Purification and Characterization of an Acid Amidase Selective for N-Palmitoylethanolamine, a Putative Endogenous Anti-inflammatory Substance*

N-Arachidonoylethanolamine (anandamide) is cannabimimetic, and N-palmitoylethanolamine is anti-inflammatory and immunosuppressive. We found an amidase that is more active with the latter than the former in contrast to the previously known anandamide amidohydrolase for whichN-palmitoylethanolamine is a poor substrate. Proteins solubilized by freezing and thawing from the 12,000 ×g pellet of various rat organs hydrolyzed [14C]N-palmitoylethanolamine to palmitic acid and ethanolamine. The specific enzyme activity was higher in the order of lung > spleen > small intestine > thymus > cecum, and high activity was found in peritoneal and alveolar macrophages. The enzyme with a molecular mass of 31 kDa was purified from rat lung to a specific activity of 1.8 μmol/min/mg protein. Relative reactivities of the enzyme with variousN-acylethanolamines (100 μm) were as follows:N-palmitoylethanolamine, 100%;N-myristoylethanolamine, 48%;N-stearoylethanolamine, 21%; N-oleoylethanolamine, 20%;N-linoleoylethanolamine, 13%; anandamide, 8%. The enzyme was the most active at pH 5 and was activated 7-fold by Triton X-100. The enzyme was almost insensitive to methyl arachidonyl fluorophosphonate, which inhibited anandamide amidohydrolase potently. Thus, the new enzyme referred to asN-palmitoylethanolamine hydrolase was clearly distinguishable from anandamide amidohydrolase.

Very recently we found an anandamide-hydrolyzing enzyme, which was catalytically distinct from the previously known anandamide amidohydrolase, in a human megakaryoblastic leukemia cell line (CMK) (36). The enzyme showed an optimal pH value at 5 rather than 8.5-10 and was much less sensitive to phenylmethylsulfonyl fluoride and MAFP. Interestingly, N-palmitoylethanolamine was 1.5-fold more active than anandamide. For further characterization of the enzyme and elucidation of its physiological significance, we attempted to purify the enzyme and examined its tissue distribution.

EXPERIMENTAL PROCEDURES
Materials- [1-14 C]Arachidonic, [1-14 C]linoleic, and [1-14 C]oleic acids, Phenyl-Sepharose CL-4B, HiTrap Heparin HP, and HiTrap Butyl FF were purchased from Amersham Pharmacia Biotech, and [1-14 C]palmitic, [1-14 C]stearic, and [1-14 C]myristic acids were from PerkinElmer Life Sciences. Linoleic and oleic acids were from Nu-Chek-Prep (Elysian, MN), stearic and myristic acids and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine were from Sigma, anandamide, N-palmitoylethanolamine, and MAFP were from Cayman Chemical Company (Ann Arbor, MI), sphingosine was from Biomol (Plymouth Meeting, PA), dithiothreitol (DTT) and p-chloromercuribenzoic acid were from Wako Pure Chemical (Osaka, Japan), Triton X-100 and Tween 20 were from Nacalai Tesque (Kyoto, Japan), CHAPS was from Dojindo (Kumamoto, Japan), n-octyl-␤-D-glucoside was from Katayama Chemical (Osaka, Japan), hydroxyapatite Bio-Gel HTP gel and protein assay dye reagent concentrate were from Bio-Rad, and precoated Silica Gel 60 F 254 glass plates for thin-layer chromatography (TLC) (20 cm ϫ 20 cm, 0.25-mm thickness) were from * This work was supported by grants-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, Human Frontier Science Program, ONO Medical Research Foundation, the Japan Foundation for Applied Enzymology, Ono Pharmaceutical Company, Kissei Pharmaceutical Company, Japan Tobacco Incorporation, and Takeda Pharmaceutical Industry. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Enzyme Preparation-Male Wistar rats (220 -280-g weight; Clea Japan or Charles River Japan) were anesthetized by diethyl ether and sacrificed by cervical dislocation. Various organs were removed and homogenized in 9 times the volume (v/w) of ice-cold 20 mM Tris-HCl (pH 7.4) containing 0.32 M sucrose with a Polytron homogenizer. The homogenates were centrifuged at 800 ϫ g for 15 min and at 12,000 ϫ g for 30 min, successively. The 12,000 ϫ g pellet was suspended in phosphate-buffered saline (pH 7.4) and subjected to two cycles of freezing and thawing. The sample was then centrifuged at 105,000 ϫ g for 50 min. The resultant supernatant was used as the solubilized enzyme. The samples thus prepared were stored at Ϫ80°C until use. Protein concentration was determined by the method of Bradford (39) with bovine serum albumin as a standard.
Macrophages were obtained from the peritoneal cavity of male Wistar rats (500 -600-g body weight) and were plated on plastic dishes. The adherent cells were used as peritoneal resident macrophages. Alveolar macrophages were obtained from bronchoalveolar lavage fluid of rat lung. The macrophages were suspended in phosphate-buffered saline (pH 7.4) and subjected to sonic disruption. The homogenate was centrifuged at 12,000 ϫ g for 30 min. Solubilization of the enzyme from the 12,000 ϫ g pellet was performed as described above. Recombinant rat anandamide amidohydrolase was overexpressed in a baculovirus-Sf9 insect cell system (35), and the particulate fraction of the transfected Sf9 cells was used for enzyme assay.
Enzyme Assay-The enzyme was incubated with 100 M [ 14 C]Npalmitoylethanolamine (10,000 cpm in 10 l of dimethyl sulfoxide) at 37°C for 30 min in 100 l of 50 mM citrate-sodium phosphate (pH 5.0) containing 3 mM DTT and 0.1% Triton X-100. The reaction was terminated by addition of 0.35 ml of a mixture of diethyl ether/methanol/1 M citric acid (30:4:1, v/v). The ethereal extract was spotted on a Silica Gel 60 F 254 glass plate (10-cm height) and subjected to TLC with a mixture of chloroform/methanol/25% ammonium hydroxide (80:20:2, v/v) for 20 min at 4°C. Radioactivity on the plate was scanned with a BAS 2000 bioimaging analyzer (Fujix, Tokyo, Japan). Assays were performed in triplicate.
Purification of the Enzyme-The solubilized protein of lung (160 mg of protein in 100 ml) was adjusted to pH 4 with HCl and kept at 4°C for 1 h. Resultant-insoluble proteins were removed by centrifugation at 260,000 ϫ g for 40 min. The supernatant was further kept at 4°C overnight and was subjected to centrifugation again. The supernatant (10 ml) was then loaded onto Phenyl-Sepharose CL-4B (1-ml bed volume) pre-equilibrated with 20 mM Tris-HCl (pH 7.4). After the column was washed with 10 ml of 20 mM Tris-HCl (pH 7.4), adsorbed proteins were eluted in 1.0-ml fractions with 1% octyl glucoside. This chromatography was performed 10 times using 100 ml of the sample in total, and all the active fractions were pooled. The sample was then loaded on a HiTrap Heparin column (1-ml bed volume) pre-equilibrated with 20 mM Tris-HCl (pH 7.4) containing 1% octyl glucoside. The flowthrough fraction was applied onto a hydroxyapatite column (Bio-Gel HTP; 1-ml bed volume) pre-equilibrated with 20 mM Tris-HCl (pH 7.

Organ Distribution of "N-Palmitoylethanolamine
Hydrolase"-The soluble fractions prepared by freezing and thawing from the 12,000 ϫ g pellet of various rat organs were assayed for the N-palmitoylethanolamine-hydrolyzing activity. To rule out a possible contribution of the contaminated membranebound anandamide amidohydrolase, we pretreated the enzyme preparation with 1 M MAFP, which almost completely inhibited the N-palmitoylethanolamine hydrolysis catalyzed by anandamide amidohydrolase as shown later in Fig. 4B. The results revealed a wide distribution of the MAFP-insensitive enzyme (Fig. 1). The highest activity was observed in lung with a specific activity of 7.1 nmol/min/mg protein at 37°C, followed by spleen, small intestine, thymus, and cecum. The other organs tested showed lower activities. The radioactive product was identified as [ 14 C]palmitic acid by TLC using four different solvent systems. It was likely that the enzyme was derived from blood cells contained in the organs. However, the removal of blood from the lung by the perfusion with saline did not decrease the specific enzyme activity. Furthermore, the leukocyte-rich fraction of rat blood cells showed a low activity (about 0.3 nmol/min/mg protein), and the platelet-rich fraction was almost inactive. Interestingly, we detected a high activity in the solubilized proteins from alveolar macrophages (13.1 Ϯ 1.9 nmol/min/mg protein; n ϭ 3) and peritoneal macrophages (19.6 Ϯ 0.8 nmol/min/mg protein; n ϭ 3).
Characterization of the Lung Enzyme-The subcellular distribution of the enzyme was examined with the lung homogenate. As shown in Table I, 46% of the total activity was recovered in the 12,000 ϫ g pellet, which showed the highest specific activity. When the enzyme was solubilized from the 12,000 ϫ g pellet by two cycles of freezing and thawing without detergent, the yield through this step was 174%, suggesting the presence of endogenous inhibitors for the enzyme in the 12,000 ϫ g pellet. The specific enzyme activity increased 10-fold by the solubilization. The heat-treated enzyme was totally inactive.
The reaction rate of the solubilized enzyme increased depending on the concentrations of N-palmitoylethanolamine, and the apparent K m value was about 35 M (Fig. 2). When the pH value was changed between 3 and 11, the activity was the highest at pH 5.0 and was hardly detectable above pH 8 (Fig. 3). At pH 7.0, the activity was considerably higher with Tris-HCl than with citrate-sodium phosphate. The recombinant rat anandamide amidohydrolase showed the highest activity at pH 9 with N-palmitoylethanolamine as substrate under the same assay conditions (Fig. 3).
The addition of DTT increased the enzyme activity dosedependently up to 6-fold (Fig. 4A). Other sulfhydryl-reducing agents (2-mercaptoethanol, glutathione, and cysteine) showed weakly stimulatory effects. In accordance with this finding, the activity was inhibited almost completely by low concentrations of sulfhydryl blockers, namely 10 M p-chloromercuribenzoic acid and 20 M HgCl 2 . The lung enzyme was hardly inhibited by MAFP at least up to 10 M, whereas MAFP inhibited N-palmitoylethanolamine-hydrolyzing activity of recombinant anandamide amidohydrolase with an IC 50 of 30 nM (Fig. 4B).
We also examined the effect of several detergents on the N-palmitoylethanolamine hydrolysis by the lung enzyme (Fig. 4C). Triton X-100 increased the enzyme activity 7-fold at maximum. The reaction rate was the highest in the presence of 0.1-0.3% of Triton X-100. Interestingly, Tween 20 acted as a potent inhibitor. The enzyme was little affected by CHAPS or octyl glucoside.
The enzymes prepared from peritoneal and alveolar macrophages were also characterized (Fig. 5). The enzymes were almost inactive at pH 9 and were only slightly affected by 1 M MAFP. When DTT or Triton X-100 was removed from the standard reaction mixture, the enzyme activity was markedly reduced. Furthermore, the enzymes were much less active with anandamide. These results strongly suggest that the macrophage enzyme is identical to the lung enzyme.
Purification of the Lung Enzyme-In an attempt to purify the enzyme, the solubilized lung enzyme was subjected sequentially to acid treatment and chromatographies using Phenyl-Sepharose, HiTrap Heparin, Hydroxyapatite, and HiTrap Butyl as described under "Experimental Procedures." Through this procedure the specific enzyme activity was increased 310fold from 5.9 nmol/min/mg protein to 1.8 mol/min/mg protein.
The yield was 0.5% starting from the solubilized protein. The purified enzyme preparation gave a major protein band around 31 kDa as analyzed by SDS-polyacrylamide gel electrophoresis (Fig. 6).
The purified enzyme was allowed to react with various ethanolamides of long-chain fatty acids (Fig. 7). In the presence of 0.1% Triton X-100, N-palmitoylethanolamine was found to be the best substrate for the lung enzyme. Relative reactivity was as follows: N-palmitoylethanolamine, 100%; N-myristoylethanolamine, 48%; N-stearoylethanolamine, 21%; N-oleoylethanolamine, 20%; N-linoleoylethanolamine, 13%; N-arachidonoylethanolamine (anandamide), 8%. When Triton X-100 was removed from the assay mixture, the reactivity with all the ethanolamides of saturated fatty acids was decreased to negligible levels (Fig. 7). In contrast, considerable reactivities with the ethanolamides of unsaturated fatty acids were observed in the absence of Triton X-100. The purified enzyme also required DTT for its full activity, was inactive at pH 9, and was insensitive to 1 M MAFP. Oleamide (30), 2-arachidonoylglycerol (42), and methyl arachidonate (43) were reported to be good substrates of anandamide amidohydrolase. However, the lung enzyme hydrolyzed oleamide only at a low rate (12% of the N-palmitoylethanolamine hydrolysis), and 2-arachidonoylglycerol and methyl arachidonate were almost inactive with this enzyme. The reverse reaction, namely the condensation of 100 M [ 14 C]palmitic acid with 250 mM ethanolamine to form N-palmitoylethanolamine, occurred at only about 2% of the rate of the N-palmitoylethanolamine hydrolysis.
The purified enzyme still showed a ceramidase activity (38 nmol/min/mg protein), which corresponded to about 3% of  the N-palmitoylethanolamine hydrolase activity. However, DTT hardly affected the ceramidase activity, and the activity at pH 9 was as high as that at pH 5. Moreover, because a stimulatory effect of Triton X-100 was also reported for a phosphodiesterase hydrolyzing N-acylphosphatidylethanolamine to N-acylethanolamine and phosphatidic acid (38,41,44), we examined whether the purified enzyme possesses the phosphodiesterase activity. However, the production of N-[ 14 C]palmitoylethanolamine from N-[ 14 C]palmitoyl-1,2-dioleoyl-snglycero-3-phosphoethanolamine was not observed with the purified enzyme. The phosphodiesterase partially purified from rat heart (41) was used as a positive control in this assay.

DISCUSSION
Biological functions of N-acylethanolamines are currently the subjects of active investigations, e.g. anandamide as a cannabimimetic compound (2) and N-palmitoylethanolamine as an anti-inflammatory compound (14,25). Enzymatic hydrolysis of these N-acylethanolamines has been observed in a variety of mammalian organs and cell lines and attributed mostly to an enzyme termed as anandamide amidohydrolase (31,45). As presented in this study, we isolated and purified for the first time an N-acylethanolamine hydrolase that hydrolyzes N-palmitoylethanolamine actively and is distinct from the previously known anandamide amidohydrolase. The enzyme was distributed widely in various mammalian organs and could be efficiently solubilized from the 12,000 ϫ g pellet of the homogenate of rat lung by freezing and thawing without detergent. The enzyme activity was the highest at pH 5, increased by DTT, and hardly inhibited by 10 M MAFP. These catalytic properties were similar to those of the hydrolase, which we recently found in a human megakaryoblastic cell line (CMK), and were clearly distinguished from the properties of anandamide amidohydrolase, which has an optimal pH at 8.5-10, is not activated by DTT, and is inhibited completely by 1 M MAFP (36). Furthermore, the molecular mass of the enzyme purified from rat lung was estimated as 31 kDa. This value was different from that of anandamide amidohydrolase (about 63 kDa) (30).
Substrate specificity was also different between this new enzyme and anandamide amidohydrolase. Among the tested N-acylethanolamines, the lung enzyme was the most active with N-palmitoylethanolamine. N-Palmitoylethanolamine was hydrolyzed 13 times faster than anandamide. This substrate specificity was in contrast with that of anandamide amidohydrolase, which prefers anandamide to N-palmitoylethanolamine (9,28,29,34,35). Thus, we refer to the enzyme with higher reactivity for N-palmitoylethanolamine as N-palmitoylethanolamine hydrolase. The difference in substrate specificity between the two enzymes suggests that anandamide and N-palmitoylethanolamine are degraded in different manners in vivo. In addition, anandamide amidohydrolase was reported to have a high esterase activity for 2-arachidonoylglycerol (42) and methyl arachidonate (43), whereas these compounds were almost inactive with the lung enzyme.
Furthermore, we should note that the distribution of the new enzyme in rat organs is considerably different from that of anandamide amidohydrolase. Namely, the former enzyme is abundant in lung, spleen, small intestine, thymus, and cecum, but shows low activities in liver and brain (Fig. 1) whereas the latter enzyme is highly expressed in liver, small intestine, brain, and testis (26,28,30,31). The different distribution of the two enzymes may reflect their different physiological roles. The abundance of the new enzyme in the organs related to immune system is interesting in the light of anti-inflammatory activity of N-palmitoylethanolamine (14 -17). In relation to this unique organ distribution, we found a high activity of the enzyme in alveolar and peritoneal resident macrophages. Therefore, the enzyme found in lung and immune organs may be derived at least in part from the resident macrophages of each organ. It is known that a large amount of N-acylethanolamine, especially N-palmitoylethanolamine, is produced in degenerating tissues (4,5,10,13). The enzyme of macrophages may be responsible for the degradation of N-acylethanolamines in the degenerating tissues. N-Palmitoylethanolamine, together with smaller amounts of other N-acylethanolamines, was also found in mouse peritoneal macrophages (8).
The enzyme activity with ethanolamides of saturated fatty acids, but not with ethanolamides of unsaturated fatty acids, was increased by Triton X-100. Such different effects of Triton X-100 on the reactivity with different substrates may be explained in terms of solubility of the substrates. It was discussed previously that low solubility of N-palmitoylethanolamine compared with anandamide may cause a low affinity for anandamide amidohydrolase (25). Thus, an appropriate concentration of Triton X-100 may enhance the solubility of N-palmitoylethanolamine and hence its affinity for the enzyme. Interestingly, only Triton X-100 stimulated the N-palmitoylethanolamine hydrolysis among the detergents tested.
As reported earlier, ceramidase was inhibited by N-oleoylethanolamine (46). Because acid ceramidase shows optimal pH around 5, a possibility could not be ruled out that the Npalmitoylethanolamine-hydrolyzing activity was derived from acid ceramidase. Our results clearly showed a different organ distribution between the two enzymes. Previously, Spence et al. (47) also measured the ceramidase activity at pH 5 in the homogenates of various rat organs and showed that the specific activity was higher in the order of kidney Ͼ brain stem Ͼ cerebrum Ͼ cerebellum Ͼ liver Ͼ spleen Ͼ cardiac muscle Ͼ lung Ͼ psoas muscle (47). Furthermore, the molecular mass of the previously reported acid ceramidase (13-and 40-kDa subunits) (48) was different from that of our enzyme (31 kDa). Our purified enzyme still showed a low but detectable ceramidase activity (about 3% of the N-palmitoylethanolamine hydrolase activity). Therefore, the enzyme may also act as a ceramidase that is different from the known acid ceramidase. However, the contamination of ceramidase in the final preparation could not be rigorously ruled out.
In summary, we purified and characterized N-palmitoylethanolamine hydrolase as a catalytically distinct enzyme from the previously known anandamide amidohydrolase and revealed the expression of the enzyme in a variety of rat organs and macrophages. The presence of a hydrolase with N-palmitoylethanolamine as a predominant substrate suggests a physiological or pathophysiological role of this compound.