A Role for Caveolae/Lipid Rafts in the Uptake and Recycling of the Endogenous Cannabinoid Anandamide*

The mechanisms responsible for the uptake and cellular processing of the endogenous cannabinoid anandamide are not well understood. We propose that anandamide uptake may occur via a caveola/lipid raft-related endocytic process in RBL-2H3 cells. Inhibitors of caveola-related (clathrin-independent) endocytosis reduced anandamide transport by (cid:1) 50% compared with the control. Fluorescein derived from fluorescently labeled anandamide colocalized with protein markers of caveolae at early time points following transport. In this study, we have also identified a yet unrecognized process involved in trafficking events affecting anandamide following its uptake. Following uptake of [ 3 H]anandam-ide by RBL-2H3 cells, we found an accumulation of tritium in the caveolin-rich membranes. Inhibitors of both anandamide uptake and metabolism blocked the observed enrichment of tritium in the caveolin-rich membranes. Mass spectrometry of subcellular membrane fractions revealed that the tritium accumulation observed in the caveolin-rich membrane fraction was not representative of intact anandamide, suggesting that following metabolism by the enzyme fatty acid amide hydrolase (FAAH), anandamide metabolites are rapidly enriched in caveolae. Furthermore, HeLa cells, which do not express high levels of found mainly in arachidonic acid-containing plasmenyleth-anolamines, which are the major source of released arachidonic acid (19). Under basal conditions, low levels of NAPE were observed in lipid rafts, but this finding would be expected, as NAPE has been reported to be formed only in response to elevation of intracellular Ca 2 (cid:5) and thus would be found at very low levels under basal conditions (19, 38, 44). We propose that caveolae could serve the function of both regulating the membrane organization of phospholipid precursors of AEA as well as restricting the spatial distribution of the enzymes involved in synthesis and release.

The endocannabinoid anandamide (arachidonylethanolamide (AEA)), 1 a long-chain fatty acid amide, was first shown to be an agonist of the brain and peripheral cannabinoid recep-tors (CB1 and CB2, respectively) in the early 1990s (1,2). AEA has received much attention in the last decade due to its ability to mimic the effects of the plant-derived cannabinoids such as ⌬ 9 -tetrahydrocannabinol, the major active component of marijuana (3)(4)(5). The mechanisms responsible for the biosynthesis and release of AEA as well as the termination of endocannabinoid signaling have not been fully defined (4,5). A better understanding of the process by which endocannabinoid signaling is regulated may identify novel drug targets for regulating the neuromodulatory actions of AEA and provide benefits in the treatment of pain, appetite loss, nausea, asthma, autoimmune disease, arthritis, fever, and glaucoma (4,5).
To terminate endocannabinoid signaling, AEA undergoes a carrier-mediated uptake process that is rapid, temperature-dependent, saturable at 37°C, inhibited by select fatty acid amide derivatives or cannabinoids in a concentration-dependent fashion, and independent of ion gradients or ATP (ATP hydrolysis) (6 -8). The actual protein(s) involved in AEA uptake have not been completely identified, with much debate regarding the existence of a putative AEA transporter (9,10). Even less is understood about the cellular processing and cellular fate of AEA following uptake. Fatty acid amide hydrolase (FAAH) has been shown to carry out the enzymatic hydrolysis of AEA (3,11,12). We (14) and others (13) have suggested that FAAH plays a critical role in maintaining the inward concentration gradient needed for uptake of AEA. However, AEA uptake occurs in cells that do not express FAAH, indicating that mechanisms other than hydrolysis by FAAH are involved in AEA transport (7). In cells that do express FAAH, FAAH is localized to intracellular membrane compartments, suggesting that an efficient means of trafficking to the intracellular compartments containing FAAH exists for rapid metabolism of AEA to take place (9,11,15,16).
The concept of organized trafficking of lipids is not well defined, and the issues surrounding the concept remain somewhat controversial (17). Lipid rafts are specialized plasma membrane microdomains that are enriched in cholesterol, sphingolipids, arachidonic acid, and plasmenylethanolamine (18,19). Related domains, caveolae, are non-clathrin-coated flask-shaped invaginations in the plasma membrane that have been identified by electron microscopy and other biochemical techniques (19,20). The lipid composition of caveolae is similar to that of lipid rafts (19 -21). The existence of caveolae is well substantiated, and they have been implicated in serving many functions, including the organization of key signaling proteins, cholesterol transport, endocytosis, and potocytosis (22).
This study shows that the cellular accumulation of AEA is possibly linked to a caveola-related process. Disruption of caveolae/lipid rafts by cholesterol depletion as well as treat-ment with agents known to inhibit caveola-related endocytic processes reduced [ 3 H]AEA uptake in RBL-2H3 cells. Furthermore, fluorescence derived from a fluorescently labeled AEA analog was colocalized with markers for caveolae following uptake.
We next determined how the intracellular processing of AEA might take place following uptake by utilizing immunocytochemistry studies, biochemical analysis of subcellular fractions, and molecular approaches. Our data show that the metabolites of AEA accumulated in the caveolin-1-rich membrane fraction obtained from sucrose gradient centrifugation. This phenomenon was not observed in cells lacking detectable FAAH. Immunocytochemistry studies of the RBL-2H3 cells confirmed that FAAH is localized to intracellular membrane compartments (11). We propose that AEA is rapidly trafficked to intracellular compartments, where it is metabolized by FAAH, and that following hydrolysis, AEA metabolites undergo recycling to the plasma membrane, where they are enriched in domains containing caveolin-1. This study suggests a new model for AEA uptake that includes endocytic processes and furthermore implicates specialized membrane microdomains in the uptake as well as possible recycling of endocannabinoids.

EXPERIMENTAL PROCEDURES
Cell Culture-RBL-2H3 and HeLa cells were maintained in Dulbecco's modified Eagle's medium and 10% fetal bovine serum supplemented with penicillin, streptomycin, and L-glutamine. Cells were grown in a humidified environment containing 5% CO 2 and held at a constant temperature of 37°C.
[ 3 H]AEA Uptake following Endocytic Inhibitor Treatments-RBL-2H3 cells were plated at 2 ϫ 10 5 cells/well in a 24-well culture plate. Twenty-four hours later, cells were washed with Krebs-Ringer Hepes (KRH) buffer (120 mM NaCl, 4.7 mM KCl, 2.2 mM CaCl 2 , 10 mM Hepes, 1.2 mM KH 2 PO 4 , and 1.2 mM MgSO 4 , pH 7.4) and incubated for 2 h in Opti-MEM I (Invitrogen) containing genistein (200 M) or for 30 min in KRH buffer containing nystatin (25 g/ml)/progesterone (10 g/ml), N-ethylmaleimide (NEM; 500 M), or chlorpromazine (28 M). Cells were also treated with potassium-free KRH buffer for 30 min. Control cells were incubated for 30 min in 1ϫ KRH buffer. AM404 (100 M) was added to the indicated groups for the final 10 min of each incubation. Following the incubations, 1 nM [ 3 H]AEA was added to all samples and allowed to incubate for 5 min at 37°C. Cells were then washed three times with 1ϫ KRH buffer, and Microscint-20 was added to each well. The amount of 3 H present was determined using a Packard TopCount microplate scintillation and luminescence counter.
Immunofluorescent Localization of SKM4-45-1-The AEA analog SKM4-45-1 was synthesized as described by Muthian et al. (23). RBL-2H3 cells (1.5 ϫ 10 5 ) were plated in a 6-well cell culture plate containing a coverslip that had been coated with poly-D-lysine overnight. The next day, the cells were treated with 100 nM SKM4-45-1 for 30 s at 37°C and then immediately fixed for 30 min at room temperature with ice-cold 4% paraformaldehyde in phosphate-buffered saline (PBS). Cells were washed four times with PBS. Primary antibodies directed against caveolin-1 (RDI), flotillin-1 (Transduction Laboratories), and the transferrin receptor (TfR; Pharmingen) were prepared in 0.3% Triton X-100 and 0.5% bovine serum albumin in PBS at dilutions of 1:2000 and added to the appropriate wells. Cells were incubated for 24 h at 4°C. The cells were then washed three times with PBS, and the secondary antibodies were added. The Alexa Fluor 568-conjugated goat anti-mouse antibody (Molecular Probes, Inc.) was added at a 1:1500 dilution. After incubation for 1 h, cells were washed six times with PBS. Coverslips were removed from the wells, mounted on slides using ProLong Antifade, and allowed to dry overnight. Cells were imaged by oil immersion confocal microscopy at ϫ60 magnification with a Nikon Diaphot 300 microscope. A Bio-Rad MRC1024 confocal system was used with a krypton (488 nm)/argon (568 nm) laser, a 522-535-nm band-pass filter, and a 588-nm long-pass filter. Images were produced as an accumulation of three scans.
Subcellular Fractionation of [ 3 H]AEA-treated Cells-Cells were grown ϳ85% confluence in 150-mm culture plates and washed once with 1ϫ KRH buffer. Cells were then incubated for 10 min at 37°C in 100 M AM404, 500 nM methyl arachidonyl fluorophosphonate, or 1ϫ KRH buffer. [ 3 H]AEA was added to each culture dish to yield a final concentration of 1 nM, and culture plates were incubated for 5 min at 37°C. Culture plates were washed three times with ice-cold KRH buffer.
On ice, treated cells were washed with cold KRH buffer and exposed to 2.5 ml of ice-cold lysis buffer (0.5% Triton X-100, 50 mM Tris-HCl, 150 mM NaCl, and 5 mM EDTA) for 20 min. On ice, cell lysates were then mixed with 2.5 ml of 80% sucrose buffer and overlaid with 4 ml of 30% sucrose buffer followed by 2 ml of 5% sucrose buffer in a Beckman Model 344059 centrifuge tube (80, 30, and 5% sucrose solutions were made in lysis buffer minus detergent). The sucrose gradients were centrifuged at 200,000 ϫ g for 18 h at 4°C. Fractions (1 ml) were collected, with the detergent-insoluble caveolin-rich component appearing as a milky white band in approximately the second to third fractions from the top of the gradient. A total of 11 fractions were collected from the gradient. To determine the tritium content of the fractions, 50 l of each fraction was added to individual wells of a Packard 24-well flat-bottom white polystyrene tissue culture microplate. Microscint-20 was added to each well, and tritium was determined using a Packard TopCount microplate scintillation and luminescence counter. Protein content of individual fractions was determined using the Bradford protein assay reagent (Bio-Rad) and used to normalize fractions for protein levels. In subcellular fractionation experiments performed on HeLa cells, cells were transiently transfected in 150-mm culture plates either with rat FAAH cDNA/pBluescript II SK Ϫ (generous gift from Dr. Benjamin Cravatt, Scripps Research Institute) or with vector only using the vaccinia virus T7 expression system (24,25).
Western Blot Analysis-Fractions were prepared for gel electrophoresis by adding a 1:1 volume of Laemmli buffer (5% bromphenol blue, 5% ␤-mercaptoethanol, 62.5 mM Tris-HCl, 20% glycerol, and 2% SDS). Samples were electrophoresed for 50 min at 145 V using the Mini-Protean 3 system (Bio-Rad) after being loaded onto a 10% SDS-polyacrylamide gel. Only fractions 1-9 were analyzed, as previous Western blot analysis revealed that subcellular fractions 10 and 11 taken from RBL-2H3 cells contained no significant levels of the proteins of interest (data not shown). Proteins were then transferred to a polyvinylidene difluoride membrane using the Bio-Rad Mini Trans-Blot system. The membrane was blocked for 24 h in PBS containing 0.1% Tween 20 and 5% dry milk at 4°C. The presence of FAAH, caveolin-1, flotillin-1, the TfR, and BiP/GRP78 was detected using rabbit anti-FAAH polyclonal primary antibody (epitope VGYYETDNYTMPSPAMR), rabbit anticaveolin-1 polyclonal primary antibody (Santa Cruz Biotechnology), mouse anti-flotillin-1 monoclonal primary antibody, mouse anti-TfR monoclonal primary antibody, and mouse anti-BiP/GRP78 monoclonal primary antibody (Transduction Laboratories), respectively, followed either by horseradish peroxidase-labeled goat anti-rabbit secondary antibody (Bio-Rad) or by horseradish peroxidase-labeled goat antimouse secondary antibody and ECL detection reagents (Amersham Biosciences). Membranes were then exposed to x-ray film.
Immunofluorescent Localization of FAAH-RBL-2H3 cells were plated in a 6-well cell culture plate containing a coverslip that had been coated with poly-D-lysine overnight. The next day, the cells were washed two times with PBS and then fixed for 30 min at room temperature with 4% paraformaldehyde. Cells were washed four times with PBS. Primary antibodies directed against FAAH and caveolin-1 were prepared in 0.3% Triton X-100 in PBS at dilutions of 1:3000 and 1:2000, respectively, and added to the appropriate wells. Cells were incubated for 48 h at 4°C. The cells were then washed six times with PBS, and the secondary antibodies were added. The Alexa Fluor 488-conjugated donkey anti-rabbit IgG (Molecular Probes, Inc.) was added at a dilution of 1:4000, and the Alexa Fluor 568-conjugated goat anti-mouse antibody was added a dilution of 1:3000. After incubation for 1 h, cells were washed six times with PBS. Coverslips were removed from the wells, mounted on slides using ProLong Antifade, and allowed to dry overnight. Cells were imaged by oil immersion confocal microscopy at ϫ60 magnification with a Nikon Diaphot 300 microscope. The Bio-Rad MRC1024 confocal system was used with a krypton (488 nm)/argon (568 nm) laser, a 522-535-nm band-pass filter, and a 588-nm long-pass filter.
Determination of Endogenous Cannabinoids by Liquid Chromatography/Tandem Mass Spectrometry (LC/MS/MS)-A quantitative bioanalytical method utilizing LC/MS/MS was used to measure AEA in membrane fractions (26). Following liquid extraction with organic solvent and further purification by solid-phase extraction, samples were separated using a Phenomenex Prodigy ODS-3 column with acetonitrile/water (50:50) as the mobile phase at a flow rate of 0.2 ml/min. Quantification of AEA was accomplished by using m/z 348 ([M ϩ H] ϩ ) as a precursor ion and m/z 62 as a product ion in a selected reaction monitoring mode using [ 3 H 8 ]anandamide as an internal standard. The lower limit of quantification for AEA was 25 pg/ml.

Effect of Endocytic Inhibitors on AEA Uptake-Pretreatment
of cells with nystatin and progesterone disrupts synthesis of cholesterol and cholesterol transport to the membrane, thereby disrupting caveolae (27). To assess the effects of cholesterol depletion and the disruption of caveolae on AEA uptake, RBL-2H3 cells were pretreated with nystatin and progesterone, and [ 3 H]AEA uptake assays were performed (Fig. 1A). Nystatin and progesterone pretreatment reduced the specific uptake of AEA in RBL-2H3 cells by ϳ50%. Genistein and NEM are both known to inhibit caveola-related endocytosis (17,28,29). We pretreated RBL-2H3 cells with genistein or NEM and then, as with the nystatin and progesterone treatment, assessed [ 3 H]AEA transport. Both genistein and NEM were found to reduce AEA uptake similar to the nystatin/progesterone treatment (Fig. 1A). Previous studies have suggested that a 20°C temperature block may interfere with caveola-related trafficking from the plasma membrane to intracellular compartments (30,31). When uptake assays were performed at 18°C, the specific uptake of AEA was again reduced by ϳ50% (Fig. 1A). Treatment of cells with either chlorpromazine or potassiumfree buffer, both of which are known to inhibit clathrin-dependent endocytosis, had no effect on the uptake of AEA by RBL-2H3 cells (Fig. 1B).
Localization of the Fluorescent AEA Derivative SKM4-45-1-The fluorescent AEA derivative SKM4-45-1 has been shown to undergo carrier-mediated uptake presumably via the same process that transports AEA. Once taken up into cells, SKM4-45-1 undergoes rapid cleavage by nonspecific esterases to release fluorescein (23). RBL-2H3 cells were treated with SKM4-45-1, and uptake was allowed to take place for a period of 30 s. This early time point was chosen to provide an indication for the initial entry point of AEA into cells. After fixing SKM4-45-1-treated cells, flotillin-1, caveolin-1, or the TfR was immunofluorescently labeled, and the cells were analyzed by confocal microscopy. Fluorescein derived from SKM4-45-1 appeared to be colocalized with flotillin-1 ( Fig. 2A) and caveolin-1 (Fig. 2B), markers for caveolae, when cells were fixed 30 s after treatment with SKM4-45-1. Fluorescein did not colocalize with the TfR, which is a marker for clathrin-coated pits (Fig. 2C). These data directly imply that the site of SKM4-45-1 and thus AEA entry into the cell may be at caveola/lipid raft domains and support the hypothesis that AEA uptake occurs via a caveolarelated endocytic process.
Accumulation of Tritium in the Caveolin-rich Membrane Fraction of RBL-2H3 Cells following AEA Uptake-Subcellular fractionation of RBL-2H3 cells was performed by sucrose gradient centrifugation (19). A milky white band representing the detergent-insoluble caveolin-rich membrane fraction typically appeared in the second fraction taken from the top of the gradient. Western blot analysis of individual fractions confirmed an enrichment of caveolin-1 in fraction 2 (Fig. 3), suggesting that we had successfully isolated caveolin-rich membranes. Large amounts of the protein flotillin-1 are also found in caveolae/lipid rafts (32). To further verify the isolation of caveola-related membranes, we probed for flotillin-1 and were able to observe the expected enrichment in fraction 2 (Fig. 3). We also verified that both the endoplasmic reticulum marker BiP/GRP78 and the TfR were excluded from the caveolin-1-rich membrane fraction (Fig. 3).
Based upon the endocytic inhibitor data above, we speculated that AEA may accumulate in cells via a caveola-related process. Thus, we determined whether AEA or its metabolites could be isolated in the caveolin-1-rich cellular fractions after uptake. Subcellular fractionation of RBL-2H3 cells was performed by sucrose gradient centrifugation following the uptake of [ 3 H]AEA (AEA was labeled on the arachidonic acid backbone of the molecule). We quantified the amount of tritium present in each fraction collected from the gradient and normalized for the protein concentration of each fraction. Tritium profiles revealed a 20-fold enrichment of tritium in the caveolin-rich membrane fraction compared with AM404-treated cells (Fig.  4A). The anandamide uptake inhibitor AM404 and the FAAH inhibitor methyl arachidonyl fluorophosphonate were both found to inhibit the enrichment of tritium in the caveolin-rich membrane fraction. RBL-2H3 cells pretreated with [ 3 H]AEA labeled on the ethanolamine portion of the molecule also displayed a tritium profile similar to that observed in Fig. 4A (data not shown).
RBL-2H3 cells are known to natively express the serotonin transporter and exhibit robust serotonin uptake (33). Following uptake by RBL-2H3 cells, serotonin should be sequestered in secretory granules and would not be expected to appear in caveolin-containing microdomains. As a control experiment, we performed fractionation experiments on RBL-2H3 cells pretreated with [ 3 H]serotonin to exclude the possibility that the observed tritium accumulation in the caveolin-rich membrane fraction was a nonspecific general phenomenon. We observed no accumulation of serotonin-associated tritium in the caveolin-1-rich membrane fraction when RBL-2H3 cells were treated with [ 3 H]serotonin (Fig. 4B). Cells treated with [ 3 H]serotonin demonstrated an accumulation of tritium that appeared in more dense fractions from the gradient.
To determine whether the tritium accumulation observed in the caveolin-1-rich membrane fraction represented intact AEA, we performed mass spectrometry on samples collected from the fractionation of RBL-2H3 cells that had undergone uptake of AEA at 37°C. RBL-2H3 cells were treated with unlabeled AEA at a concentration equal to [ 3 H]AEA used in Fig. 3 above. Membrane fractions were isolated, and the AEA content in each fraction was determined by LC/MS/MS (Fig. 5). To verify successful isolation of caveolin-rich membranes, these gradient fractions were also analyzed by Western blot analysis in parallel experiments (data not shown). In the non-AM404-treated cells (representing total uptake), we observed only small amounts of intact AEA in the caveolin-rich membranes and the other fractions (Fig. 5). These data indicate that the tritium observed in the caveolin-rich membranes (without AM404) (Fig. 4A) was most likely arachidonate derived from AEA me-tabolism that had been targeted to the caveolae, and not intact AEA. Those fractions isolated from cells treated with AM404, which inhibits both specific uptake and metabolism by FAAH, contained intact AEA at levels that paralleled the observed tritium profile prior to normalization (Fig. 5). Since the tritium profiles in membranes from cells treated with [ 3 H]AEA labeled on the ethanolamine moiety displayed a profile similar to that observed for the arachidonate-labeled AEA (data not shown), both arachidonate and ethanolamine may be enriched in caveolin-rich membranes following the metabolism of AEA.
Trafficking from intracellular compartments to the plasma membrane is inhibited at 18°C (34). Cellular trafficking of the metabolites of AEA to the caveola/lipid raft domains of the plasma membrane following hydrolysis by FAAH should be inhibited if uptake (prior to fractionation) is performed at 18°C. Although specific uptake was reduced at 18°C (Fig. 1), ϳ50% of the uptake remained. RBL-2H3 cells were treated with [ 3 H]AEA at 18°C and then fractionated by sucrose gradient centrifugation. There was no significant difference in the enrichment of tritium found in the caveolin-1-rich membrane fractions of control and AM404-treated cells from 18°C uptake experiments (Fig. 6). This result suggests that cellular trafficking machinery involved in export to the plasma membrane is involved in the enrichment of AEA metabolites in caveolin-rich domains.

Requirement of FAAH for Accumulation of Tritium in Caveolin-rich Membrane
Fractions-Caveolin-rich membrane fractions obtained from RBL-2H3 cells following the uptake of [ 3 H]AEA showed no accumulation of tritium in the presence of the FAAH inhibitor methyl arachidonyl fluorophosphonate (Fig. 4A). We predicted that FAAH activity is necessary for the enrichment of AEA metabolites in the caveolin-1-rich membrane fractions. To test this hypothesis, HeLa cells, which lack detectable FAAH activity (14), were transiently transfected with either FAAH cDNA/pBluescript II SK Ϫ or vector using the vaccinia virus T7 expression system. An enrichment of tritium in the caveolin-1-rich membrane fraction following exposure to [ 3 H]AEA was not observed in HeLa cells transiently transfected with vector only. However, HeLa cells transfected with the FAAH cDNA revealed an accumulation of tritium in the caveolin-1-rich membrane fraction similar to RBL-2H3 cells (Fig. 7). This finding suggests that AEA metabolism must occur for the enrichment of tritium in caveolin-1-rich membrane fractions. Interestingly, both FAAH-and vector-transfected HeLa cells revealed a peak of tritium in fraction 5. This peak was greater in the FAAH-transfected cells, most likely because more AEA accumulated in the cells due to the expression of FAAH. Since this peak was observed in vector-transfected cells, we believe that it represents intact [ 3 H]AEA and that this fraction may contain remnants of an intermediate compartment involved in AEA transport and trafficking.
Exclusion of FAAH from Caveolin-rich Membrane Fractions-Previous studies have shown FAAH to be located in intracellular membrane compartments (9,11). Taking this into consideration, we next determined whether FAAH was present in caveolae, or caveolin-1-rich membrane fractions, thereby indicating whether metabolism of AEA was occurring in the caveolin-rich membrane domains. Western blot analysis of FAAH confirmed that FAAH was not localized to the caveolin-1-rich membrane fraction of RBL-2H3 cells in which an enrichment of AEA metabolites was observed (Fig. 3). Bands representative of FAAH were observed in fractions 7-9. No band representing FAAH was detectable in fraction 2 (Fig. 3). To further investigate the localization of FAAH, we performed immunocytochemistry studies. We labeled both FAAH and caveolin-1 in RBL-2H3 cells and then imaged the cells by confocal microscopy. The overlay of FAAH-labeled (green) cells and caveolin-1-labeled (red) cells revealed that FAAH did not appear to be colocalized with caveolin-1 (Fig. 8), and therefore, we conclude that FAAH is most likely excluded from caveolae and that AEA metabolism occurs in non-caveolin-containing intracellular compartments. DISCUSSION The accumulation of AEA in cells has been shown to be temperature-dependent, saturable at 37°C, and inhibited by select fatty acid amide derivatives (6 -8). The presence of an AEA transporter has been proposed based on these characteristics. However, there has been no success in attempts to identify the putative AEA transporter. It has also been proposed that AEA uptake is the result of simple diffusion and that the necessary concentration gradient for diffusion is driven by the temperature-dependent hydrolysis of AEA by FAAH (9). Even if AEA enters the plasma membrane via diffusion, being a hydrophobic molecule, it should not readily enter the hydrophilic cytoplasm of the cell. The presence of potential binding proteins and the potential for AEA to associate with various intracellular membranes further complicate simple diffusion of AEA. We suggest that for AEA to be rapidly metabolized by FAAH, intracellular AEA would need to be efficiently trafficked to the intracellular domains containing FAAH.
There is emerging evidence for the tight control of lipid trafficking to intracellular compartments via endocytosis (17,35). Endocytic processes meet the criteria for the cellular uptake of AEA that have been used to imply the presence of an actual transporter protein. There are two endocytic pathways that have been identified: clathrin-dependent and clathrinindependent (associated with caveolae and lipid rafts) (36). Both types of endocytosis follow similar cellular pathways. Internalized vesicles are sorted to the early endosomes. Vesicles can then be sent to recycling endosomes for trafficking back to the plasma membrane, late endosomes, or lysosomes for degradation or to the Golgi network and endoplasmic reticulum for further processing. The notion that AEA uptake occurs via an endocytic process does not exclude the possibility that there is an AEA carrier protein. For example, transferrin is taken up by the TfR, which is a marker for clathrin-dependent endocytosis (36). Proteins such as albumin, fatty acidbinding protein, and caveolin-1 have been implied to serve as shuttles in the intracellular trafficking of long-chain fatty acids to various subcellular organelles (37). Interestingly, caveolin-1, the marker for caveolae, has been shown to display properties of a fatty acid-binding protein (37). We explored the possibility that the uptake of AEA may take place via a caveola-related endocytic process. Treatment of cells with nystatin and progesterone inhibits cholesterol synthesis and transport to the membrane, thereby disrupting caveolae/lipid rafts (27). Our data show that nystatin and progesterone pretreatment of RBL-2H3 cells reduced the specific uptake of AEA by ϳ50%, suggesting that a caveola/lipid raft-related process may be involved. However, cholesterol depletion could possibly interfere with clathrin-dependent endocytic processes as well (27). We were able to exclude the possibility that clathrin-dependent processes play a role in AEA uptake by treatment of RBL-2H3 cells with either chlorpromazine or potassium-free buffer, both of which inhibit clathrin-dependent endocytic processes. Neither treatment had any effect on AEA uptake. The tyrosine kinase inhibitor genistein and NEM have both been shown to inhibit caveola-related endocytic processes (17,28,29). Our results show that pretreatment of RBL-2H3 cells with these compounds reduced AEA transport. Each of these drug treatments may have multiple nonspecific effects on our cells. However, we have shown similar results with multiple inhibitors that each inhibit endocytosis by different mechanisms. Taken together, our data implicate a caveola/lipid raft-related endocytic process as being involved in AEA uptake in RBL-2H3 cells. The observation that fluorescein derived from the AEA derivative SKM4-45-1 colocalized with markers for caveolae at early time points following uptake further supports this hypothesis.
The peak of tritium that was observed in the caveolin-1-rich membrane fraction following the uptake of [ 3 H]AEA was not representative of intact AEA. This fact was confirmed by mass spectrometry. Note that the subcellular fractions taken from cells treated with the AEA uptake inhibitor AM404 prior to AEA treatment showed significantly higher levels of intact AEA than those fractions collected from control (total uptake) cells. This result was not surprising because the 100 M AM404 used in these experiments would also act as an inhibitor of FAAH, preventing the metabolism of any AEA that had accumulated in the cell. AM404 could also cause an increase in the levels of intact AEA observed in the caveolin-rich membrane fraction by preventing the internalization of AEA that had partitioned into the caveolae or lipid raft domains. The RBL-2H3 fractionation experiments presented in this study would require FAAH or other enzymatic activity to observe the enrichment of tritium (representing AEA metabolites) in the caveolin-rich membrane fraction. The necessity of FAAH activity for this phenomenon was confirmed by our studies in which we transiently transfected HeLa cells, which express no detectable levels of FAAH, with FAAH cDNA. Without FAAH, AEA metabolites should not be produced, and thus, we would not expect to see an enrichment of tritium in the caveolin-rich membrane fraction. Our data confirm this hypothesis, as only those HeLa cells transfected with FAAH displayed a tritium profile similar to that of RBL-2H3 cells, with an accumulation of tritium found in the caveolin-rich fraction.
Our results support the notion that following metabolism, the AEA metabolites are rapidly trafficked to caveolin-containing membrane microdomains. Prior to the enrichment in caveolin-rich membranes, the AEA metabolites (arachidonic acid and ethanolamine) are most likely incorporated into phospholipids or re-esterified. It would be advantageous for the cell to enrich AEA metabolites in a domain such as caveolae because they may also serve as precursors for the synthesis of new AEA. The widely accepted biosynthetic mechanism for AEA centers FIG. 8. Immunocytochemical localization of FAAH and caveolin-1. Confocal microscopy analysis revealed that FAAH was excluded from the caveolin-1-containing compartments of RBL-2H3 cells. A, FAAH labeling was achieved with rabbit anti-FAAH antibody followed by Alexa Fluor 488-conjugated donkey anti-rabbit IgG (green). B, caveolin-1 was labeled with mouse anti-caveolin-1 antibody (RDI) followed by Alexa Fluor 568-conjugated goat anti-mouse antibody (red). C, the overlay of FAAH and caveolin-1 illustrates the distinct localization of the two proteins. D, shown is a bright-field image of cells.
around the synthesis and release from a membrane precursor, N-arachidonylphosphatidylethanolamine (NAPE) (38). NAPE is formed by the transfer of arachidonate from the sn-1 position of a phospholipid to the primary amine of phosphatidylethanolamine (38). This process is catalyzed by a Ca 2ϩ -dependent N-acyltransferase (6,38,39). The formation and release of AEA occur via a phosphodiesterase-mediated cleavage of NAPE (6,38). The enzymatic cleavage of NAPE takes place in a Ca 2ϩdependent manner and appears to display properties similar to the activity of phospholipase D (38). A novel phospholipase D was isolated from rat heart and brain by Petersen and Hansen (40), and it was suggested that this enzyme might be specific for the production of N-acylethanolamides. Okamoto et al. (41) recently reported the molecular characterization of a phospholipase D isolated from rat heart that is specific for the production of AEA and other N-acylethanolamides. Interestingly, phospholipase D activity has been shown to be enriched in caveolin-rich membranes (42). The biosynthesis of the hydrophobic amide signaling molecule ceramide appears to be compartmentalized to caveolae, establishing a precedence for the localized production of lipid signaling molecules in caveolae (43). Could the metabolites of AEA be available for incorporation into newly synthesized AEA? Certainly, the recycling of re-esterified arachidonic acid could have functions other than to form AEA. This arachidonic acid could be involved in receptor-stimulated arachidonic acid release or the conversion of arachidonic acid by phospholipase A 2 into other eicosanoid signaling molecules such as the prostaglandins. Interestingly, analysis of lipid raft composition revealed that arachidonic acid is found mainly in arachidonic acid-containing plasmenylethanolamines, which are the major source of released arachidonic acid (19). Under basal conditions, low levels of NAPE were observed in lipid rafts, but this finding would be expected, as NAPE has been reported to be formed only in response to elevation of intracellular Ca 2ϩ and thus would be found at very low levels under basal conditions (19,38,44). We propose that caveolae could serve the function of both regulating the membrane organization of phospholipid precursors of AEA as well as restricting the spatial distribution of the enzymes involved in synthesis and release.
Having established that AEA metabolites are enriched in caveolae following hydrolysis by FAAH, it was next necessary to address the question of where AEA metabolism occurs. The presence of AEA metabolites in caveolin-rich membranes could suggest that FAAH may be localized to caveolae. However, previous studies have reported that FAAH is found in intracellular membranes and not the plasma membrane, making it unlikely that FAAH would be present in caveolae (9,16). Both Western blot and immunocytochemistry analyses of the RBL-2H3 cells confirmed that FAAH localization did not correspond with caveolin-1. Thus, following uptake, AEA would be targeted to the intracellular domains containing FAAH. This process remains unclear; however, our data could be interpreted as implicating a caveola-related trafficking process. There is also the possibility that FAAH is trafficked to the domains containing AEA, but we have observed no evidence of AEA-stimulated FAAH redistribution (data not shown).
The findings of this study suggest that the uptake of AEA takes place via a caveola-related endocytic process. Although caveolin-1 is not found in neurons, caveolin-1 is found in brain tissue (possibly glia) (45). Detergent-insoluble membranes lacking caveolin-1 are found in brain tissue (45). These membranes, characterized as lipid rafts, are capable of endocytosis in the absence of caveolin-1 and therefore could be involved in endocannabinoid uptake and signaling in the brain (46).