Abstract
Activation of group I metabotropic glutamate (mGlu) receptors recruits the endocannabinoid system to produce both short- and long-term changes in synaptic strength in many regions of the brain. Although there is evidence that the endocannabinoid 2-arachidonoylglycerol (2-AG) mediates this process, the molecular mechanism underlying 2-AG mobilization remains unclear. In the present study, we used a combination of genetic and targeted lipidomic approaches to investigate the role of the postsynaptic membrane-associated lipase, diacylglycerol lipase type-α (DGL-α), in mGlu receptor-dependent 2-AG mobilization. DGL-α overexpression in mouse neuroblastoma Neuro-2a cells increased baseline 2-AG levels. This effect was accompanied by enhanced utilization of the 2-AG precursor 1-stearoyl,2-arachidonoyl-sn-glycerol and increased accumulation of the 2-AG breakdown product arachidonic acid. A similar, albeit less marked response was observed with other unsaturated and polyunsaturated monoacylglycerols, 1,2-diacylglycerols, and fatty acids. Silencing of DGL-α by RNA interference elicited lipidomic changes opposite those of DGL-α overexpression and abolished group I mGlu receptor-dependent 2-AG mobilization. Coimmunoprecipitation and site-directed mutagenesis experiments revealed that DGL-α interacts, via a PPxxF domain, with the coiled-coil (CC)-Homer proteins Homer-1b and Homer-2, two components of the molecular scaffold that enables group I mGlu signaling. DGL-α mutants that do not bind Homer maintained their ability to generate 2-AG in intact cells but failed to associate with the plasma membrane. The findings indicate that DGL-α mediates group I mGlu receptor-induced 2-AG mobilization. They further suggest that the interaction of CC-Homer with DGL-α is necessary for appropriate function of this lipase.
The endocannabinoids are a family of lipid-derived messengers that are involved in a variety of short-range signaling events in the brain (for review, see Piomelli, 2003). Physiological experiments suggest that these lipid mediators may be released from postsynaptic neurons, diffuse across the synaptic cleft, and bind to CB1 cannabinoid receptors on presynaptic terminals to regulate calcium and potassium channel activities, and inhibit neurotransmitter release (for review, see Freund et al., 2003; Chevaleyre et al., 2006). This retrograde signaling process seems to be widespread in the central nervous system (CNS), but two instances of its occurrence have attracted most attention. The first is a form of short-term synaptic plasticity, termed depolarization-induced suppression of inhibition (DSI), in which depolarization of a postsynaptic cell induces the transient suppression of neurotransmitter release from GABAergic nerve terminals contacting that cell (for review, see Alger, 2002; Freund et al., 2003; Chevaleyre et al., 2006). There is ample evidence that DSI—along with its counterpart at excitatory synapses, DSE—is mediated through the Ca2+-dependent mobilization of an endocannabinoid messenger; however, the chemical identity of this compound remains unclear (Chevaleyre et al., 2006; Szabo et al., 2006). Another important form of endocannabinoid-mediated retrograde signaling is initiated by postsynaptic Gq-coupled receptors. In the hippocampus and other brain regions, activation of group I metabotropic glutamate receptors (mGlu1 and mGlu5) or cholinergic muscarinic receptors depresses inhibitory and/or excitatory neurotransmission, either transiently or persistently, through a CB1-dependent mechanism (Chevaleyre et al., 2006). This receptor-driven form of retrograde signaling is probably due to the postsynaptic release of 2-arachidonoylglycerol (2-AG).
2-AG is the most abundant endocannabinoid substance found in the mammalian CNS (Mechoulam et al., 1995; Sugiura et al., 1995; Stella et al., 1997). It is thought to be produced through phospholipase C (PLC)-catalyzed cleavage of membrane phosphatidylinositol-4,5-bisphosphate (PIP2)to produce 1,2-diacylglycerol (DAG), followed by diacylglycerol lipase (DGL)-mediated hydrolysis of DAG to 2-AG (Scheme 1) (Stella et al., 1997). Newly produced 2-AG is broken down by monoacylglycerol lipase (Dinh et al., 2002, 2004; Muccioli et al., 2007), which is found in excitatory and inhibitory axon terminals (Gulyas et al., 2004). Electrophysiological studies have revealed that genetic deletion of type-β PLC or pharmacological inhibition of DGL activity each abrogates group I mGlu receptor-induced depression of excitatory transmission in the hippocampus (Chevaleyre and Castillo, 2003; Hashimotodani et al., 2005; Maejima et al., 2005). Biochemical experiments have further demonstrated that 1) activation of group I mGlu receptors, most likely mGlu5, stimulates 2-AG formation in hippocampal and corticostriatal slice cultures and 2) PLC or DGL inhibitors prevent this response (Jung et al., 2005). Finally, electron microscopic studies have shown that type-α DGL (DGL-α)—the most abundant of the two DGL isoforms expressed in adult brain tissue (Bisogno et al., 2003)—is localized to dendritic spines of glutamatergic synapses throughout the hippocampus and cerebellum (Katona et al., 2006; Yoshida et al., 2006). It is noteworthy that DGL-α is concentrated in a perisynaptic compartment of the dendritic spine, which is also known to contain mGlu5 receptors (Luján et al., 1996, 1997).
To understand the functions served by 2-AG in retrograde signaling, it is essential to identify the molecular mechanism responsible for its mobilization. In the present study, we have combined genetic and targeted lipidomics approaches to explore the role of DGL-α in mGlu receptor-dependent 2-AG formation. Using the mouse neuroblastoma cell line Neuro2a, which constitutively expresses mGlu5 receptors, we examined whether changes in DGL-α expression alter 2-AG mobilization and influence the cellular profile of 2-AG-related lipids. We also sought to determine whether Homer proteins containing the coiled-coil (CC) domain (Ehrengruber et al., 2004)—which are obligatory components of the postsynaptic mGlu5 receptor complex (Ango et al., 2001; Bockaert et al., 2004; Sala et al., 2005)—regulate the cellular localization and enzymatic activity of DGL-α.
Materials and Methods
Chemicals
(S)-3,5-Dihydroxyphenylglycine (DHPG) was purchased from Tocris (Ellisville, MO). Hexadecanoic acid, heptadecanoic acid (HDA), octadecanoic acid, Δ9-octadecenoic acid, Δ9,12-octadecadienoic acid, Δ9,12,15-octadecatrienoic acid, Δ8,11,14-eicosatrienoic acid, eicosatetraenoic acid, eicosapentaenoic acid, and docosahexaenoic acid were from Nu-Chek Prep (Elysian, MN); 1(3)-palmitoyl-sn-glycerol, 1(3)-heptadecanoyl-sn-glycerol (HDG), 1(3)-stearoyl-sn-glycerol, 1(3)-oleoyl-sn-glycerol, 1(3)-linoleoyl-sn-glycerol, 1(3)-linolenoyl-snglycerol, 1(3)-eicosatrienoyl-sn-glycerol, 1(3)-arachidonoyl-sn-glycerol, and 1(3)-docosahexaenoyl-sn-glycerol were from Nu-Chek Prep; 2-arachidonoyl-sn-glycerol and 2-linoleoyl-sn-glycerol were from Cayman Chemical (Ann Arbor, MI), and 2-oleoyl-sn-glycerol was from Sigma-Aldrich (St. Louis, MO). 1,3-Diheptadecanoyl-sn-glycerol (DHDG) was from Nu-Chek Prep; 1-stearoyl,2-arachidonoyl-snglycerol (SAG), 1-stearoyl,2-linoleoyl-sn-glycerol, 1-oleoyl,2-palmitoyl-sn-glycerol, 1-palmitoyl,2-stearoyl-sn-glycerol, 1,2-dipalmitoylsn-glycerol, and 1,2-dioleoyl-sn-glycerol were from Avanti Polar Lipids (Alabaster, AL). Solvents were purchased from Honeywell Burdick & Jackson (Meskegon, MI).
Plasmids
We amplified the full-length coding sequence of DGL-α by polymerase chain reaction (PCR) using High Fidelity PCR Master Mix (Roche, Indianapolis, IN) and first-strand mouse brain cDNAs as templates. We designed two primers using sequences obtained from the NCBI database (Bisogno et al., 2003): 5′-DGL-α (5′-GTTATGCCCGGGATCGTGGTGTTCCG-3′) and 3′-DGL-α (5′-GCGTGCCGAGATGACCAGGTCATCC-3′). The PCR product was subcloned into a pEF-V5/His vector by TOPO cloning (Invitrogen, Carlsbad, CA) to construct a mammalian expression vector encoding V5-tagged DGL-α. A construct encoding an enhanced green fluorescent protein (EGFP) fusion protein to the C-terminal of DGL-α was cloned in the pEGFP-N2 vector (Clontech, Mountain View, CA) using a SalI site. Constructs encoding Homer isoforms were cloned by PCR amplification using first-strand rat brain cDNAs as template and subsequent subcloning in pFlag-CMV vector (Stratagene, La Jolla, CA) using BamHI and EcoRI sites. The primers used were: Homer1b, 5′-GAAGGATCCATGGGGAACAACCTATCTTCAGCAC-3′ (sense) and 5′GCAGAATTCTTAGCTGCATTCTAGTAGCTTGGCCAA-3′ (antisense) and Homer2, 5′-GAAGGATCCATGGGAGAGCAGCCCATCTTCACC-3′ (sense) and 5′-GCAGAATTCCTAGTTATCCGTGCCTAGCTTGGAGA-3′ (antisense). Site-directed mutagenesis was performed using QuikChange II XL Site-Directed Mutagenesis Kit (Stratagene), following the manufacturer's instructions. The primers used were: DGL-α P975L, 5′AATCTGGTACCCAAGCCCCTCAGGCTCTTTGCAGGATCTG-3′ (sense) and 5′CAGATCCTGCAAAGAGCCTGAGGGGCTTGGGTACCAGATT-3′ (antisense); DGL-α F978R, 5′ACCCAAGCCCCCCAGGCTCCGTGCAGGATCTGCCGAACCC-3′ (sense) and 5′-GGGTTCGGCAGATCCTGCACGGAGCCTGGGGGGCTTGGGT-3′ (antisense). All constructs obtained through PCR amplifications were verified by DNA sequencing.
Cloning and Screening of shRNA Constructs
We generated small-hairpin RNA (shRNA) constructs that contained both a cytomegalovirus promoter-driven GFP and a U6 promoter-driven shRNA expression system. The shRNAs were designed using BLOCK-iT RNAi Designer (https://rnaidesigner.invitrogen.com/rnai-express/), and the corresponding oligonucleotides were synthesized. Sense and antisense oligonucleotides were annealed and ligated into pENTR entry vector (Invitrogen) to generate U6 promoter-shRNA-Pol III terminator cassette. The oligonucleotides used were as follows: 191i, 5′-CACCGCATCCTTCTGAGCTGTATGACGAATCATACAGCTCAGAAGGATGC-3′ (top) and 5′-AAAAGCATCCTTCTGAGCTGTATGATTCGTCATACAGCTCAGAAGGATGC-3′ (bottom); 788i (α-i), 5′-CACCGCTACAAAGAAGTCTGCTACTCGAAAGTAGCAGACTTCTTTGTAGC-3′ (top) and 5′-AAAAGCTACAAAGAAGTCTGCTACTTTCGAGTAGCAGACTTCTTTGTAGC-3′ (bottom); 377i, 5′-CACCGCAATGACCTCACTGCTAAGACGAATCTTAGCAGTGAGGTCATTGC-3′ (top) and 5′-AAAAGCAATGACCTCACTGCTAAGATTCGTCTTAGCAGTGAGGTCATTGC-3′ (bottom); LacZ-i, 5′-CACCGCTACACAAATCAGCGATTTCGAAAAATCGCTGATTTGTGTAG-3′ (top) and 5′-AAAACTACACAAATCAGCGATTTTTCGAAATCGCTGATTTGTGTAGC-3′ (bottom); Unrl-i, 5′-CACCGGATATATCCCGAACTAGACACGAATGTCTAGTTCGGGATATATCC-3′ (top) and 5′-AAAAGGATATATCCCGAACTAGACATTCGTGTCTAGTTCGGGATATATCC-3′ (bottom). RNAi silencing was tested by Western blotting and quantitative PCR using the pENTR constructs. Three different shRNA sequences were tested, and one was selected for further experiments. The selected shRNA expression cassette was amplified by PCR using M13F and M13R primers containing an MluI overhang at both 5′ and 3′ ends. The PCR product was ligated into an adeno-associated virus vector (pAAV-hrGFP; Stratagene) using MluI sites. Constructs were confirmed by DNA sequencing.
Reverse Transcriptase-PCR and Quantitative PCR
We extracted total RNA with TRIzol (Invitrogen) from Neuro-2a cells (American Type Culture Collection, Manassas, VA) cultured for 3 days after transfection, and synthesized first-strand complementary DNA using SuperScript II RNaseH reverse transcriptase (Invitrogen). Reverse transcription of total RNA (0.2 μg) was carried out using oligo(dT)12–18 primers for 50 min at 42°C. For the identification of endogenous DGL-α mRNA, full-length DGL-α was amplified from the first-strand cDNA by using primers 5′-DGL-α (5′-GTTATGCCCGGGATCGTGGTGTTCCG-3′) and 3′-DGL-α (5′-GCGTGCCGAGATGACCAGGTCATCC-3′). Real-time quantitative PCR was conducted using an Mx3000P system (Stratagene). We designed primer/probe sets with Primer Express software (Applied Biosystems, Foster City, CA) and mouse DGL-α sequence obtained from NCBI database (GenInfo Identifier 37674270). Primers and fluorogenic probes were synthesized by TIB Molbiol (Adelphia, NJ). DGL-α mRNA levels were normalized using glyceraldehyde-3-phosphate dehydrogenase as an internal standard. The primer/probe sequences for DGL-α were: forward, 5′-CCAGGCCTTTGGGCG-3′; reverse, 5′GCCTACCACAATCAGGCCAT-3′; and TaqMan probe, 5′-ACCTGGGCCGTGGAACCAAACA-3′.
Cell Transfections and Lipid Extractions
We cultured Neuro-2a cells (passage in culture: 4–20 after purchase from American Type Culture Collection) at 37°C with 5% CO2 in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen) supplemented with 10% fetal bovine serum (fetal bovine serum; Invitrogen) and antibiotics (100 unit/ml Penicillin G and 100 μg/ml Streptomycin; Gibco-Invitrogen). We transfected cells at approximately 30% confluence using Superfect reagent (QIAGEN, Valencia, CA) and incubated them for 72 h at 37°C with 5% CO2 in DMEM supplemented with 10% fetal bovine serum and antibiotics. In shRNA-induced silencing experiments, transfection efficiency was approximately 50%, based on GFP monitoring. We used shRNAs designed against unrelated sequences, LacZ-i and Unrl-i, as controls (Invitrogen). Cells were washed with ice-cold phosphate-buffered saline and scraped into 1 ml of methanol/water [1:1 (v/v)]. Protein concentration was measured using the BCA protein assay (Pierce, Rockford, IL). Lipids were extracted using a chloroform/methanol mixture (2:1, 1.5 ml) containing DHDG, HDG, and HDA as internal standards (100 pmol/sample). The organic phase was collected, dried under N2 and dissolved in methanol/chloroform (3:1) for analyses. Extraction recovery efficiencies were measured from spiked biological samples compared with direct injection of nonextracted lipid standards into high-performance liquid chromatography/mass spectrometry (LC/MS): DAG, 72 ± 5%, monoacylglycerols, 75 ± 3%; and fatty acids, 95 ± 2%.
Lipid Analyses
Monoacylglycerols. We used an Agilent 1100-LC system (Agilent Technologies, Palo Alto, CA) coupled to a 1946A-MS detector equipped with an electrospray ionization (ESI) interface (Agilent Technologies). MAGs were separated on a XDB Eclipse C18 column (50 × 4.6 mm i.d., 1.8 μm; Zorbax; Agilent Technologies). They were eluted with a gradient of methanol in water (from 85% to 90% methanol in 2.0 min and 90% to 100% in 3.0 min) at a flow rate of 1.5 ml/min. Column temperature was kept at 40°C. MS detection was in the positive ionization mode, capillary voltage was set at 3 kV, and fragmentor voltage was 120 V. N2 was used as drying gas at a flow rate of 13 liters/min and a temperature of 350°C. Nebulizer pressure was set at 60 psi. Commercial MAGs were used as reference standards. For quantification purposes, we monitored the Na+ adducts of the molecular ions [M+Na]+ in the selected ion-monitoring (SIM) mode, using HDG (m/z 367) as an internal standard. Calibration curves were generated using commercial MAGs. Limit of quantification was 1.25 pmol.
Fatty Acids. We used a reversed-phase XDB Eclipse C18 column (50 × 4.6 mm i.d., 1.8 μm; Zorbax; Agilent Technologies) eluted with a linear gradient from 90% to 100% of A in B for 2.5 min at a flow rate of 1.5 ml/min with column temperature at 40°C. Mobile phase A consisted of methanol containing 0.25% acetic acid and 5 mM ammonium acetate; mobile phase B consisted of water containing 0.25% acetic acid and 5 mM ammonium acetate. ESI was in the negative mode, capillary voltage was set at 4 kV, and fragmentor voltage was 100 V. N2 was used as drying gas at a flow rate of 13 liters/min and a temperature of 350°C. Nebulizer pressure was set at 60 psi. We used commercially available fatty acids as reference standards, monitoring deprotonated molecular ions [M-H]– in the SIM mode. HDA (m/z = 269) was the internal standard. Calibration curves were generated using commercial fatty acids. Limit of quantification was 0.5 pmol.
DAGs. We used an Agilent 1100-LC system coupled to a MS detector Ion-Trap XCT interfaced with ESI (Agilent Technologies). DAGs were separated using a reversed-phase Poroshell 300SB C18 column (2.1 × 75 mm i.d., 5 μm; Agilent). We applied a linear gradient from 60% to 100% of mobile phase A in B in 5 min at a flow rate of 1.0 ml/min with column temperature at 30°C. Mobile phase A consisted of methanol containing 0.25% acetic acid, 5 mM ammonium acetate; mobile phase B consisted of water containing 0.25% acetic acid, 5 mM ammonium acetate. The capillary voltage was set at 4.0 kV and skimmer voltage at 40 V. N2 was used as drying gas at a flow rate of 12 liters/min, temperature at 350°C, and nebulizer pressure at 80 psi. Helium was used as collision gas, and fragmentation amplitude was set at 1.2 V. MS detection was in the positive ionization mode. DAGs were identified based on their retention times and MS3 properties, using synthetic standards as references. Multiple reaction monitoring was used to acquire full-scan tandem MS spectra of selected DAG ions. Extracted ion chromatograms were used to quantify each DAG by selecting appropriate precursor>product ions. DHDG (m/z 619.8 > 349.5) was the internal standard. Calibration curves were generated using SAG. Limit of quantification was 1 pmol.
DGL Assay
We harvested Neuro-2a cells in 50 mM Tris-HCl (pH 7.0, 1 ml/dish) and homogenized them with a Dounce homogenizer on ice. Homogenates were centrifuged at 800g for 5 min at 4°C, and the resulting supernatants were used for the assay. DGL activity was measured at 37°C for 30 min in 50 mM Tris-HCl, pH 7.0, containing 0.1% Triton X-100, tissue homogenate protein (100 μg) and DHDG (50 μM) as substrate. The reactions were stopped by adding chloroform-methanol (1:1) containing [2H8]-2-AG. After centrifugation at 1500g at 4°C for 5 min, the organic layers were collected and dried under N2. The residues were suspended in chloroform/methanol (1:3, 50 μl) and analyzed by LC/MS. For quantification purposes, we monitored the reaction product HDG [M+Na]+ (m/z 367) along with standard [2H8]2-AG (m/z 409).
Immunostaining and Western Blotting
We prepared lysates from Neuro-2a cells in a buffer containing Tris-HCl (10 mM, pH 7.4), NaCl (150 mM), Triton X-100 (1%), Nonidet P-40 (0.25%), and EDTA (2 mM) supplemented with a mixture of protease inhibitors (Roche). Lysates were centrifuged at 14,000g for 10 min. Proteins (20 μg) from supernatants were separated by 4–20% SDS-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes, and subjected to Western blotting. We used the following antibodies: monoclonal anti-V5 (1: 5000; Invitrogen), polyclonal anti-V5 (1:3000; Covance, Berkeley, CA), monoclonal anti-Flag M2 (1:3000; Sigma-Aldrich, St. Louis, MO), and polyclonal anti-DGL-α-int (1:2000; Katona et al., 2006). We cultured Neuro-2a cells on Lab-Tek chamber slides (Nalge Nunc International, Roskilde, Denmark), fixed them, and immunostained them as described previously (Jung et al., 2005), using monoclonal anti-Flag M2 (1:1000) and polyclonal anti-V5 (1:1000) as primary antibodies. Alexa 488-labeled anti-mouse or Alexa 546-labeled antirabbit secondary antibodies (1:1000; Invitrogen) were used for detection, which was conducted using a Eclipse E600 fluorescence microscope (Nikon, Japan) equipped with a digital camera (Diagnostic Instruments, Sterling Heights, MI).
Immunoprecipitation
We lysed the cells on ice with a buffer composed of 50 mM Tris-HCl, pH 7.0, 1 mM EDTA, 150 mM NaCl, and 1% Triton X-100 supplemented with protease inhibitor mixture (Roche). The lysates were centrifuged at 14,000g for 10 min at 4°C, and the supernatants (0.6 ml, 1 mg/ml protein) were incubated with monoclonal anti-Flag M2 antibody (1:150; Sigma-Aldrich) overnight at 4°C. The immunocomplexes were precipitated by incubation with protein G-Sepharose (40 μl; GE Healthcare, Piscataway, NJ) for 1 h at room temperature. The immunoprecipitates were washed five times with buffer, eluted with SDS sample buffer, and analyzed by Western blotting.
Statistical Analyses
Results are expressed as means ± S.E.M. Analyses were conducted with GraphPad Prism ver. 4.0, and differences were considered significant if P < 0.05 (Student's t test). For DGL-α silencing experiments shown in Fig. 3, Table 1, and Table 3, one-way analysis of variance (ANOVA) followed by Dunnett's multiple comparison test was used. Differences were considered significant if comparison with both controls gives significant changes. A representative result from at least three independent experiments was shown. N represent the number of replicates within a single experiment.
Results
DGL-α Overexpression Increases Cellular 2-AG Content. To investigate the role of DGL-α in 2-AG mobilization, we transiently overexpressed V5-His-tagged DGL-α in mouse Neuro-2a cells. We confirmed DGL-α expression using either an anti-V5 antibody (Fig. 1A, top) or a polyclonal antibody generated against a DGL-α-GST fusion protein (Fig. 1A, middle) measuring actin as a loading control (Fig. 1A, bottom). Cells transfected with DGL-α contained significantly higher 2-AG levels than did pEF vector-transfected cells (P < 0.001, Student's t test) (Fig. 1B and Table 1). Levels of other unsaturated and polyunsaturated MAGs were also elevated in Neuro-2a-DGL-α cells compared with pEF vector-transfected controls (Table 1). By contrast, the cellular content of saturated MAGs remained unchanged (Table 1). It is noteworthy that MAG levels in cells transfected with either the empty pEF vector or the pEF-DGL-α vector were slightly different from those of nontransfected cells (Table 1): for example, 18:1 Δ9 MAG was significantly lower in pEF-transfected cells than in nontransfected ones (P < 0.001, Student's t test). We attribute this difference, which was consistently observed in our experiments, to changes induced by transfection with the pEF vector. These minor differences notwithstanding, 2-AG levels were significantly higher in DGL-α-transfected cells than nontransfected ones (P < 0.001, Student's t test).
To identify the substrates used by DGL-α to produce 2-AG (Scheme 1), we profiled DAG species in Neuro-2a cells by LC coupled to an ion trap MS (LC/MSn). Under our analytical conditions, most endogenous DAGs were eluted from the LC column with retention times of 5 to 6 min (Fig. 2A) and yielded informative fragmentation patterns by collision-induced dissociation of their Na+ adducts (Fig. 2B). Such patterns were used to identify and quantify cellular 1-acyl,2-arachidonoyl-sn-glycerol species, which were expected to serve as metabolic precursors for 2-AG. Significant levels of SAG were found in unstimulated Neuro-2a cells, where this DAG species accounted for approximately 84% of all 1-acyl,2-arachidonoyl-sn-glycerols present (Table 2). It is noteworthy that SAG content was significantly reduced in cells overexpressing DGL-α, indicating an enhanced breakdown of this 2-AG precursor (Fig. 1C).
If DGL-α cleaves SAG to generate 2-AG, unesterified stearic acid should transiently accumulate in cells. Consistent with this prediction, stearic acid levels were significantly higher in Neuro-2a-DGL-α than control cells (Fig. 1D). Other fatty acids derived from the DGL-catalyzed hydrolysis of DAGs (or from other lipolytic pathways such as the monoacylglycerol lipase-catalyzed hydrolysis of MAGs) were increased by DGL-α overexpression (Table 3). The results indicate that 1) DGL-α overexpression increase 2-AG levels in Neuro-2a cells; and 2) unstimulated Neuro-2a cells contain significant amounts of preformed DAGs, which are used by DGL-α to produce 2-AG.
DGL-α Silencing Reduced Cellular 2-AG Content. Next, to examine whether DGL-α silencing affects 2-AG mobilization, we engineered three distinct shRNAs and screened their ability to suppress heterologous DGL-α expression in HEK293 cells. One of the shRNAs tested, 788i (α-i), produced a marked decrease in DGL-α levels (Fig. 3A). This shRNA was selected for further studies. Two control shRNAs were used, each containing an oligodeoxynucleotide sequence that is not present in the mouse or human genome. The first (LacZ-i) targeted the bacterial lacZ gene, whereas the second (Unrl-i) consisted of a random sequence of nucleotides that did not match any existing sequence in the human or mouse genome.
Neuro-2a cells constitutively express detectable, albeit low levels of DGL-α mRNA (Fig. 3B). Incubation of these cells with shRNA α-i caused a significant decrease in DGL-α mRNA (Fig. 3C). No such effect was observed in cells transfected with the two control shRNA (Fig. 3C). DGL-α silencing lowered DGL activity (Fig. 3D) and produced a set of lipidomic changes opposite those caused by DGL-α overexpression: it decreased cellular levels of 2-AG (Fig. 3E and Table 1), increased levels of the 2-AG precursor SAG (Fig. 3F), and decreased levels of unesterified stearic acid (Fig. 3G) and other fatty acids (Table 3). By contrast, no such changes in 2-AG levels were observed in Neuro-2a cells expressing control LacZ-i shRNA (Fig. 3D-G) or Unrl-i shRNA (Fig. 3E). Thus, DGL-α activity may contribute in an important way to constitutive 2-AG mobilization in Neuro-2a cells.
DGL-α Silencing Impairs mGlu Receptor-Dependent 2-AG Mobilization. Previous studies have shown that the group I mGlu receptor agonist DHPG stimulates 2-AG formation in organotypic cultures of rat brain slices (Jung et al., 2005). A similar effect was observed in Neuro-2a cells expressing control LacZ-i shRNA (Fig. 4A) as well as nontransfected cells (data not shown). Consistent with this finding, Western blot analyses demonstrated that Neuro-2a cells contained both mGlu5 receptors and PLC-β1 (Fig. 4B). Under the same conditions, no mGlu1 receptor immunoreactivity was detected (Fig. 4B). To examine whether DGL-α participates in mGlu receptor-mediated generation of 2-AG, we tested the effects of DHPG (100 μM) in Neuro-2a cells in which DGL-α had been silenced with shRNA α-i. DHPG did not change 2-AG content in these cells (Fig. 4A), suggesting that DGL-α is required for mGlu5 receptor-dependent 2-AG mobilization.
Subcellular Localization of DGL-α. We expressed a DGL-α-GFP fusion protein in either Neuro-2a or HEK293 cells and visualized it using fluorescence microscopy. In contrast to GFP, which was associated with the cytosol, DGL-α-GFP was primarily found on the cell surface (Fig. 5, A and B). These findings confirm previous observations indicating that DGL-α is localized to the neuronal plasma membrane (Katona et al., 2006; Yoshida et al., 2006).
DGL-α Interacts with CC-Homer Proteins. A bioinformatics search revealed that the cytoplasmic C-terminal end of DGL-α contains the PPxxF motif, which is characteristic of Homer-binding proteins such as mGlu1, mGlu5, or IP3 receptors (Fig. 6A) (Brakeman et al., 1997; Xiao et al., 2000). To determine whether DGL-α interacts with CC-Homer proteins, we conducted coimmunoprecipitation studies in HEK293 cells that heterologously expressed DGL-α along with Homer-2 or Homer-1b. We found that DGL-α coimmunoprecipitates with either protein (Fig. 6, B and C). As previously shown for other Homer-binding proteins (Tu et al., 1998), such coimmunoprecipitation persisted in the presence of a high concentration of the nonionic detergent Triton X-100 (1%). It is noteworthy that single point mutations of proline (P975L) and phenylalanine (F978R) in the PPxxF motif of DGL-α, which are known to prevent the interaction of group I mGlu receptors with CC-Homer (Beneken et al., 2000), abolished DGL-α/Homer-1a coimmunoprecipitation (Fig. 6C), suggesting that the two proteins interact selectively through this domain.
To determine whether Homer-2 colocalizes with DGL-α in cells, we visualized both proteins using fluorescent doubleimmunostaining. As illustrated in Fig. 7A, Homer-2 was associated with plasma membranes and endoplasmic-like structures in the cytosol. Merged images showed that DGL-α was localized to plasma membranes together with Homer-2 (Fig. 7A). In contrast, the DGL-α mutant P975L, which is unable to bind Homer (Fig. 6C), remained associated with intracellular structures (Fig. 7A).
We next asked whether CC-Homer proteins affect the ability of DGL-α to mobilize 2-AG in intact cells. Coexpression of DGL-α with Homer-1b or Homer-2 produced only small increases in cellular 2-AG content that were not affected by the P975L mutation (Fig. 7B). We interpret these results to indicate that, in our heterologous expression system, the interaction of CC-Homer with DGL-α is required for the membrane localization of this lipase, but not for its enzymatic activity.
Discussion
Group I mGlu receptors regulate excitatory transmission throughout the CNS by releasing a diffusible signal that leaves the dendritic spine, crosses the synaptic cleft, and activates CB1 cannabinoid receptors on adjacent glutamatergic axon terminals (for review, see Chevaleyre et al., 2006). Although biochemical and pharmacological evidence suggests that 2-AG may be the messenger—or at least one of the messengers—responsible for this receptor-operated retrograde signaling process (Jung et al., 2005; Maejima et al., 2005), many aspects of the molecular mechanism governing 2-AG mobilization remain unknown. The present findings identify DGL-α, a serine lipase localized to the plasma membrane of glutamatergic dendritic spines (Bisogno et al., 2003; Katona et al., 2006; Yoshida et al., 2006), as a key component of the enzymatic machinery responsible for mGlu receptor-dependent 2-AG mobilization. The results further show that DGL-α selectively interacts with the CC-Homer proteins Homer-1b and Homer-2.
Group I mGlu receptors, consisting of the mGlu 1 and mGlu 5 subtypes, are Gq-coupled receptors that are concentrated in the perisynaptic compartment of glutamatergic dendritic spines (Nusser et al., 1994; Luján et al., 1996, 1997). They stimulate PLC-β to cleave membrane PIP2 and produce inositol-1,4,5-trisphosphate (IP3)—which binds to IP3 receptors and triggers calcium release from intracellular stores—and DAG—which activates a variety of intracellular targets, such as protein kinase C and C-type transient receptor potential cation (TRPC) channels (Hofmann et al., 1999). Within the spine, group I mGlu receptors combine in a multiprotein complex that includes two of their downstream signaling effectors, IP3 receptors (Tu et al., 1998) and TRPC-1 channels (Kim et al., 2003). A polymeric scaffold of CC-Homer proteins holds this complex together and connects it to other components of the postsynaptic density (Brakeman et al., 1997).
Our RNAi experiments suggest that silencing of recombinant DGL-α expression in Neuro-2a cells reduces total DGL activity, decreases baseline 2-AG levels, and abrogates group I mGlu-dependent 2-AG mobilization. These results are consistent with previous studies showing that the DGL inhibitor RHC80267 prevents mGlu5-induced 2-AG formation (Jung et al., 2005) and indicate that DGL-α may be responsible for the release of this endocannabinoid messenger. Moreover, our immunoprecipitation experiments show that DGL-α selectively interacts with CC-Homer proteins. This interaction seems to be functionally important, in that single-nucleotide mutations that prevent DGL-α binding to CC-Homer also disrupt the localization of DGL-α to the plasma membrane. Although we used a transfected cell line to test our hypothesis, this result acquires significance in light of the neuroanatomical distribution of DGL-α, which is localized to the same presynaptic compartment of the dendritic spine that contains the group I mGlu receptor complex (Nusser et al., 1994; Luján et al., 1996, 1997; Katona et al., 2006; Yoshida et al., 2006). A functional association of mGlu receptors with DGL-α would enable this lipase to become rapidly activated in response to receptor occupation, and generate significant amounts of 2-AG in close proximity to presynaptic CB1 receptors (Katona et al., 2006; Yoshida et al., 2006). Furthermore, the localization of DGL-α may account for the differential contribution of mGlu5 and muscarinic acetylcholine receptor to endocannabinoid-mediated plasticity, as recently observed at synapses of striatal medium spiny neurons (Uchigashima et al., 2007).
The LC/MSn analyses conducted here, which included several key substrates and products of DGL-α-catalyzed DAG hydrolysis, allowed us to obtain a broad view of the lipidomic changes induced by DGL-α overexpression or silencing. Two noteworthy points emerge from these analyses. First, unsaturated MAGs, such as 2-AG and, to a minor extent, 2-oleoylglycerol, are the primary products of DGL-α activity in intact cells. This product selectivity, which was not observed in broken cell preparations (Bisogno et al., 2003), may result from the availability in intact Neuro-2a cells of substrate DAG species containing arachidonic and oleic acid at the sn-2 position. Second, unstimulated Neuro-2a cells contain significant amounts of preformed DAGs, such as SAG, which may serve as 2-AG precursors. In separate experiments, we have also found high SAG levels in various regions of snap-frozen rat brain, including the amygdala (6.6 ± 0.3 nmol/g; n = 3) and the cerebellum (0.6 ± 0.02 nmol/g; n = 3) (G. Astarita and D. Piomelli, unpublished observations), where SAG is presumably derived from PIP2 hydrolysis (Wenk et al., 2003). These results suggest that elevation in intracellular Ca2+ levels in dendritic spines might initiate 2-AG mobilization through a direct activation of DGL-α, rather than through a traditional route involving the sequential activation of PLC-β and DGL-α. Consistent with this possibility, electrophysiological experiments have shown that DHPG-induced suppression of inhibition in the hippocampus and DSI in the cerebellum is inhibited by the DGL inhibitors RHC80267 and tetrahydrolipstatin but is not affected by the PLC inhibitor U73122 (Edwards et al., 2006; Szabo et al., 2006).
Our experiments have not addressed the possible functions served by DGL-β, a second DGL isoform present in the mammalian CNS (Bisogno et al., 2003). Clear developmental and structural differences exist between these two isoforms (Bisogno et al., 2003). It is noteworthy that DGL-β lacks the CC-Homer-binding motif present in DGL-α and, when expressed in Neuro-2a cells, it associates with intracellular structures rather than with the plasma membrane (K.-M. Jung and D. Piomelli, unpublished observations). However, the significance of these differences to neuronal 2-AG mobilization remains to be determined. This question notwithstanding, the present results provide new evidence for a primary role of DGL-α in mGlu5 receptor-dependent 2-AG mobilization and highlight the value of targeted lipidomic analyses in the study of brain endocannabinoid signaling.
Acknowledgments
We thank Dr. Faizy Ahmed for helpful discussion and Christopher Stapleton and Mariam Behbehani for their invaluable experimental help. The contribution of the Agilent Technologies/University of California, Irvine Analytical Discovery Facility, Center for Drug Discovery is gratefully acknowledged.
Footnotes
-
This work was supported by National Institute on Drug Abuse grants DA12447 and DA12413 (to D.P.) and DA11322 and DA00286 (to K.M.). K.-M.J. and G.A. contributed equally to this work.
-
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
-
doi:10.1124/mol.107.037796.
-
ABBREVIATIONS: CNS, central nervous system; DSI, depolarization-induced suppression of inhibition; 2-AG, 2-arachidonoylglycerol; PLC, phospholipase C; PIP2, phosphatidylinositol-4,5-bisphosphate; DAG, 1,2-diacylglycerol; DGL, diacylglycerol lipase; DGL-α, diacylglycerol lipase type-α; CC, coiled-coil; DHPG, (S)-3,5-dihydroxyphenylglycine; PCR, polymerase chain reaction; EGFP, enhanced green fluorescent protein; shRNA, small-hairpin RNA; RNAi, RNA interference; DMEM, Dulbecco's modified Eagle's medium; GFP, green fluorescent protein; DHDG, diheptadecanoyl-sn-glycerol; HDG, 1(3)-heptadecanoyl-sn-glycerol; HDA, heptadecanoic acid; mGlu, metabotropic glutamate; LC/MS, liquid chromatography/mass spectrometry; ESI, electrospray ionization; MAG, monoacylglycerol; SIM, selected ion monitoring; SAG, 1-stearoyl,2-arachidonoyl-sn-glycerol; HEK, human embryonic kidney; IP3, inositol-1,4,5-trisphosphate; RHC80267, 1,6-bis(cyclohexyloximinocarbonylamino)hexane; U73122, 1-(6-((3-methoxyestra-1,3,5(10)-trien-17-yl)amino)hexyl)-1H-pyrrole-2,5-dione.
- Received May 4, 2007.
- Accepted June 21, 2007.
- The American Society for Pharmacology and Experimental Therapeutics