pellet suspended in water. After 10 min on ice, the homogenate was centrifuged for 20 min at 18,000 rpm in a Sorvall SS-34 rotor. This step was repeated, and the pellet was suspended in 50 mM Tris-HCl (pH 9) and stored at -70°C until used. Subcellular fractionation by discontinous sucrose density gradient centrifugation was performed by the proto¬col of Whittaker et al. (17). To remove the sucrose, the various layers were suspended in 20 volumes of 50 mM Tris-HCI (pH 9) and centrifuged at 35,000 rpm for 60 min in a Beckman type 35 rotor. The pellets were resuspended in 50 mM Tris-HCl (pH 9) and stored at -70°C until used. Protein measurements were determined with the Pierce bicichommc acid (BCA) kit.
Measurement of the Enzymatic Formation of Anaadamide. A typical reaction mixture for measurement of the enzymatic formation of anandamide contained =70 µg of protein from brain tissue, [14C]ethanolamine hydrochloride (550 nCi), 20 mM ethanolamine, 100 µM arachidonic acid, and 125 mM Tris-HC1(pH 9) in a total volume of 2001d in a 13 mm x 100 mm glass tube. After 20-40 min of incubation at 37°C, 400 A of 0.15 M HCl was added, and the mixture was extracted with 3 ml of toluene. After vortexing and centrifugation for 5 min at 900 x g to facilitate phase separation, 2 ml of the toluene phase was transferred to a 12 mm x 75 mm glass tube. The solvent was removed under reduced pressure in a rotary evaporator. Twenty microliters of toluene was added to the dried residue and the product, along with an authentic ['4C]anandamide standard, was spotted onto a silica-gel TLC plate (10 x 20 cm) which was developed to 7 cm from the origin with solvent (chloroform/petroleum ether/methanol, 80
10 by volume). Under these conditions, anandamide had an Rf value ranging from 0.5 to 0.65, which varied with room temperature and tank saturation. For identification purposes, reverse-phase chromatography was also per-formed; KC18F TLC plates were developed 7 cm with solvent (acetonitrile/water, 92:8), and anandamide exhibited an Rf value of 0.4.To quantify the amount of [14C]anandamide formed, stan-dards of several concentrations of [14C]ethanolamine were spotted on the plate after chromatography was complete. The radioactivity on the TLC plates was quantified by phosphor-imaging with the FUJIX BAS 2000 bioimaging analyzer (Fuji), which provides luminescence values proportional to the radiation dose (dpm) present. The data reported were obtained as follows. A curve of the luminescence values of the [14C]ethanolamine standards versus their dpm was used to transform the sample luminescence value into dpm. The sample dpm value was converted to nmol of anandamide formed by using the final specific activity of [14C]ethanol¬amine in the assay (usually 305 dpm/nmol). The overall recovery, through chromatography, of a test sample of [3H]-anandamide was about 60%; data were not corrected for recovery. Linear results were obtained with 25-150 µg of protein per tube. Experiments were performed in duplicate. As examples of the intraassay variability, the variance as percent of the mean of the duplicates in Fig. 2 was 5.3 ± 4.6% and in Fig. 3 was 7.3 ± 5.1% (mean ± SD).
When [3H]arachidonic acid was substituted for [14C]etha¬nolamine as the labeled compound incorporated into anan¬damide, P2 membranes (70 µg of protein) were incubated with [3H]arachidonic acid (1 MCi), 100 µM arachidonic acid, and 20 mM ethanolamine for 40 NaOH (400 A, 0.36 M) was added to the mixture instead of HC1 before toluene extrac¬tion. After chromatography, the radioactive products were visualized by autoradiography.
RESULTS
Enzymatic Formation of [14C]Anandamide by Brain Mem¬branes. To determine synthesis of anandamide, bovine cor¬tical membranes (P2 fraction) were incubated with arachi¬donic acid and [14C]ethanolamine at 37°C for 40 After incubation the reaction mixture was extracted with toluene and the solvent was evaporated. To determine the identity of the apparent [14C]anandamide formed enzymatically, the residue was subjected to TLC. A radioactive product having the same Rf value as authentic anandamide was detected after both normal-phase (silica) and reverse-phase (Co) chroma¬tography (Fig. 1 A and B). When the radioactive product identified after normal-phase chromatography was extracted and rechromatographed under reverse-phase conditions, the radioactivity migrated as one spot having the same Rf value as authentic anandamide (data not shown). When brain membranes were preheated at 100°C for 5 min and then incubated with substrates, extracted, and subjected to TLC, there was no radioactive product formed having the same Rf value as authentic anandamide (Fig. 1B). Further evidence for the enzymatic formation of anandamide was obtained when [3H]arachidonic acid was substituted for [14C]ethanol¬amine in the reaction mixture. The reaction mixture was extracted into toluene and subjected to silica chromatogra¬phy. The 3H-labeled product formed had the same Rfvalue as anandamide in both normal-phase (Fig. 1C) and reverse-phase (data not shown) chromatography.
PMSF, a serine protease inhibitor, has been shown to inhibit the enzymatic cleavage of anandamide which yields arachidonic acid and ethanolamine (18, 19). Including PMSF in our incubation mixture, however, greatly inhibited the production of anandamide (Fig. IA). The inclusion of PMSF also resulted in the formation of radioactive products which had different Rf values than synthetic anandamide (Fig. 1A).
The enzymatic formation of anandamide was linear for 50 min and then began to level off (Fig. 2). The pH optimum for enzymatic activity was between 9 and 10 (data not shown). Taken together, these observations indicate that incubation of brain membranes with ethanolamine and arachidonic acid results in the enzymatic formation of anandamide.
Substrate Specificity of Fatty Acid Ethanolamide Formation. The ability of a number of saturated and unsaturated fatty acids to form amides of ethanolamine was investigated. Enzyme activity was determined by incubation of a range of concentrations of fatty acids with [14C]ethanolamine and hippocampal membranes, followed by extraction into toluene and silica chromatography. Radiolabeled enzymatic products had the same Rf values as their respective synthetic fatty acid ethanolamides. The enzymatic activity of fatty acid ethanol-amide formation was quantitated with a phosphor-imaging plate scanner. Of all the fatty acids examined, arachidonic acid (C20:4) was the best substrate, having both the lowest EC50 (30-50 ) and the highest V. over the concentration range tested (Fig. 3). The closely related unsaturated fatty acids dihomo-y-linolenic acid (C20:3; cis-8,11,14-eicosa¬trienoic acid) and cis-11,14-eicosadienoic acid (C20:2) were the next best substrates, followed by palmitic acid (C16:0). The poorest fatty acid substrates were adrenic acid (C22:4; cis-7,10,13,16-docosatetraenoic) and docosahexaenoic acid (C22:6). Maximal activity was achieved with 100-300 µM arachidonic acid; with higher concentrations there was a marked inhibition in enzyme activity. None of the other substrates examined showed this inhibitory behavior. Arachidonyl-CoA was also examined as a substrate and was found to be almost 10-fold less potent than arachidonic acid (data not shown).
When 70 µg of hippocampal P2 membranes was incubated with 150 µM arachidonic acid and various concentrations of unlabeled ethanolamine for 20 min, the apparent K. for ethanolamine was 27 ± 4 mM and the apparent Vmax was 2230 ± 260 pmol/min per mg of protein (mean ± SD, determined by Lineweaver-Burk analysis).
Localization of Anandamide Synthase Activity in Brain Areas. To compare the enzymatic activity in various brain areas, washed P2 fractions of brain membranes were made from various regions of bovine brain. Tissue preparations from each brain area were incubated with [14C]ethanolamine and various concentrations of arachidonic acid (100-2400 AM), and [14C]anandamide synthesis was measured. A con¬centration of 100 µM arachidonic acid produced peak enzy¬matic activity in all brain areas. The amounts of anandamide produced enzymatically differed in the various brain regions examined (Fig. 4). The hippocampus had the highest enzy¬matic activity. Thalamus, striatum, and frontal cortex had a similar enzymatic activity, whereas pons, cerebellum, and medulla had the lowest activity.
To examine the subcellular distribution of the enzymatic activity, cortical P2 membranes were fractionated by density gradient centrifugation (step gradient of 0.4, 0.6, 0.8, and 1.2 M sucrose). Although enzymatic activity was observed in all fractions, enzymatic activity was highest in fractions of lower density than 0.8 M sucrose (370-530 pmol/min per mg of protein, for incubation with 1 mM arachidonic acid and 20 mM ethanolamine). These fractions contain synaptic vesi¬cles, microsomes, myelin fragments, and synaptosome ghosts (17). Lowest activity was observed in the pellet of 1.2 M sucrose (120 pmol/min per mg of protein), a fraction enriched in mitochondria (17). The cytosolic fraction exhib¬ited negligible activity.
DISCUSSION
The brain was shown to enzymatically synthesize ananda¬mide, a cannabimimetic eicosanoid, using arachidonic acid and ethanolamine as substrates. When incubated with either [14C]ethanolamine or [3H]arachidonic acid, bovine brain membranes synthesized a radioactive compound which had the same Rfvalue as anandamide following toluene extraction and chromatography. A previous report had described the enzymatic synthesis of palmitoylethanolamide from palmitic acid and ethanolamine in various tissues (20). The protocol described for the measurement of palmitoylethanolamide was not adequate to specifically measure the enzymatic synthesis of anandamide in brain, because the incorporation of [14C]ethanolamine into toluene-extractable products other than anandamide varied with both subcellular fraction and brain structure investigated, as well as with the length of incubation and the concentration of the substrates. Deter-mining whether the enzyme capable of synthesizing ananda¬mide in the brain is the same as that which synthesizes palmitoylethanolamide in other tissues must await their pu-rification.
The conjugation of ethanolamine with a fatty acid to form an amide may require a source of energy. The impurity of our preparation did not permit an extensive investigation into the possible cofactors involved. Arachidonyl-CoA was a poor substrate, and the addition of other potential energy-generating compounds such as acetyl-CoA or ATP had little effect on enzymatic activity. Enzymatic activity was retained after the membranes were washed several times in water, indicating that water-soluble cofactors are probably unnec¬essary. There are many competing reactions for the incor¬poration and metabolism of both ethanolamine and arachi¬donic acid in brain membrane fractions making further char¬acterization of the amide-forming enzyme difficult without greater purification.
Several unsaturated fatty acids have been shown to be substrates for the fatty acid ethanolamide-forming enzyme. Of the substrates examined, arachidonic acid had the highest
and the lowest ECso. In contrast to other fatty acid substrates, high concentrations of arachidonic acid were inhibitory. This inhibitory phenomena was found in the various subcellular fractions and in all areas of the brain examined. This type of kinetics suggests that anandamide synthase is regulated by the substrate arachidonic acid. It was previously found that the ethanolamides of dihomo-y-linolenic acid (C20:3) and adrenic acid (C22:4) had a potency similar to anandamide with respect to their ability to bind to the cloned human brain cannabinoid receptor and to inhibit cAMP production (13). However, dihomo-y-linolenyl etha-nolamide was less potent than anandamide at inhibiting the N-type calcium channel (13). The ethanolamides of dihomo-y-linolenic and adrenic acids have also recently been detected in porcine brain (21). These fatty acids, however, appear to be poorer substrates than arachidonic acid for the enzymatic synthesis of their respective ethanolamides.
The substrates of anandamide synthase, arachidonic acid and ethanolamine, can be generated by the activation of phospholipase A2 and phospholipase D, respectively, sug-gesting that they may arise from the hydrolysis of phospho-lipids. Preliminary evidence has indicated that the synthesis of anandamide in cultured cerebellar granule cells is evoked by a receptor-mediated influx of calcium, the activation of phospholipase A2, and the liberation of arachidonic acid
(C. C. Felder, E. M. Briley, G. -M. Yan, and J.A., unpub-lished data).
Anandamide synthase activity was present in all areas of bovine brain examined. The highest enzymatic activity was found in the hippocampus, and the lowest levels were found in the cerebellum and medulla. The cannabinoid receptor is also widely distributed with high levels in the striatum, cerebellum, and hippocampus, but the relative levels of cannabinoid receptor vary with species (22). A more detailed study of the distribution of anandamide synthesis in various areas of the brain is needed.
Anandamide, a highly lipophilic compound, could cross the synapse and be taken up presynaptically, thus serving as a retrograde transmitter. Lipids such as arachidonic acid (23) and platelet-activating factor (24) have been shown to be retrograde transmitters involved in regulating long-term po-tentiation. It is possible that anandamide may play a similar role. That the enzymatic activity is high in the hippocampus, a brain area considered to be important in the regulation of long-term potentiation, supports this suggestion.
Behavior studies in mice have shown that anandamide can affect locomotor activity (15, 16). The presence of ananda-mide synthase activity in striatum and cerebellum, areas of the brain involved in movement, suggests that anandaergic neurons may also be involved in the control of movement. Since cannabinoid drugs produce euphoria and paranoia, as well as alter the perception of time (25), anandaergic neurons may also be involved in these processes.
We thank Drs. D. C. Button, C. C. Felder, and J. K. Northup for helpful discussions and for critical reading of the manuscript.
1. Gaom, Y. & Mechoulam, R. (1964) J. Am. Chem. Soc. 86, 1646-1647.
2. Devane, W. A., Dysarz, F. A., III, Johnson, M. R., Melvin,
L. S. & Howlett, A. C. (1988) Mol. Pharmacol. 34, 605-613.
3. Matsuda, L. A., Lolait, S. J., Brownstein, M., Young, A. &
Bonner, T. I. (1990) Nature (London) 346, 561-564.
4. Gerard, C. M., Mollereau, C., Vassart, G. & Parmentier, M.
(1991) Biochem. J. 279, 129-134.
5. Howlett, A. C. & Fleming, R. M. (1984) Mol. Pharmacol. 27, 429-436.
6. Howlett, A. C., Bidaut-Russell, M., Devane, W. A., Melvin, L. S., Johnson, M. R. & Herkenham, M. (1990) Trends Phar¬macol. Sci. 13, 420-423.
7. Felder, C. C., Veluz, J. S., Williams, H. L., Briley, E. M. & Matsuda, L. A. (1992) Mol. Pharmacol. 42, 838-845.
8. Mackie, K. & Halle, B. (1992) Proc. Natl. Acad. Sci. USA 89, 3825-3829.
9. Caulfield, M. P. & Brown, D. A. (1992) Br. J. Pharmacol. 106, 231-232.
10. Devane, W. A., Hanus, L., Breuer, A., Pertwee, R. G.,
Stevenson, L. A., Griffin, G., Gibson, D., Mandelbaum, A.,
Etinger, A. & Mechoulam, R. (1992) Science 258, 1946-1949.
11. Johnson, D. E., Heald, S. L., Dally, R. D. & Janis, R. A.
(1993) Prostaglandins Leukotrienes Essent. Fatty Acids 48,
429-437.
12. Vogel, Z., Barg, J., Levy, R., Saya, D., Heldman, E. & Mechoulam, R. (1993) J. Neurochem. 61, 352-355.
13. Felder, C. C., Briley, E. M., Axelrod, J., Simpson, J. T., Mackie, K. & Devane, W. A. (1993) Proc. Natl. Acad. Sci. USA 90, 7656-7660.
14. Mackie, K., Devane, W. A. & Hille, B. (1993) Mol. Pharmacol. 44,498-503.
15. Fride, E. & Mechoulam, R. (1993) Eur. J. Pharmacol. 231, 313-314.
16. Crawley, J. N., Corwin, R. L., Robinson, J. K., Felder, C. C., Devane, W. A. & Axelrod, J. (1993) Pharmacol. Biochem. Behav. 46, 967-972.
17. Whittaker, V. P., Michaelson, I. A. & Kirkland, R. J. A. (1964) Biochem. J. 90, 293-303.
18. Deutsch, D. G. & Chin, S. A. (1993) Biochem. Pharmacol. 46, 791-7%.
19. Childers, S. R., Sexton, T. & Roy, M. B. (1994) Biochem. Pharmacol. 47, 711-715.
20. Bachur, N. R. & Udenfriend, S. (1966) J. Biol. Chem. 241, 1308-1313.
21. Hanus, L., Gopher, A., Almog, S. & Mechoulam, R. (1993) J. Med. Chem. 36, 3032-3034.
22. Herkenham, M., Lynn, A. B., Little, M. D., Johnson, M. R., Melvin, L. S., de Costa, B. R. & Rice, K. C. (1990) Proc. Natl. Acad. Sci. USA 87, 1932-1936.
23. Williams, J. H., Errington, M. L., Lynch, M. A. & Bliss, T. V. P. (1991) Nature (London) 341, 739-742.
24. to, K., Clark, G. D., Bazars, N. G. & Zorumski, C. F. (1994) Nature (London) 367, 175-179.
25. Hollister, L. E. (1986) Pharmacol. Rev. 38, 1-20.
Source: Enzymatic synthesis of anandamide, an endogenous ligand for the cannabinoid receptor, by brain membranes
