Modulation of Glucagon-like Peptide-1 (GLP-1) Potency by Endocannabinoid-like Lipids Represents a Novel Mode of Regulating GLP-1 Receptor Signaling*

Background: Glucagon-like peptide-1 receptors (GLP-1Rs) are expressed in many tissues that are accessed only by a basal circulating level of GLP-1. Results: GLP-1 potency is enhanced by specific endocannabinoid-like lipids. Conclusion: Enhancing GLP-1 potency is a novel mechanism regulating GLP-1R signaling. Significance: Spatiotemporal regulation of GLP-1R becomes possible at basal physiological levels of GLP-1 because endocannabinoid-like lipids are known to be physiologically regulated. Glucagon-like peptide-1 (GLP-1) analogs are approved for treatment of type 2 diabetes and are in clinical trials for disorders including neurodegenerative diseases. GLP-1 receptor (GLP-1R) is expressed in many peripheral and neuronal tissues and is activated by circulating GLP-1. Other than food intake, little is known about factors regulating GLP-1 secretion. Given a normally basal circulating level of GLP-1, knowledge of mechanisms regulating GLP-1R signaling, which has diverse functions in extrapancreatic tissues, remains elusive. In this study, we found that the potency of GLP-1, not exendin 4, is specifically enhanced by the endocannabinoid-like lipids oleoylethanolamide (OEA) and 2-oleoylglycerol but not by stearoylethanolamide (SEA) or palmitoylethanolamide. 9.2 μm OEA enhances the potency of GLP-1 in stimulating cAMP production by 10-fold but does not affect its receptor binding affinity. OEA and 2-oleoylglycerol, but not SEA, bind to GLP-1 in a dose-dependent and saturable manner. OEA but not SEA promoted GLP-1(7–36) amide to trypsin inactivation in a dose-dependent and saturable manner. Susceptibility of GLP-1(7–36) amide to trypsin inactivation is increased 40-fold upon binding to OEA but not to SEA. Our findings indicate that OEA binds to GLP-1(7–36) amide and enhances the potency that may result from a conformational change of the peptide. In conclusion, modulating potency of GLP-1 by physiologically regulated endocannabinoid-like lipids allows GLP-1R signaling to be regulated spatiotemporally at a constant basal GLP-1 level.

established therapeutic target for type 2 diabetes. In addition to human pancreatic islet ␤ cells, GLP-1R is expressed in a wide array of tissues, including lung, heart, kidney, blood vessels, neurons, and lymphocytes (1)(2)(3)(4). Mice deficient in GLP-1R expression or with blunted GLP-1R function show impairment of physiologic features not limited to glucose homeostasis but also including learning and memory (4). Clinical trials targeting GLP-1 signaling to treat non-metabolic diseases include those for psoriasis, heart disease, and neurodegenerative diseases (5)(6)(7). Despite encouraging outcomes with GLP-1 analogs in reducing myocardial infarct size in acute coronary occlusion (7) and improving clinical symptoms in patients with Parkinson disease (5), the mechanisms of physiological regulation of GLP-1R signaling beyond energy homeostasis remain largely unknown.
GLP-1 is an incretin peptide hormone derived from posttranslational processing of the precursor proglucagon in intestinal L cells (8). On food intake, the biologically active forms of GLP-1  amide and GLP-1  are secreted, thus increasing the basal plasma level by 3-4-fold, to maintain normoglycemia by enhancing glucose-dependent insulin secretion and suppressing glucagon function (8,9). Circulating GLP-1 has a short plasma half-life of only a few min due to renal clearance after rapid enzymatic inactivation by a plasma enzyme, dipeptidyl peptidase 4 (10). Other cells outside of the gut shown to produce GLP-1 include pancreatic ␣ cells and neurons in the localized area of the brain stem (4,(11)(12)(13)(14), but our knowledge of the physiological regulation of GLP-1 secretion by these cells is limited.
In the brain, GLP-1 is synthesized primarily by a discrete group of neurons located in the nucleus of the solitary tract (12). These neurons send abundant projections to other regions of the brain, including the forebrain, hypothalamus, amygdala, stria terminalis, and thalamus, where GLP-1Rs are expressed; this neuronal circuit of GLP-1 signaling is considered relevant to satiety and energy homeostasis (2,11). GLP-1R is also expressed in neurons in the hippocampus (1) and dopaminergic neurons in substantia nigra (3), where no known GLP-1-secreting neuron innervation is found (3,14). It has been suggested that the basal circulating GLP-1 level is the primary source of ligands accessible to GLP-1Rs in these brain regions and probably in the heart as well. Therefore, determining the mechanism by which the basal level of GLP-1 can activate receptors in these brain regions is germane.
The well delineated functions of GLP-1 are mainly mediated by activation of GLP-1R (4). GLP-1R, as a member of the class B G protein-coupled receptor (GPCR) family, is the only known receptor with high specific affinity for GLP-1. GLP-1R activation leads to two major signaling pathways, namely G␣ s coupling and recruitment of ␤-arrestin to the agonist-occupied receptor; the former mainly leads to activation of adenylyl cyclase, with subsequent generation of cAMP (15), and the latter leads to receptor endocytosis and activation of extracellular signal-regulated kinase 1/2 (ERK1/2) signaling (4). In pancreatic ␤ cells, the increased cAMP level is responsible for glucosedependent insulin release (16) and contributes to maintaining glucose homeostasis. Thus, cAMP production is measured and used as a GLP-1R-mediated functional response in properly designed assays.
Analysis of food intake-induced increase in GLP-1 level and subsequent activation of GLP-1R has provided valuable insights into the role of GLP-1R signaling in energy homeostasis. However, the short half-life and low basal level of circulating GLP-1(7-36) amide do not permit assessment of the physiological relevance of GLP-1R signaling other than energy homeostasis. During initial screening of GLP-1 potentiation activity from edible plants, we found that olive oil enhances GLP-1-dependent cAMP production in RINm5F cells. Here, we report that the specific endogenous endocannabinoid-like lipids oleoylethanolamide (OEA) and 2-oleoylglycerol (2-OG) can bind to GLP-1 peptide and are accompanied by an enhancement of cAMP production. Physiological levels of these lipids are subject to temporal and spatial regulation (17)(18)(19), and alone, they cannot activate GLP-1R signaling. Here, we show that the potency of GLP-1 is modulated by specific lipids, and this may represent a novel mode of spatiotemporal regulation of GLP-1R signaling.

Experimental Procedures
Reagents-A U2OS osteosarcoma cell line stably expressing a ␤-arrestin 2-GFP fusion protein was obtained from Norak Biosciences (Morrisville, NC) (now Molecular Devices, a part of MDS (Mississauga, Ontario, Canada)). The construct pcDNA-Rluc8, which expresses a variant of Renilla luciferase (Rluc8) (20), and a green fluorescent protein (GFP2) cDNA were kind gifts of Drs. Sanjiv Gambhir (Stanford University, Stanford, CA) and Szu-Hao Kung (21), respectively. cDNA encoding human exchange protein activated by cAMP 1 (Epac1; Rap guanine nucleotide exchange factor 3) and cDNA of human GLP-1R were purchased from OriGene Technologies (Rockville, MD). All oligonucleotides were from Mission Biotech (Taipei, Taiwan). The Lipofectamine LTX transfection kit and TOPO TA cloning kit were from Invitrogen. Peptides of GLP-1(7-36) amide, His-tagged GLP-1 , glucagon, and gastric inhibitory polypeptide (GIP) were from LifeTein (Hillsborough, NJ). Exendin 4 (Ex-4) and exendin 9 (Ex-9) were synthesized by Genomics BioSci & Technology (Taipei, Taiwan Construction of RG-cAMP Sensor-We replaced the energy donor and acceptor of the previously described Epac1 biosensor (22,23) with Rluc8 and GFP2, respectively, for enhanced bioluminescence resonance energy transfer 2 (BRET2) (24). Primers used to generate the fusion protein construct are shown in Table 1. Construction of expression vector for fusion protein RG-cAMP sensor is described in the legend to Fig. 1. Briefly, overlap extension PCR using Rluc8-pcDNA and pCMV-XL5-Epac1 as templates was used to generate Rluc8-Epac1 148 -430 pcDNA*, encoding Epac1 amino acids 148 -430 (Fig. 1A). The fragment of pCMV-XL5-Epac1 encoding the C-terminal portion of Epac1 was then inserted into Rluc8-Epac1 148 -430 pcDNA* to yield construct Rluc8-Epac1 148 -881 pcDNA* (Fig. 1B). Epac1 750 -881(T781A,F782A) , encoding Epac1 amino acids 148 -881, where Thr 781 and Phe 782 have been changed to alanine, and GFP2 gene sequence were independently generated by PCR. Overlap extension PCR was used to generate the Epac1 750 -881(T781A,F782A) -GFP2 fusion gene, followed by generation of the expression construct Rluc8-Epac1 148 -881(T781A,F782A) -GFP2-pcDNA, which codes for the TABLE 1 Primers used in the generation of RG-cAMP sensor Primer 1 is the forward primer from the 5Ј-untranslated sequence to the start codon of Rluc8. Primer 2 is the reverse primer containing the sequence complementary to nucleotides 916 -938 of the rluc8 gene and 444 -442 of the epac1 gene, separated by a linker sequence (boldface type). Primer 3 is a forward primer containing the forward sequence of nucleotides 922-933 of the rluc8 coding sequence and 442-459 of the epac1 coding sequence separated by a linker (boldface type). Primer 4 is a reverse primer covering the epac1 coding sequence from nucleotide 1270 to 1290. Primers 2 and 3 possess a 5Ј-extension that is complementary to the 5Ј-end of the epac1 sequence coding for amino acids 148 -430 of epac1 and the 3Ј-end of the rluc8 coding sequence. Primer 5 is the forward primer spanning nucleotides 2234 -2254 of the epac1 coding sequence. Primer 6 is a reverse primer spanning the 5Ј-end of the GFP2 coding sequence and the 3Ј-end of the epac1 coding sequence nucleotides 2626 -2643, separated by a linker (boldface type). Primer 7 is a reverse primer spanning nucleotides 2632-2643 of the epac1 coding sequence fused to the sequence spanning nucleotides 1-11 of the GFP2 coding sequence, separated by a linker (boldface type). Primer 8 is a reverse primer spanning nucleotides 701-719 of the GFP2 3Ј coding sequence with sequence containing an XbaI site (underlined) at the 5Ј-end.
Cell Culture and Transfection of Cell Lines-The RINm5F cell line is derived from a rat islet cell tumor (25,26) and expresses receptors for GLP-1, GIP, and glucagon (27,28). RINm5F cells were seeded at a density of 3 ϫ 10 5 cells/well in 24-well plates and cultured at 37°C and 5% CO 2 in complete RPMI 1640 medium supplemented with 10% (v/v) FBS, 1 mM sodium pyruvate, 14 mM glucose, 1:1000 penicillin/streptomycin, 2 mM glutamine, and 35 mM sodium bicarbonate, pH 7.4, for 16 h. Before transfection, the medium was removed and replaced with 500 l of fresh complete RPMI 1640. Transfection was carried out using the Lipofectamine LTX transfection kit according to the manufacturer's protocol. Briefly, 0.5 g of plasmid RG-cAMP sensor in 0.1 ml of Opti-MEM containing 1 l of Plus reagent and 3 l of LTX reagent was used to transfect the cells. Transfection complexes were added dropwise to the cells. After a 12-h incubation, the medium containing the lipoplexes was removed and replaced with 500 l of complete RPMI 1640. The cultures were incubated for another 48 h, the population of green fluorescent cells was counted, and transfection efficiency was estimated. Cells from each well were transferred to a 15-cm culture dish with 20 ml of complete RPMI 1640 containing 400 g/ml G418 to select the stably transfected cells.
BRET and cAMP Response Assay-RINm5F cells stably expressing the RG-cAMP sensor were seeded at 3 ϫ 10 4 cells/ well in 96-well white plates in 0.15 ml of RPMI 1640 containing 400 g/ml G418. On the next day, cells were washed twice with 0.1 ml of phenol red-free MEM containing 5 mM HEPES and incubated in the same medium for 1 h. The medium was replaced with 90 l of the same medium containing 1 mg/ml BSA and 5 M Deep Blue C. The whole plate was immediately loaded onto a SpectraMax paradigm detection platform equipped with a dual-color luminescence detection cartridge and SoftMax Pro 6.2.2 (Molecular Devices, Sunnyvale, CA) to obtain the background BRET signal based on the sequential integration of the luminescence detected at 370 -450 and 500 -530 nm over 60 -150 s. Each well was then stimulated by adding 10 l of 10ϫ solutions of peptide and lipids and 1 mg/ml BSA in phenol red-free MEM containing 5 mM HEPES, and BRET signals were obtained immediately under identical settings. The BRET ratio is the ratio of light emitted between 90 and 300 s at 500 -530 nm to that emitted at 370 -450 nm. The cAMP response was expressed as a percentage of cAMP production and was calculated as 100 ϫ (BRET ratio from 0.01 nM GLP-1(7-36) amide Ϫ BRET ratio from the indicated concentration of peptide with or without lipids)/(BRET ratio from 0.01 nM GLP-1(7-36) amide Ϫ BRET ratio from 250 nM GLP-1(7-36) amide). The dose-response curve, maximal response, and concentration of peptide needed to yield half-maximal response (EC 50 ) were obtained by nonlinear regression to fit the data to the agonist versus response equation using Prism software version 5.0 (GraphPad, San Diego, CA). Unless otherwise specified, all cAMP response data are mean Ϯ S.E. of three independent experiments performed in triplicate.
GLP-1 and OEA Binding Assay-The putative GLP-1 peptide and lipid interaction were examined with a binding assay between tritium-labeled OEA ([ 3 H]OEA) and His-tagged GLP-1(7-36) peptide (Fig. 6A (29) and separated by two alanine residues as linker. The GLP-1R-V2R chimeric construct was inserted into the EcoRI site of pcDNA3 (pcDNA3-GLP-1R-V2R) such that expression of the chimeric protein was under the control of the CMV promoter. pcDNA3-GLP-1R-V2R was used to transfect U2OS osteosarcoma cells stably expressing ␤-arrestin-2-GFP to obtain cell lines stably co-expressing GLP-1R-V2R and ␤-arrestin-2-GFP. U2OS cells stably expressing GLP-1R-V2R were grown to 90% confluence (about 10 7 cells/15-cm dish). The media was removed and washed twice with 30 ml of PBS, followed by adding 2.2 ml of ice-cold homogenization buffer (20 mM HEPES, 1 mM EDTA, 0.7% protease inhibitor mixture (Sigma-Aldrich, P8340) per dish. The cells were scraped immediately, and the scraped cells were centrifuged at 3000 ϫ g and 4°C for 30 min. The pellet was resuspended with 5 ml of homogenization buffer and then homogenized with the maximal speed of a Polytron 3000 system for 10 s on ice with a 30-s interval rest three times. The homogenized cells were centrifuged for 10 min at 4°C and 1000 ϫ g. The supernatant was transferred to a fresh transparent centrifuge tube and centrifuged for 60 min at 4°C and 55,000 ϫ g. The pellets were resuspended with resuspension buffer (20 mM HEPES, 1 mM MgCl 2 , 0.7% protease inhibitor) by passing through the 22-gauge needle one time, a 25-gauge needle three times, and then a 26-gauge needle one time. The protein concentration of the membrane preparation was determined according to the instructions for the Qubit fluorometer (Life Technologies). 4 -5 g of this membrane will yield greater than 5-fold signal/background with 125 I-labeled GLP-1 at 5 nM in a 0.22-ml assay volume.
Receptor Binding Assay-Assays were conducted in 0.22 ml of 50 mM HEPES, pH 7.4, 5 mM MgCl 2 , 1 mM CaCl 2 , and 1 mg/ml BSA containing 3.9 g of GLP-1 receptor membranes, varying concentrations of [ 125 I]GLP-1(7-36) radioactive ligand and in the absence (total binding) or presence (nonspecific binding) of a 500-fold excess of unlabeled exendin 4. The reactions were incubated for 90 min. Prior to filtration, an FC 96-well harvest plate (Millipore, MAFC N0B 10) was coated with 0.5% polyethyleneimine for 30 min and then washed with 50 mM HEPES, pH 7.4. 0.1 ml of the binding reaction was transferred to the filter plate and washed 15 times (0.1 ml/well/wash) with ice-cold 25 mM HEPES, pH 7.4, and 50 mM NaCl. The plate was dried, followed by the addition of 30 l of Microscint 20 (PerkinElmer Life Sciences) per well, and the activity was determined by using single-photon counting (60 s/well read) on a TopCount scintillation counter (PerkinElmer Life Sciences). Specific bindings were obtained by subtracting nonspecific binding from total binding. The dissociation constant (K d ) for GLP-1(7-36) was obtained using Prism software version 5.0 (GraphPad). Data shown are the means Ϯ S.E. of three independent experiments performed in duplicates.
In Vitro Assay of Trypsin Activity in the Presence of OEA-To determine whether OEA affects trypsin activity, we used the Trypsin Activity Colorimetric Test Kit (BioVision, Milpitas, CA). Briefly, 2 l of substrate was added to a well in a 96-well plate containing 48 l of 0.000125% trypsin in the presence or absence of 92 M OEA. Reactions were conducted at room temperature, and the extent of cleavage was monitored by real-time reading of optical density at 405 nm.
Determination of Concentrations of GLP-1  Amide and Glucagon and GIP Peptides-Molar concentrations were determined using an equation whereby peptide concentration (M) ϭ (A 280 ϫ -fold dilution)/(1200 ϩ 5560), where 1200 and 5560 are the molar extinction coefficients for tyrosine and tryptophan, respectively.

Results
Construction of RG-cAMP Sensor-To reduce BRET background and obtain better separation of the donor and acceptor energy emission peaks, we replaced the energy donor and acceptor in the early version Epac1 cAMP biosensor (22,23) with Rluc8 (20) and GFP2, respectively, for use in enhanced BRET2 (24) (Fig. 1, A-C). For a detailed description of generation of the expression construct for the RG-cAMP sensor, see "Experimental Procedures." Validation of the RG-cAMP Sensor in RINm5F Cells-The use of the RG-cAMP sensor ( Fig. 2A) as an intracellular cAMP biosensor in the RINm5F cell line, which is derived from a rat islet cell tumor, in the BRET assay was validated using the membrane-permeable cAMP analog 8-Br-2Ј-O-Me-cAMP-AM, the adenylyl cyclase activator forskolin, and the adenylyl cyclase inhibitor MDL-12330A. As shown in Fig. 2B, increasing the level of cAMP analog resulted in reduction of the BRET ratio in a concentration-dependent and saturable manner. The BRET ratio started to decrease from 1. This observation led us to further test whether this cAMP biosensor would respond to forskolin, which activates adenylyl cyclase and leads to cAMP production. As shown in Fig. 2C, forskolin treatment resulted in reduction of the BRET ratio in a dose-dependent and saturable manner; the BRET ratio started to decrease at a concentration of 1.8 nM forskolin, and reduction became saturated as the forskolin concentration reached 5000 nM. The forskolin-elicited BRET response was eliminated in the presence of 250 M MDL-12,330A, a potent adenylyl cyclase inhibitor. The change in basal cAMP level to that maximally stimulated by forskolin corresponds to a BRET ratio reduction of 0.3 units. The EC 50 for forskolin was determined to be 70 nM. This observation demonstrated that activation of adenylyl cyclase leads to increased intracellular cAMP. MDL-12,330A, which in turn inhibits the enzyme, abolished the activating effect of forskolin. Both processes can be sensed by the RG-cAMP sensor in the cells. RINm5F cells endogenously express G␣ s -coupled receptors for GLP-1, GIP, and glucagon (27,28). cAMP production in RINm5F cells expressing the RG-cAMP sensor showed dose-dependent and saturable responses to GLP-1(7-36) amide, GIP, and glucagon (Fig. 3, A-C). These responses were curtailed by the corresponding antagonists, Ex-9, GIP(8 -42) (30), and [des-His 1 ,Glu 9 ]glucagon amide (31), respectively, indicating that the cAMP responses were mediated by specific activation of the cognate receptors (Fig. 3,  A-C). The EC 50 was 1.3, 6, and 43 nM for GLP-1(7-36) amide, GIP, and glucagon, respectively. These results validated the

GLP-1R Signaling Regulation by Modulation of GLP-1 Potency
applicability of the RG-cAMP sensor in detecting intracellular cAMP level in response to incretin and glucagon stimulation in RINm5F cells.
OEA and 2-OG Potentiate cAMP Production in Response to GLP-1  Amide-In a preliminary test of the ability of the 18-carbon fatty acids and their lipid derivatives to enhance the GLP-1R cAMP response using a BRET assay, OA (18:1) and linoleic acid (18:2) at 106 and 107 M, respectively, consistently but mildly left-shifted the dose-response curve of GLP-1(7-36) amide (Fig. 4, A and B), whereas stearic acid, ␣-linolenic acid, and ␥-linolenic acid did not show such an effect (Fig. 4C). Other lipids tested that did not affect GLP-1 potency in receptor signaling included stearoylethanolamide (SEA), palmitoylethanolamide, and n-oleoyldopamine (Fig. 4, D and E). We then examined endogenous lipid derivatives containing OA as their acyl-lipid backbone. OEA and 2-OG, at a concentration of 9 to 8 M, markedly left-shifted the dose-response curve for GLP-1(7-36) amide to stimulate cAMP production in RINm5 F cells (Fig. 5, A and B). Responses were enhanced when the concentration of GLP-1 reached 20 pM, and the EC 50 of GLP-1(7-36) amide was reduced from 1.29 nM to 0.12 and 0.11 nM by 9.2 M OEA and 8.4 M 2-OG, respectively ( Table 2). The enhancement was dependent on the GLP-1(7-36) amide, because the effect diminished when the concentration of GLP-1(7-36) was reduced to 2 pM. To examine whether this enhancement is specific for GLP-1(7-36) amide, we further examined whether OEA and 2-OG could enhance the potency of exendin 4 (Ex-4), an analog of mammalian GLP-1. Ex-4 dose-dependently stimulated cAMP production and reached saturation at a concentration of 10 nM, with an EC 50 of 0.5 Ϯ 0.13 nM (Fig. 5, C and D); however, the potency of Ex-4 was unaltered by the addition of OEA or 2-OG to a concentration up to 27 M (Fig. 5, C and D). Therefore, OEA and 2-OG do not enhance the Ex-4-stimulated cAMP response in RINm5F cells. RINm5F is a rat insulinoma cell line, which should expresses many receptors involved in cAMP production other than GLP-1R. In order to investigate whether receptors other than GLP-1 receptor participate in OEA-or 2-OG-mediated enhancement of cAMP production by GLP-1, we examined whether this enhancement and cAMP responses still occur when GLP-1 receptor signaling was blocked by its antagonist exendin 9. As shown in Fig. 5E, 9 M OEA or 8.4 M 2-OG remarkably enhanced the cAMP responses to 200 pM GLP-1(7-36) amide in RINm5F cells (Student's t test; p Ͻ 0.001). This enhancement and cAMP responses were barely detectable when GLP-1 receptor signaling was blocked by the presence of 500 nM exendin 9. This analysis clearly illustrated that most of OEA-or 2-OG-mediated enhancement of cAMP response by GLP-1 is through GLP-1 receptors. This analysis revealed the specificity of these lipids in enhancing GLP-1(7-36) amide potency and not that of structurally related peptides or receptors other than GLP-1R.
GLP-1 Specifically Binds to OEA and 2-OG-Because both OEA and 2-OG enhance the potency of GLP-1(7-36) but not that of exendin 4, one possible explanation is that these lipids may interact specifically with GLP-1(7-36) amide. We further looked into the binding between His-tagged GLP-1(7-36) (Fig.  6A) and [ 3 H]OEA. His-tagged GLP-1(7-36) displays comparable response to OEA to enhance cAMP response as GLP-1(7-36) amide does. In a binding reaction, the peptide-bound [ 3 H]OEA together with the free peptide were captured by Cu 2ϩ -NTA resin and separated from free [ 3 H]OEA after centrifugation. 3 l of Cu 2ϩ -NTA resin in an 80-l binding reaction volume has been validated to be sufficient to capture all of the His-tagged GLP-1(7-36). Fig. 6B shows that the binding of 0. Because the potency of GLP-1 to stimulate cAMP production is enhanced by OEA and 2-OG but not by SEA, we also analyzed binding of 2-OG and SEA to His-tagged GLP-1(7-36) in a competition assay. As shown in Fig. 6D, the His-tagged GLP-1(7-36)-bound [ 3 H]OEA decreased as the concentration of 2-OG increased from 0.125 to 2.0 M, and the K i is estimated to be 0.1 Ϯ 0.03 M. SEA, on the other hand, did not affect the binding of [ 3 H]OEA to His-tagged GLP-1(7-36). These observations indicate that OEA and 2-OG specifically bind to Histagged GLP-1 .
Effect of OEA on Saturation Binding of GLP-1 to GLP-1 Receptor-Because OEA binds to and enhances the potency of GLP-1(7-36), we further investigated whether this effect may result from an enhanced binding affinity of GLP-1(7-36) to GLP-1 receptor. We performed a saturation binding assay to examine the effect of OEA on GLP-1 receptor binding affinity. As shown in Fig. 7A, binding of [ 125 I]GLP-1 to GLP-1 receptor membrane is dose-dependent and saturable; the dissociation constant K d and the maximal binding (B max ) were measured to be 1.5 Ϯ 0.25 nM and 51 Ϯ 2.9 pmol/mg of membrane, respectively. The saturation binding curve was barely affected by 9.2 M OEA, which maximally enhances the potency of cAMP production by GLP-1 (Fig. 7B). The K d and B max measured in the presence of 9.2 M OEA were 1.4 Ϯ 0.5 nM and 51 Ϯ 6.5   (Fig. 7C).
This analysis revealed that OEA has no discernible effects on the affinity or capacity of GLP-1 to bind GLP-1R. It is concluded that 9.2 M OEA, although it maximally enhances the  potency of GLP-1  to stimulate cAMP response, does not affect the physical affinity between GLP-1 and its cognate receptor.

TABLE 2 Effect of endocannabinoid-like lipids on the EC 50 (nM) of GLP-1 peptides to elicit cAMP response
Assay of the effects of the indicated concentrations of OEA and 2-OG on cAMP response to the titration of GLP-1(7-36) amide was carried out in RINm5F cells expressing RG-cAMP sensor. The cAMP response is determined by the reduction in BRET ratio, which is the ratio of light emitted at 500 -530 nm to that emitted at 370 -450 nm. The concentration needed to give half-maximal response (EC 50 ) was derived using nonlinear regression to fit the data to the agonist versus response equation from the Prism software version 5.0 (GraphPad, Inc., San Diego, CA). Data are means Ϯ S.E. of triplicate determinations from three independent experiments. tion of trypsin for the limited digestion of 2 M GLP-1(7-36) amide, we titrated trypsin concentration from 2.5 ϫ 10 Ϫ5 to 0.002% versus GLP-1(7-36) amide activity to stimulate cAMP production in RINm5F cells. As shown in Fig. 8C, the residual activity of GLP-1  to stimulate cAMP production was gradually reduced as trypsin concentration increased from 2.5 ϫ 10 Ϫ5 to 0.002%. The activity was decreased by a factor of 10 when GLP-1  was digested with 0.002% of trypsin. This analysis demonstrated that trypsin inactivates GLP-1(7-36) amide in a dose-dependent manner. When the trypsin cleavage reactions were carried out in the presence of 92 M OEA (Fig.  8D), GLP-1 activity was dramatically reduced by a factor of more than 300 (Fig. 8D). This analysis revealed that OEA remarkably facilitated GLP-1  to trypsin digestion. This is not due to an activation of trypsin by OEA, because OEA does not affect the innate trypsin activity (Fig. 8B). To analyze whether the effect of OEA on trypsin inactivation of GLP-1 was dependent on the concentration of OEA, digestion of GLP-1(7-36) amide with 0.00067% trypsin was carried out with increasing concentrations of OEA. Fig. 8E shows that as the concentration of OEA increased from 1 to 92 M, the residual activity of GLP-1(7-36) amide after trypsin digestion was reduced by a factor of more than 40. The effect of OEA became saturated as it reached a concentration of 9 M. These analyses revealed that OEA dose-dependently and saturably promotes GLP-1  peptide to trypsin cleavage. In contrast, SEA at concentrations ranging from 1 to 92 M did not affect trypsin digestion of GLP-1(7-36) amide (Fig. 8F). Our findings clearly revealed that trypsin digestion of GLP-1(7-36) amide was facilitated by OEA. This is not due to the activation of trypsin by OEA, because OEA did not affect the innate enzymatic activity of trypsin (Fig. 8B). Taken together, these data show that the susceptibility of GLP-1 peptide to trypsin cleavage is specifically increased by OEA, most likely due to a conformational change in GLP-1  amide.

Discussion
The real-time intracellular cAMP assay used in our study is an enhanced BRET2 (24) assay that has better separation of the energy emitted by the donor and acceptor (22,33) and yields data comparable with that obtained by other previously reported methods (22,33). This enhanced assay has the advantage of being able to monitor intracellular cAMP production (not cAMP accumulation) in the absence of phosphodiesterase inhibition (e.g. by 3-isobutyl-1-methylxanthine) as well as being a higher throughput assay, based on the use of 96-or 384-well plates.
OEA and 2-OG are both endogenous fatty acids reported to be ligands for GPCR (GPR119) but, when present alone in our assay, did not elicit cAMP responses even with concentrations as high as 10 times (33 M for OEA and 28 M for 2-OG) the reported EC 50 (2.9 M) for these lipids to activate GPR119 (34). These data agree with the observation that the RINm5F cell line is devoid of GPR119 receptor signaling (35,36), and the assay is thus free from the interference of GPR119 signaling.
In the present work, we present evidence that specific endocannabinoid-like lipids (OEA and 2-OG) increase the potency of GLP-1 to stimulate cAMP production, whereas SEA showed little effect. Lipid modification of glucagon has been reported to alter receptor selectivity between GLP-1R and glucagon receptor (GCGR). It is also possible that binding of OEA or 2-OG to GLP-1 peptide may activate receptors other than GLP-1R in RINm5F cells and results in enhanced cAMP production. The present studies show that the enhancement effect of OEA or 2-OG and cAMP response was barely detected when GLP-1Rs in RINm5F cells were blocked by 500 nM of GLP-1R antagonist (Fig. 5E). This observation indicated that lipid enhancement and cAMP response is most likely mediated by GLP-1R. The result also shows that other receptors in RINm5F cells will not respond to the enhancement of potency by OEA or 2-OG. Several mechanisms may account for this phenomenon. At least three possible mechanisms depend on GLP-1R signaling. Binding of lipids to GLP-1R could enhance its coupling to G␣ s . Alternatively, the GLP-1 peptide, which may also interact with lipids, could enhance its potency to stimulate the coupling of GLP-1R to G␣ s . Alternatively, both of the mechanisms mentioned above could account for the observation. Several lines of evidence imply that endocannabinoid-like lipids interact with GLP-1(7-36) peptide. First, we are not able to observe detectable specific binding of [ 3 H]OEA to GLP-1 receptor. Second, both OEA and 2-OG selectively enhance the potency of the GLP-1(7-36) amide, but not that of Ex-4, indicating interaction with selective peptide. Third, a signal-enhancing lipid, OEA, but not the non-signal-enhancing lipid, SEA, facilitates trypsin inactivation of the peptide, also suggesting a lipid-peptide interaction. Finally, GLP-1(7-36) binds to OEA and 2-OG but not to SEA, revealing a specific binding of OEA or 2-OG to GLP-1(7-36) peptide. Thus the available experimental data are in line with a simple model that specific endocannabinoid-like lipids can bind to GLP-1 and form a complex with enhanced potency to stimulate cAMP response. Surprisingly, interaction of OEA with GLP-1 does not affect the binding affinity to GLP-1 receptor, and the lipid-peptide interactions are mostly dedicated to enhancing the potency to stimulate the coupling of GLP-1 receptor to cAMP production. It has been demonstrated that the binding energy of single point mutant GLP-1 to GLP-1R may not be realized to its signaling potency (37). OEA but not SEA facilitates susceptibility of GLP-1 to trypsin cleavage, which is consistent with the specific binding of OEA and not SEA to GLP-1. Putting these findings together, it is most likely that binding of OEA to GLP-1 may induce a conformational change of the peptide and is accompanied with an enhancement in potency. This proposed model is evocative of a scenario in which peptide ligands for class B GPCRs can be induced to undergo a conformational change in the presence of lipids (38) or upon interaction with their cognate receptors (39).
Apart from the well known effects of postprandial GLP-1 secretion on pancreatic ␤ cells for glycemic control, our knowledge of how GLP-1Rs broadly expressed in tissues serving different functions can be differentially regulated is far from complete (for a review, see Ref. 4). By changing the potency, our results present a novel way to regulate GLP-1 receptor signaling without altering the extracellular level of GLP-1 peptide. GLP-1Rs in the brain stem are considered activated by the centrally derived GLP-1, which is transmitted through neural fibers in most of brain areas (11,12,40). However, neurons expressing GLP-1Rs in notable areas, such as caudal hippocampus and the dopaminergic neurons in the substantia nigra, have no apparent innervations from GLP-1-producing neurons (1,3,14). Both areas are anatomic sites of deficient functions in disease that are ameliorable by GLP-1 analog therapy, which improves memory/learning in an Alzheimer mouse model and Parkinson symptoms in humans (5,6,41). It has been suggested that GLP-1Rs in these areas might be activated by the circulating GLP-1 in cerebrospinal fluid and blood (40), which is normally at a level that is probably below the threshold concentration for GLP-1R activation (42). Given that a 2-5-fold postprandial surge in plasma GLP-1 level can activate pancreatic ␤ cells to release insulin, induction of a 10-fold enhancement in potency by lipids may be sufficient to allow the basal level of GLP-1 to activate its receptors.
The lipids OEA and 2-OG are described as endocannabinoid-like because they share structural homology with the endocannabinoids but do not activate the cannabinoid receptors. Our knowledge of the biology of these endocannabinoid-like lipids is sketchy because they have attracted research interest only in recent years, because each is known to activate other receptors, including GPR119 and peroxisome proliferator-activated receptor (for a review, see Ref. 43). In general, endocannabinoid-like lipids are present in variable amounts in different tissues (44). Their levels are also subject to change following chronic feeding of a diet rich in select types of oil (e.g. olive oil predominantly yields OEA and 2-OG (43)). Because the levels of OEA and 2-OG are much higher (19,45,46), these endocannabinoid-like lipids would seem similarly well situated as signal modulators in the brain. OEA is generated in tissues in a stimulus-dependent manner and is quickly removed by enzymatic hydrolysis, indicating a function in signaling (for a review, see Ref. 17). The concentration of 2-OG is significantly elevated in the blood of hibernating animals compared with that of active animals (18). The level of 2-OG is in the range of 2-3 nmol/g of whole brain tissues but in specific regions of the brain can go up to 100 nmol/g of tissue and is also subject to physiological regulation (19). Distinct from the mode of all lipid mediators that activate receptors in their own right, these endocannabinoid-like lipids regulate GLP-1R signaling by enhancing the potency of GLP-1. In conclusion, local tissue levels of these lipids are varied and regulated physiologically; thus, they can modulate the potency of GLP-1 in a tissue-and physiology-dependent manner. This will allow temporal and spatial regulation of GLP-1R signaling without changing the basal level of GLP-1.
Finally, deterioration in GLP-1 signaling response in type 2 diabetes rather than reduced GLP-1 secretion is associated with the pathogenesis of the disease (47). Such patients have impaired incretin effect, despite normal GLP-1 secretion (48) and normal GLP-1 secretion in response to oral glucose or meal tests (47). Clinically, the administration of GLP-1 analog has been successful in ameliorating reduced GLP-1 signaling in type 2 diabetes; likewise, similar therapeutic effects may be achieved with positive modulators for GLP-1 peptide, which may restore the function of GLP-1 signaling that is more in line with physiological demands. Because cAMP is sufficient for glucose-dependent insulin release (49), one could investigate whether compounds capable of enhancing the cAMP response to GLP-1(7-36) amide stimulation in RINm5F cells may represent a new class of therapeutic agents. Or nutrient derivatives display similar activities can be used to treat or prevent type 2 diabetes and other GLP-1 signaling-related diseases.
In summary, our present study shows that the potency of a GLP-1 ligand is modulated by specific endogenous lipids, which in turn are subjected to temporal and spatial regulation. This novel mode of regulation allows GLP-1R signaling in different tissues to be differentially activated without changing the basal level of GLP-1. This mode of regulation is distinct from the traditional model, which requires that the extracellular levels of ligand be the primary factors contributing to regulation of GPCR signaling. Our results show that the potency of the ligand itself can be a target for modulation. This finding could further lead to exploration of a distinct class of endogenous peptide ligands that represent novel targets for drug discovery.