Truth Seeker
New Member
Abstract
An amidohydrolase activity present in rat brain microsomes catalyzes the hydrolysis of N-arachidonoyl[3H]ethanolamine ([3H]anandamide), an endogenous cannabimimetic substance, forming [3H]ethanolamine and arachidonic acid. Amidohydrolase activity is maximal at pH 6 and 8, is independent of divalent cations, has an apparent K for [3H]anandamide of 12.7 ± 1.8 μM, and has a Vmax of 5630 ± 200 pmol/min/mg of protein. Phenylmethylsulfonyl fluoride, a serine protease inhibitor, and p-bromophenacyl bromide, a histidine-alkylating reagent, inhibit the activity, whereas N-ethylmaleimide and various nonselective peptidase inhibitors (EDTA, o-phenanthroline, bacitracin) have no effect. Brain amidohydrolase activity exhibits high substrate specificity for [3H]anandamide; N--linolenoyl-, N-homo--linolenoyl-, and N-11,14-eicosadienoyl- are hydrolyzed at markedly slower rates. Moreover, N-11-eicosaenoyl- and N-palmitoyl-[3H]ethanolamine are not hydrolyzed. [3H]Anandamide hydrolysis is inhibited competitively by nonradioactive anandamide and by other N-acylethanolamines with the following rank order of potency: anandamide > N-linoleoyl- = N-cis-linolenoyl- = N--linolenoyl- = N-homo--linolenoyl- > N-11,14-eicosadienoyl- > N-oleoyl- > N-docosahexaenoyl- > N-docosatetraenoyl > N-linoelaidoyl- > N-eicosaenoyl- > N-palmitoyl ≥ N-elaidoyl- = N-eicosanoyl-ethanolamine = no effect. Amidohydrolase activity is high in liver and brain and low in heart, kidney, intestine, stomach, lung, spleen, and skeletal muscle. Within the central nervous system, highest activity is found in globus pallidus and hippocampus, two regions rich in cannabinoid receptors, and lowest activity is found in brainstem and medulla, where cannabinoid receptors are sparse. The results, showing that brain amidohydrolase activity is selective for anandamide and enriched in areas of the central nervous system with high density of cannabinoid receptors, suggest that this activity may participate in the inactivation of anandamide at its sites of action.
INTRODUCTION
The discovery of a G protein-coupled membrane receptor that recognizes Δ9-tetrahydrocannabinol, the major psychoactive principle of Cannabis sativa, has prompted the search for an endogenous substance with cannabimimetic properties (for review, see Howlett et al.(1990)). This search has recently led to the isolation from porcine brain of a cannabimimetic lipid derivative, identified as N-arachidonoylethanolamine and named anandamide (Devane et al., 1992). Like Δ9-tetrahydrocannabinol, anandamide binds with high affinity to brain cannabinoid receptors, reduces electrically evoked contractions in mouse vas deferens, and modulates the activities of adenylyl cyclase and voltage-dependent Ca channels in neuroblastoma cell lines (Devane et al., 1992; Vogel et al., 1993; Felder et al., 1993; Mackie et al., 1993). Moreover, anandamide produces, in vivo, a series of behavioral responses typical of cannabinoid drug administration, including catalepsy, hypothermia, analgesia, and activation of the hypothalamo-pituitary axis (Smith et al., 1994; Weidenfeld et al., 1994).
Recently, two other cannabimimetic N-acylethanolamines (N-homo--linolenoyl- and N-docosatetraenoyl ethanolamine) have been identified in brain tissue (Hanuset al., 1993; Pertwee et al., 1994), suggesting that anandamide may belong to a family of lipid mediators serving as endogenous cannabimimetic messengers in the central nervous system. Biochemical studies lend further support to this hypothesis. Rat brain preparations were shown to catalyze the synthesis of anandamide via energy-independent condensation of arachidonic acid with ethanolamine (Kruszka and Gross, 1994; Devane and Axelrod, 1994). Furthermore, studies in our laboratory have shown that rat brain neurons in primary culture produce and release anandamide and other N-acylethanolamines when they are stimulated with membrane-depolarizing agents or Caionophores. This reaction, which is both Ca-dependent and neuron-specific, involves the phosphodiesterase-mediated cleavage of a membrane phospholipid precursor, N-acylphosphatidylethanolamine (Di Marzo et al., 1994). Together, these results suggest that multiple pathways of anandamide formation may coexist in nervous tissue.
Despite these important advances, the fate of endogenous anandamide in the central nervous system is still poorly documented. Before the discovery of anandamide, the pioneering studies of Schmid and co-workers have demonstrated that ethanolamides of saturated fatty acids are hydrolyzed in tissues by an amidohydrolase (amidase) activity with broad substrate specificity (Natarajan et al., 1984; Schmid et al., 1985, 1990). That anandamide may be a substrate for amidohydrolase activity was suggested by recent reports showing hydrolytic cleavage of exogenous [3H]anandamide by rat brain homogenates (Deutsch and Chin, 1993) or by intact brain neurons in primary culture (Di Marzo et al., 1994).
We report now that rat brain microsomes contain an amidohydrolase activity that acts with high selectivity on [3H]anandamide and other polyunsaturated N-acylethanolamines. This "anandamide amidohydrolase" activity is discretely distributed in rat central nervous system and is abundant in regions where cannabinoid receptors are densely expressed. Our results suggest therefore that anandamide amidohydrolase activity may participate in the degradation of endogenous anandamide at synaptic sites.
EXPERIMENTAL PROCEDURES
Materials
[3H]Ethanolamine (28 Ci/mol) and [3H]arachidonic acid (214 Ci/mmol) were from Amersham Corp., and fatty acid chlorides were from Nu-Chek Prep (Elysian, MN). Ethanolamine, phenylmethylsulfonyl fluoride, p-bromophenacyl bromide, N-ethylmaleimide, thimerosal, bacitracin, and o-phenanthroline were from Sigma.
Syntheses and Purification of Substrates
Syntheses of both radioactive and nonradioactive N-acylethanolamines were performed essentially as described previously (Devane et al., 1992). Briefly, ethanolamine (300 μmol) was allowed to react with the appropriate fatty acyl chloride (30 μmol) in dichloromethane (4 ml) at 0−4°C, and the reaction was stopped after 15 min by extracting excess ethanolamine with water (15 ml). To prepare radioactive N-acylethanolamines, [3H]ethanolamine hydrochloride (50 μCi) was dried under vacuum, resuspended in the appropriate volume of nonradioactive ethanolamine, and added to the reaction mixtures. N-[3H]Arachidonoylethanolamine was synthesized by allowing [3H]arachidonic acid (28 μmol, 100 μCi) to react with ethanolamine (164 μmol) in the presence of N,N′-diisopropylcarbodiimide (8.4 mg) in 1 ml of dichloromethane for 40 min at 0−4°C. N-Acylethanolamines were purified by semipreparative reversed-phase HPLC, (1)by using a Nova-Pak HR C18 column (6 μm, 7.8 × 300 mm) eluted at 3 ml/min with a gradient of methanol in water (from 70 to 100% over 40 min). Fractions containing N-acylethanolamines were dried, resuspended in methanol to a final concentration of ≈10 mM, and stored at −80°C. Specific radioactivities of the N-acyl-[3H]ethanolamines used in the present study were as follows: N-arachidonoyl-[3H]ethanolamine, 0.31 mCi/mmol; N-dihomo--linolenoyl-[3H]ethanolamine, 0.52 mCi/mmol; N-11,14-eicosadienoyl-[3H]ethanolamine, 0.51 mCi/mmol; N-11-eicosamonoenoyl-[3H]ethanolamine, 0.62 mCi/mmol; N-palmitoyl-[3H]ethanolamine, 0.31 mCi/mmol; N--linolenoyl-[3H]ethanolamine, 0.69 mCi/mmol.
Preparation of Rat Tissue Homogenates and Microsome Fraction
Wistar rats were sacrificed by cervical dislocation, and brain and other tissues were homogenized in 20 mM Tris, pH 7.5, containing 0.32 M sucrose and 1 mM EGTA. Tissue extracts were centrifuged sequentially at 1,000 × g (1 min), 22,000 × g (30 min), and 105,000 × g (60 min). The soluble fraction obtained in the last centrifugation step (microsome fraction) was stored at −80°C until use. In some experiments, the brains were cut manually into ≈4-mm slices with a razor blade, and individual brain structures were dissected and homogenized in 1 ml of Tris buffer (50 mM, pH 7.5).
Amidohydrolase Assay
Standard amidohydrolase assays were carried out for 5 min at 37°C in 1 ml of Tris buffer (50 mM, pH 7.5) containing microsomes (0.2 mg of protein) and [3H]anandamide (11-14 μM, 10,000 dpm/ml). [3H]Anandamide and other N-acyl-[3H]ethanolamines were added in methanol to yield a final concentration of 0.2% (v/v). Identical incubations were carried out in the absence of tissue: these "blank" samples contained ≈40 dpm/sample, which were subtracted from values obtained with tissue samples. Amidohydrolase activity in brain regions was measured in freshly dissected tissue homogenates (1 mg/ml), which were incubated for 10 min at 37°C in the presence of [3H]anandamide (11 μM, 10,000 dpm). Under these conditions, [3H]anandamide hydrolysis was found in preliminary experiments to be linear with respect to time and protein concentration (data not shown). The reaction mixtures were diluted with methanol (4 ml) and centrifuged to eliminate precipitated protein, and samples (1 ml) were applied onto glass wool-plugged Pasteur pipettes, which had been previously packed with a suspension (100 mg/ml, 1 ml) of polydivinylbenzene (Porapak type Q, Waters) in methanol. [3H]Ethanolamine was eluted in the void volume, and measured by liquid scintillation counting. Under these elution conditions, unreacted [3H]anandamide and [3H]ethanolamine-labeled phospholipids were quantitatively retained on the columns, which could be regenerated with ethanol (≈20 ml).
Additional Analytical Procedures
To determine whether [3H]anandamide undergoes oxidative metabolism in brain microsomes, samples of amidohydrolase assays carried out for 10 min were diluted with an equal volume of methanol and extracted with chloroform (2 ml). The organic phases were applied onto Pasteur pipettes packed with a slurry of silica gel G in chloroform to yield a final column volume of ≈0.5 ml. Radioactive products were eluted from the columns with 1 ml of ethyl acetate and brought to dryness under vacuum. Reversed-phase HPLC analysis was performed using a Nova-Pak C18 column (3.9 × 150 mm, Waters) eluted at 1 ml/min with a gradient of methanol in water (from 60 to 100% over 20 min). Elution of products was followed by monitoring UV absorbance at 214 nm. 1-min fractions were collected for liquid scintillation counting. Thin layer chromatography analysis was carried out on silica gel G-coated plastic plates (Merck) eluted with a solvent system of chloroform/methanol (9:1). The RF of anandamide in this solvent system was ≈0.5. After having visualized the lipids with iodine vapors, 0.5-cm bands were cut, and radioactivity in the bands was determined by liquid scintillation counting.
Data Presentation
Each experiment was carried out in quintuplicate and repeated at least three times with identical results. Data are expressed as mean ± S.E.
RESULTS AND DISCUSSION
Hydrolysis of [3H]Anandamide
Exogenous [3H]anandamide, labeled radioactively on the ethanolamine moiety, was rapidly degraded by rat brain microsomes forming a product which, on polydivinylbenzene columns, had the chromatographic behavior of [3H]ethanolamine (Fig. 1A). When [3H]anandamide was labeled on the arachidonate moiety, similar incubations resulted in the formation of a product which comigrated with [3H]arachidonic acid on thin layer chromatography. In contrast, no degradation occurred in boiled tissue samples (data not shown). The results suggest that rat brain microsomes contain an amidohydrolase activity that catalyzes the cleavage of exogenous anandamide to ethanolamine and arachidonic acid. The subcellular fractionation illustrated in Table 1 shows that this activity was present in both particulate and microsomal fractions and was enriched in the latter.
Mouse liver microsomes incubated in the presence of NADPH convert anandamide into several oxygenated products, possibly formed by cytochrome P-450-dependent monooxygenation (Bornheim et al., 1993). Under our experimental conditions, oxidative metabolism does not appear, however, to participate in [3H]anandamide degradation. When samples from amidohydrolase assays (10-min incubation) were analyzed by reversed-phase HPLC, no evidence was found for the formation of polar [3H]anandamide metabolites (data not shown).
Amidohydrolase activity in microsomes was dependent on the concentration of protein (Fig. 1B) and was optimal at pH 6 and 8 (Fig. 1C). Moreover, the activity was dependent on the concentration of [3H]anandamide, with an apparent Km of 12.7 ± 1.8 μM and a Vmax of 5630 ± 200 pmol/min/mg of protein (n = 3). Divalent cations were neither necessary for nor stimulatory on amidohydrolase activity; adding EGTA, EDTA, CaCl2, or MgCl2 (each at 10 mM) had little or no effect (Table 2).
The serine protease inhibitor, phenylmethylsulfonyl fluoride, was shown to prevent the degradation of anandamide in brain homogenates (Deutsch and Chin, 1993) and to improve its metabolic stability in binding assays carried out on rat brain membranes (Abadji et al., 1994; Childers et al., 1994). In our experiments, phenylmethylsulfonyl fluoride inhibited amidohydrolase activity in a concentration-dependent manner (Table 2). The effects of other protein-alkylating reagents (N-ethylmaleimide and p-bromophenacyl bromide), as well as of nonselective peptidase inhibitors (EDTA, o-phenanthroline, bacitracin) are also shown in Table 2. Among these, only p-bromophenacyl bromide, which alkylates histidine residues on proteins, produced a significant inhibition of the activity.
Substrate Specificity in Brain Microsomes
To study the substrate specificity of rat brain microsome amidohydrolase, in a first series of experiments, we compared the rate of hydrolysis of [3H]anandamide (short-hand designation of the N-acyl chain = 20:4 Δ) with those of five congeners, whose structures are depicted in Fig. 2(N-homo--linolenoyl-[3H]ethanolamine (20:3 Δ), N-11,14-eicosadienoyl-[3H]ethanolamine (20:2 Δ), N-11-eicosaenoyl-[3H]ethanolamine (20:1 Δ), N--linolenoyl-[3H]ethanolamine (18:3 Δ), and N-palmitoyl-[3H]ethanolamine (16:0)). The results indicate that, among the substrates tested, [3H]anandamide is hydrolyzed at the highest rate (Fig. 2). Although more slowly than anandamide, other polyunsaturated N-acylethanolamines are also hydrolyzed by brain amidohydrolase activity, whereas monounsaturated and saturated N-acylethanolamines are not. Indeed, very little hydrolysis of N-palmitoyl-[3H]ethanolamine was observed even when the microsomes were incubated in the presence of 30 μMN-palmitoyl-[3H]ethanolamine (39 ± 2 pmol/min/mg of protein).
In a second series of experiments, [3H]anandamide hydrolysis was measured in the presence of various nonradioactive N-acylethanolamines (Fig. 3). The concentration of [3H]anandamide in these assays was 14 μM, and that of the competing products was 200 μM. The results show that [3H]anandamide hydrolysis was best inhibited by unsaturated ethanolamides with a fatty acyl chain containing 18-20 carbon atoms and a number of cis double bonds comprised between two and four (Fig. 3). Structural modifications that resulted in reduced inhibitory efficacy included the following: 1) elongating the fatty acyl chain above 20 carbon atoms; 2) replacing cis double bonds with trans double bonds; 3) decreasing the number of double bonds to one or eliminating them (Fig. 3).
Although our experiments do not provide a complete characterization of the structural requirements of brain microsome amidohydrolase activity, which may be best accomplished on a purified enzyme preparation, they do suggest that this activity is highly selective for anandamide. That other unsaturated N-acylethanolamines may be also substrates for this activity is in agreement with studies showing that more than one anandamide may be produced and released by stimulated neurons (Hanus et al., 1993; Pertwee et al., 1994; Di Marzo et al., 1994).
Substrate Specificity in Brain Particulate Fractions
The results shown in Table 3 demonstrate that the relative rates of hydrolysis of various N-acyl-[3H]ethanolamines and the inhibition of [3H]anandamide hydrolysis by various nonradioactive N-acylethanolamines are similar in brain particulate fractions and microsomes. One notable exception is N-palmitoyl-[3H]ethanolamine, which is hydrolyzed by particulate fractions at a rate that greatly exceeds that observed in microsomes. This result suggests that component(s) of particulate fraction may contain a nonspecific short chain amidohydrolase, which may participate in the hydrolysis of endogenous saturated N-acylethanolamines (Di Marzo et al., 1994).
Tissue Distribution
High levels of amidohydrolase activity were found in microsomes prepared from liver and brain tissues, whereas activity was low in kidney, intestine, stomach, lung, and spleen and barely detectable in skeletal muscle (Fig. 4A). Preliminary experiments showed that the substrate specificity of liver microsome amidohydrolase activity was similar to that observed in brain microsomes (data not shown).
Regional Distribution in Brain
Within the central nervous system, highest levels of amidohydrolase activity were measured in homogenates of globus pallidus, hippocampus, and substantia nigra; intermediate levels were measured in striatum, thalamus, cerebellum, and cortex; and lowest levels were measured in brainstem and medulla (Fig. 4B). Quantitative autoradiographic studies have shown that cannabinoid receptor binding is dense in globus pallidus, substantia nigra, hippocampus, entopeduncular nucleus, and cerebellum; moderate in cerebral cortex and caudate-putamen; and sparse in brainstem and spinal cord (Howlett et al., 1990; Herkenham et al., 1990, 1991). Thus, the levels of amidohydrolase activity in brain were in good correlation with the distribution of cannabinoid receptors. An important exception was constituted by the cerebellum, where to high levels of receptor density corresponded intermediate levels of amidohydrolase activity. This discrepancy may be accounted for by the fact that, within the cerebellum, cannabinoid receptors are densely concentrated in the molecular layer and only sparsely present elsewhere (Herkenham et al., 1991). Because our measurements were carried out in homogenates of whole cerebella, they likely reflect average levels of amidohydrolase activity in cerebellar tissue.
Conclusions
Two lines of evidence suggest that the amidohydrolase activity identified and partially characterized in the present study participates in the physiological degradation of anandamide in the central nervous system. First, this activity, unlike those described in previous studies (Natarajan et al., 1984; Schmid et al., 1985, 1990), appears to be highly selective for [3H]anandamide. Criteria of selectivity include 1) the rate of hydrolysis, 2-100 fold faster with [3H]anandamide than with a series of closely related congeners and 2) the rank potency of nonradioactive N-acylethanolamines to compete for [3H]anandamide hydrolysis, greatest with anandamide and with its close structural congeners. Second, anandamide amidohydrolase activity is discretely distributed in the rat central nervous system where its localization parallels, by and large, that of cannabinoid receptors (Herkenham et al., 1990, 1991; Matsuda et al., 1993). As for other neurotransmitter-metabolizing enzymes (e.g. acetylcholinesterase, Butcher and Wolfe(1984)), such parallel distribution supports the possibility that anandamide amidohydrolase activity participates in terminating the actions of anandamide at its sites of action.
Source, Graphs and Figures: Anandamide Amidohydrolase Activity in Rat Brain Microsomes
An amidohydrolase activity present in rat brain microsomes catalyzes the hydrolysis of N-arachidonoyl[3H]ethanolamine ([3H]anandamide), an endogenous cannabimimetic substance, forming [3H]ethanolamine and arachidonic acid. Amidohydrolase activity is maximal at pH 6 and 8, is independent of divalent cations, has an apparent K for [3H]anandamide of 12.7 ± 1.8 μM, and has a Vmax of 5630 ± 200 pmol/min/mg of protein. Phenylmethylsulfonyl fluoride, a serine protease inhibitor, and p-bromophenacyl bromide, a histidine-alkylating reagent, inhibit the activity, whereas N-ethylmaleimide and various nonselective peptidase inhibitors (EDTA, o-phenanthroline, bacitracin) have no effect. Brain amidohydrolase activity exhibits high substrate specificity for [3H]anandamide; N--linolenoyl-, N-homo--linolenoyl-, and N-11,14-eicosadienoyl- are hydrolyzed at markedly slower rates. Moreover, N-11-eicosaenoyl- and N-palmitoyl-[3H]ethanolamine are not hydrolyzed. [3H]Anandamide hydrolysis is inhibited competitively by nonradioactive anandamide and by other N-acylethanolamines with the following rank order of potency: anandamide > N-linoleoyl- = N-cis-linolenoyl- = N--linolenoyl- = N-homo--linolenoyl- > N-11,14-eicosadienoyl- > N-oleoyl- > N-docosahexaenoyl- > N-docosatetraenoyl > N-linoelaidoyl- > N-eicosaenoyl- > N-palmitoyl ≥ N-elaidoyl- = N-eicosanoyl-ethanolamine = no effect. Amidohydrolase activity is high in liver and brain and low in heart, kidney, intestine, stomach, lung, spleen, and skeletal muscle. Within the central nervous system, highest activity is found in globus pallidus and hippocampus, two regions rich in cannabinoid receptors, and lowest activity is found in brainstem and medulla, where cannabinoid receptors are sparse. The results, showing that brain amidohydrolase activity is selective for anandamide and enriched in areas of the central nervous system with high density of cannabinoid receptors, suggest that this activity may participate in the inactivation of anandamide at its sites of action.
INTRODUCTION
The discovery of a G protein-coupled membrane receptor that recognizes Δ9-tetrahydrocannabinol, the major psychoactive principle of Cannabis sativa, has prompted the search for an endogenous substance with cannabimimetic properties (for review, see Howlett et al.(1990)). This search has recently led to the isolation from porcine brain of a cannabimimetic lipid derivative, identified as N-arachidonoylethanolamine and named anandamide (Devane et al., 1992). Like Δ9-tetrahydrocannabinol, anandamide binds with high affinity to brain cannabinoid receptors, reduces electrically evoked contractions in mouse vas deferens, and modulates the activities of adenylyl cyclase and voltage-dependent Ca channels in neuroblastoma cell lines (Devane et al., 1992; Vogel et al., 1993; Felder et al., 1993; Mackie et al., 1993). Moreover, anandamide produces, in vivo, a series of behavioral responses typical of cannabinoid drug administration, including catalepsy, hypothermia, analgesia, and activation of the hypothalamo-pituitary axis (Smith et al., 1994; Weidenfeld et al., 1994).
Recently, two other cannabimimetic N-acylethanolamines (N-homo--linolenoyl- and N-docosatetraenoyl ethanolamine) have been identified in brain tissue (Hanuset al., 1993; Pertwee et al., 1994), suggesting that anandamide may belong to a family of lipid mediators serving as endogenous cannabimimetic messengers in the central nervous system. Biochemical studies lend further support to this hypothesis. Rat brain preparations were shown to catalyze the synthesis of anandamide via energy-independent condensation of arachidonic acid with ethanolamine (Kruszka and Gross, 1994; Devane and Axelrod, 1994). Furthermore, studies in our laboratory have shown that rat brain neurons in primary culture produce and release anandamide and other N-acylethanolamines when they are stimulated with membrane-depolarizing agents or Caionophores. This reaction, which is both Ca-dependent and neuron-specific, involves the phosphodiesterase-mediated cleavage of a membrane phospholipid precursor, N-acylphosphatidylethanolamine (Di Marzo et al., 1994). Together, these results suggest that multiple pathways of anandamide formation may coexist in nervous tissue.
Despite these important advances, the fate of endogenous anandamide in the central nervous system is still poorly documented. Before the discovery of anandamide, the pioneering studies of Schmid and co-workers have demonstrated that ethanolamides of saturated fatty acids are hydrolyzed in tissues by an amidohydrolase (amidase) activity with broad substrate specificity (Natarajan et al., 1984; Schmid et al., 1985, 1990). That anandamide may be a substrate for amidohydrolase activity was suggested by recent reports showing hydrolytic cleavage of exogenous [3H]anandamide by rat brain homogenates (Deutsch and Chin, 1993) or by intact brain neurons in primary culture (Di Marzo et al., 1994).
We report now that rat brain microsomes contain an amidohydrolase activity that acts with high selectivity on [3H]anandamide and other polyunsaturated N-acylethanolamines. This "anandamide amidohydrolase" activity is discretely distributed in rat central nervous system and is abundant in regions where cannabinoid receptors are densely expressed. Our results suggest therefore that anandamide amidohydrolase activity may participate in the degradation of endogenous anandamide at synaptic sites.
EXPERIMENTAL PROCEDURES
Materials
[3H]Ethanolamine (28 Ci/mol) and [3H]arachidonic acid (214 Ci/mmol) were from Amersham Corp., and fatty acid chlorides were from Nu-Chek Prep (Elysian, MN). Ethanolamine, phenylmethylsulfonyl fluoride, p-bromophenacyl bromide, N-ethylmaleimide, thimerosal, bacitracin, and o-phenanthroline were from Sigma.
Syntheses and Purification of Substrates
Syntheses of both radioactive and nonradioactive N-acylethanolamines were performed essentially as described previously (Devane et al., 1992). Briefly, ethanolamine (300 μmol) was allowed to react with the appropriate fatty acyl chloride (30 μmol) in dichloromethane (4 ml) at 0−4°C, and the reaction was stopped after 15 min by extracting excess ethanolamine with water (15 ml). To prepare radioactive N-acylethanolamines, [3H]ethanolamine hydrochloride (50 μCi) was dried under vacuum, resuspended in the appropriate volume of nonradioactive ethanolamine, and added to the reaction mixtures. N-[3H]Arachidonoylethanolamine was synthesized by allowing [3H]arachidonic acid (28 μmol, 100 μCi) to react with ethanolamine (164 μmol) in the presence of N,N′-diisopropylcarbodiimide (8.4 mg) in 1 ml of dichloromethane for 40 min at 0−4°C. N-Acylethanolamines were purified by semipreparative reversed-phase HPLC, (1)by using a Nova-Pak HR C18 column (6 μm, 7.8 × 300 mm) eluted at 3 ml/min with a gradient of methanol in water (from 70 to 100% over 40 min). Fractions containing N-acylethanolamines were dried, resuspended in methanol to a final concentration of ≈10 mM, and stored at −80°C. Specific radioactivities of the N-acyl-[3H]ethanolamines used in the present study were as follows: N-arachidonoyl-[3H]ethanolamine, 0.31 mCi/mmol; N-dihomo--linolenoyl-[3H]ethanolamine, 0.52 mCi/mmol; N-11,14-eicosadienoyl-[3H]ethanolamine, 0.51 mCi/mmol; N-11-eicosamonoenoyl-[3H]ethanolamine, 0.62 mCi/mmol; N-palmitoyl-[3H]ethanolamine, 0.31 mCi/mmol; N--linolenoyl-[3H]ethanolamine, 0.69 mCi/mmol.
Preparation of Rat Tissue Homogenates and Microsome Fraction
Wistar rats were sacrificed by cervical dislocation, and brain and other tissues were homogenized in 20 mM Tris, pH 7.5, containing 0.32 M sucrose and 1 mM EGTA. Tissue extracts were centrifuged sequentially at 1,000 × g (1 min), 22,000 × g (30 min), and 105,000 × g (60 min). The soluble fraction obtained in the last centrifugation step (microsome fraction) was stored at −80°C until use. In some experiments, the brains were cut manually into ≈4-mm slices with a razor blade, and individual brain structures were dissected and homogenized in 1 ml of Tris buffer (50 mM, pH 7.5).
Amidohydrolase Assay
Standard amidohydrolase assays were carried out for 5 min at 37°C in 1 ml of Tris buffer (50 mM, pH 7.5) containing microsomes (0.2 mg of protein) and [3H]anandamide (11-14 μM, 10,000 dpm/ml). [3H]Anandamide and other N-acyl-[3H]ethanolamines were added in methanol to yield a final concentration of 0.2% (v/v). Identical incubations were carried out in the absence of tissue: these "blank" samples contained ≈40 dpm/sample, which were subtracted from values obtained with tissue samples. Amidohydrolase activity in brain regions was measured in freshly dissected tissue homogenates (1 mg/ml), which were incubated for 10 min at 37°C in the presence of [3H]anandamide (11 μM, 10,000 dpm). Under these conditions, [3H]anandamide hydrolysis was found in preliminary experiments to be linear with respect to time and protein concentration (data not shown). The reaction mixtures were diluted with methanol (4 ml) and centrifuged to eliminate precipitated protein, and samples (1 ml) were applied onto glass wool-plugged Pasteur pipettes, which had been previously packed with a suspension (100 mg/ml, 1 ml) of polydivinylbenzene (Porapak type Q, Waters) in methanol. [3H]Ethanolamine was eluted in the void volume, and measured by liquid scintillation counting. Under these elution conditions, unreacted [3H]anandamide and [3H]ethanolamine-labeled phospholipids were quantitatively retained on the columns, which could be regenerated with ethanol (≈20 ml).
Additional Analytical Procedures
To determine whether [3H]anandamide undergoes oxidative metabolism in brain microsomes, samples of amidohydrolase assays carried out for 10 min were diluted with an equal volume of methanol and extracted with chloroform (2 ml). The organic phases were applied onto Pasteur pipettes packed with a slurry of silica gel G in chloroform to yield a final column volume of ≈0.5 ml. Radioactive products were eluted from the columns with 1 ml of ethyl acetate and brought to dryness under vacuum. Reversed-phase HPLC analysis was performed using a Nova-Pak C18 column (3.9 × 150 mm, Waters) eluted at 1 ml/min with a gradient of methanol in water (from 60 to 100% over 20 min). Elution of products was followed by monitoring UV absorbance at 214 nm. 1-min fractions were collected for liquid scintillation counting. Thin layer chromatography analysis was carried out on silica gel G-coated plastic plates (Merck) eluted with a solvent system of chloroform/methanol (9:1). The RF of anandamide in this solvent system was ≈0.5. After having visualized the lipids with iodine vapors, 0.5-cm bands were cut, and radioactivity in the bands was determined by liquid scintillation counting.
Data Presentation
Each experiment was carried out in quintuplicate and repeated at least three times with identical results. Data are expressed as mean ± S.E.
RESULTS AND DISCUSSION
Hydrolysis of [3H]Anandamide
Exogenous [3H]anandamide, labeled radioactively on the ethanolamine moiety, was rapidly degraded by rat brain microsomes forming a product which, on polydivinylbenzene columns, had the chromatographic behavior of [3H]ethanolamine (Fig. 1A). When [3H]anandamide was labeled on the arachidonate moiety, similar incubations resulted in the formation of a product which comigrated with [3H]arachidonic acid on thin layer chromatography. In contrast, no degradation occurred in boiled tissue samples (data not shown). The results suggest that rat brain microsomes contain an amidohydrolase activity that catalyzes the cleavage of exogenous anandamide to ethanolamine and arachidonic acid. The subcellular fractionation illustrated in Table 1 shows that this activity was present in both particulate and microsomal fractions and was enriched in the latter.
Mouse liver microsomes incubated in the presence of NADPH convert anandamide into several oxygenated products, possibly formed by cytochrome P-450-dependent monooxygenation (Bornheim et al., 1993). Under our experimental conditions, oxidative metabolism does not appear, however, to participate in [3H]anandamide degradation. When samples from amidohydrolase assays (10-min incubation) were analyzed by reversed-phase HPLC, no evidence was found for the formation of polar [3H]anandamide metabolites (data not shown).
Amidohydrolase activity in microsomes was dependent on the concentration of protein (Fig. 1B) and was optimal at pH 6 and 8 (Fig. 1C). Moreover, the activity was dependent on the concentration of [3H]anandamide, with an apparent Km of 12.7 ± 1.8 μM and a Vmax of 5630 ± 200 pmol/min/mg of protein (n = 3). Divalent cations were neither necessary for nor stimulatory on amidohydrolase activity; adding EGTA, EDTA, CaCl2, or MgCl2 (each at 10 mM) had little or no effect (Table 2).
The serine protease inhibitor, phenylmethylsulfonyl fluoride, was shown to prevent the degradation of anandamide in brain homogenates (Deutsch and Chin, 1993) and to improve its metabolic stability in binding assays carried out on rat brain membranes (Abadji et al., 1994; Childers et al., 1994). In our experiments, phenylmethylsulfonyl fluoride inhibited amidohydrolase activity in a concentration-dependent manner (Table 2). The effects of other protein-alkylating reagents (N-ethylmaleimide and p-bromophenacyl bromide), as well as of nonselective peptidase inhibitors (EDTA, o-phenanthroline, bacitracin) are also shown in Table 2. Among these, only p-bromophenacyl bromide, which alkylates histidine residues on proteins, produced a significant inhibition of the activity.
Substrate Specificity in Brain Microsomes
To study the substrate specificity of rat brain microsome amidohydrolase, in a first series of experiments, we compared the rate of hydrolysis of [3H]anandamide (short-hand designation of the N-acyl chain = 20:4 Δ) with those of five congeners, whose structures are depicted in Fig. 2(N-homo--linolenoyl-[3H]ethanolamine (20:3 Δ), N-11,14-eicosadienoyl-[3H]ethanolamine (20:2 Δ), N-11-eicosaenoyl-[3H]ethanolamine (20:1 Δ), N--linolenoyl-[3H]ethanolamine (18:3 Δ), and N-palmitoyl-[3H]ethanolamine (16:0)). The results indicate that, among the substrates tested, [3H]anandamide is hydrolyzed at the highest rate (Fig. 2). Although more slowly than anandamide, other polyunsaturated N-acylethanolamines are also hydrolyzed by brain amidohydrolase activity, whereas monounsaturated and saturated N-acylethanolamines are not. Indeed, very little hydrolysis of N-palmitoyl-[3H]ethanolamine was observed even when the microsomes were incubated in the presence of 30 μMN-palmitoyl-[3H]ethanolamine (39 ± 2 pmol/min/mg of protein).
In a second series of experiments, [3H]anandamide hydrolysis was measured in the presence of various nonradioactive N-acylethanolamines (Fig. 3). The concentration of [3H]anandamide in these assays was 14 μM, and that of the competing products was 200 μM. The results show that [3H]anandamide hydrolysis was best inhibited by unsaturated ethanolamides with a fatty acyl chain containing 18-20 carbon atoms and a number of cis double bonds comprised between two and four (Fig. 3). Structural modifications that resulted in reduced inhibitory efficacy included the following: 1) elongating the fatty acyl chain above 20 carbon atoms; 2) replacing cis double bonds with trans double bonds; 3) decreasing the number of double bonds to one or eliminating them (Fig. 3).
Although our experiments do not provide a complete characterization of the structural requirements of brain microsome amidohydrolase activity, which may be best accomplished on a purified enzyme preparation, they do suggest that this activity is highly selective for anandamide. That other unsaturated N-acylethanolamines may be also substrates for this activity is in agreement with studies showing that more than one anandamide may be produced and released by stimulated neurons (Hanus et al., 1993; Pertwee et al., 1994; Di Marzo et al., 1994).
Substrate Specificity in Brain Particulate Fractions
The results shown in Table 3 demonstrate that the relative rates of hydrolysis of various N-acyl-[3H]ethanolamines and the inhibition of [3H]anandamide hydrolysis by various nonradioactive N-acylethanolamines are similar in brain particulate fractions and microsomes. One notable exception is N-palmitoyl-[3H]ethanolamine, which is hydrolyzed by particulate fractions at a rate that greatly exceeds that observed in microsomes. This result suggests that component(s) of particulate fraction may contain a nonspecific short chain amidohydrolase, which may participate in the hydrolysis of endogenous saturated N-acylethanolamines (Di Marzo et al., 1994).
Tissue Distribution
High levels of amidohydrolase activity were found in microsomes prepared from liver and brain tissues, whereas activity was low in kidney, intestine, stomach, lung, and spleen and barely detectable in skeletal muscle (Fig. 4A). Preliminary experiments showed that the substrate specificity of liver microsome amidohydrolase activity was similar to that observed in brain microsomes (data not shown).
Regional Distribution in Brain
Within the central nervous system, highest levels of amidohydrolase activity were measured in homogenates of globus pallidus, hippocampus, and substantia nigra; intermediate levels were measured in striatum, thalamus, cerebellum, and cortex; and lowest levels were measured in brainstem and medulla (Fig. 4B). Quantitative autoradiographic studies have shown that cannabinoid receptor binding is dense in globus pallidus, substantia nigra, hippocampus, entopeduncular nucleus, and cerebellum; moderate in cerebral cortex and caudate-putamen; and sparse in brainstem and spinal cord (Howlett et al., 1990; Herkenham et al., 1990, 1991). Thus, the levels of amidohydrolase activity in brain were in good correlation with the distribution of cannabinoid receptors. An important exception was constituted by the cerebellum, where to high levels of receptor density corresponded intermediate levels of amidohydrolase activity. This discrepancy may be accounted for by the fact that, within the cerebellum, cannabinoid receptors are densely concentrated in the molecular layer and only sparsely present elsewhere (Herkenham et al., 1991). Because our measurements were carried out in homogenates of whole cerebella, they likely reflect average levels of amidohydrolase activity in cerebellar tissue.
Conclusions
Two lines of evidence suggest that the amidohydrolase activity identified and partially characterized in the present study participates in the physiological degradation of anandamide in the central nervous system. First, this activity, unlike those described in previous studies (Natarajan et al., 1984; Schmid et al., 1985, 1990), appears to be highly selective for [3H]anandamide. Criteria of selectivity include 1) the rate of hydrolysis, 2-100 fold faster with [3H]anandamide than with a series of closely related congeners and 2) the rank potency of nonradioactive N-acylethanolamines to compete for [3H]anandamide hydrolysis, greatest with anandamide and with its close structural congeners. Second, anandamide amidohydrolase activity is discretely distributed in the rat central nervous system where its localization parallels, by and large, that of cannabinoid receptors (Herkenham et al., 1990, 1991; Matsuda et al., 1993). As for other neurotransmitter-metabolizing enzymes (e.g. acetylcholinesterase, Butcher and Wolfe(1984)), such parallel distribution supports the possibility that anandamide amidohydrolase activity participates in terminating the actions of anandamide at its sites of action.
Source, Graphs and Figures: Anandamide Amidohydrolase Activity in Rat Brain Microsomes
