2-Arachidonyl glyceryl ether, an endogenous agonist of the cannabinoid CB1 receptor Lumı´r Hanusˇ*, Saleh Abu-Lafi†, Ester Fride*‡, Aviva Breuer*, Zvi Vogel§, Deborah E. Shalev¶, Irina Kustanovich¶, and Raphael Mechoulam*ʈ *Department of Medicinal Chemistry and Natural Products, Hebrew University Medical Faculty, Jerusalem 91120, Israel; †Chemistry and Chemical Technology Department, Faculty of Science and Technology, P.O. Box 20002, Al-Quds University, Abu-Deis, West Bank, Palestinian Authority; ‡Department of Behavioral Sciences, College of Judea and Samaria, Ariel 88437, Israel; §Department of Neurobiology, The Weizmann Institute of Science, Rehovot 76100, Israel; and ¶Department of Biological Chemistry and The Wolfson Centre for Applied Structural Biology, Institute of Life Sciences, Hebrew University of Jerusalem, Jerusalem 91904, Israel Communicated by L. L. Iversen, University of Oxford, Oxford, United Kingdom, January 18, 2001 (received for review November 23, 2000) Two types of endogenous cannabinoid-receptor agonists have 2 kg of porcine brain, we obtained 0.5 mg of noladin ether, been identified thus far. They are the ethanolamides of polyun- approximately the amount of anandamide obtained from a saturated fatty acids—arachidonoyl ethanolamide (anandamide) is comparable extraction (1). the best known compound in the amide series—and 2-arachido- noyl glycerol, the only known endocannabinoid in the ester series. Preparation of Synaptosomal Membranes. Synaptosomal mem- We report now an example of a third, ether-type endocannabinoid, branes were prepared from the brains of Sabra male rats (250 g) 2-arachidonyl glyceryl ether (noladin ether), isolated from porcine after removal of the brainstem. The [3H]HU-243 probe, with a brain. The structure of noladin ether was determined by mass dissociation constant of 45 pM (11), was incubated with synap- spectrometry and nuclear magnetic resonance spectroscopy and tosomal membranes (3–4 ␮g) for 90 min at 30°C with the was confirmed by comparison with a synthetic sample. It binds to indicated concentrations of noladine ether or the vehicle alone ؎ ؍ the CB1 cannabinoid receptor (Ki 21.2 0.5 nM) and causes (fatty-acid-free BSA at a final concentration of 0.5 mg͞ml). sedation, hypothermia, intestinal immobility, and mild antinoci- Bound and free radioligand were separated by centrifugation ␮ ception in mice. It binds weakly to the CB2 receptor (Ki > 3 M). (see ref. 11 for full details of the assay). The data were normalized to 100% of specific binding, which was determined e have reported the isolation and identification of two with 50 nM unlabeled HU-243. Data points represent the Wtypes of endogenous cannabinoids that bind and activate average of triplicate determinations from three independent the known cannabinoid receptors CB1 and CB2. Arachidonoyl experiments. The Ki value was determined with a GraphPad (San ethanolamide (anandamide; ref. 1) and later two more polyun- Diego) PRISM 2.01 program that follows the Cheng-Prusoff equa- saturated fatty acid ethanol amides (2) were found in porcine tion (12). A sigmoid dose-response (variable slope) built-in brain. An ester, 2-arachidonoyl glycerol (2-AG) was isolated by equation in PRISMwas used to fit the curve shown in Fig. 2. us from canine gut (3) and by Sugiura et al. from brain (4). Anandamide and 2-AG have been the objects of numerous HPLC. The HPLC analysis was performed by using a Waters liquid investigations in various areas of biology and have been found to chromatograph (Millipore) equipped with a Photodiode Array affect processes in the nervous, cardiovascular, immune, and (model 996), 20- and 50-␮l loops, a Waters 600E multisolvent reproductive systems (5–8). They interact with many neurotrans- delivery system, and a Millennium 2010 chromatography man- mitters and affect hormone levels (9–10). This ubiquity of effects ager for the analysis of data. A reversed-phase Waters Symmetry led us to look for additional endocannabinoids. C18 column (5 ␮m, 250 ϫ 4.6 mm i.d.) was used. Reversed-phase We report now that we have isolated from porcine brain a third solvents were HPLC grade. Acetonitrile was purchased from endocannabinoid, 2-arachidonyl glyceryl ether, which we have J. T. Baker. Water was doubly distilled and filtered by using a named noladin ether (Fig. 1). 0.2-␮m cellulose mixed ester membrane (Schleicher & Schuell). Before injection into the HPLC, the sample was filtered by Materials and Methods loading concentrated methanol extract onto a 0.2-␮m Gelman Isolation of Noladin Ether. Porcine brain (100 g, approximately a nylon syringe filter. single brain) was added to a mixture of chloroform (200 ml) and The analyses were run under isocratic conditions at ambient methanol (200 ml) and mixed in a blender for 2 min. Water (100 temperature. The optimized mobile phase used for the isolation ml) was added, and the mixing process was continued for another of the active noladin ether was acetonitrile͞water [70:30 (vol͞ minute. The mixture was filtered. Two layers were formed. The vol)]. The flow rate was 1.5 ml͞min, and the monitoring wave- water-methanol layer was separated and evaporated under re- length was 210 nm. Under these conditions, noladin ether elutes duced pressure. The residue obtained was extracted with meth- at 22.88 min. Of 4.5 mg of column fraction, 0.5 mg of pure anol. The above isolation was repeated numerous times (from a noladin ether was obtained. total of 2.4 kg porcine brain). The extract obtained was chro- Synthetic standard (noladin ether) was injected by using the matographed on a gravity column (i.d. 2.5 cm, height 28 cm, 82 g same working conditions. Similar elution time was noticed for ICN Silica TSC, 60 A) with hexane͞acetone initially in a ratio of both compounds. After comparing the UV͞Vis spectrum of the 10:1 (vol͞vol) (400 ml), then 9:1 (100 ml), and finally 4:1 (100 synthetic compound with extracted noladin ether, excellent ml). The fractions eluted were monitored for binding to the CB1 matching was obtained (Fig. 3). Moreover, when a synthetic receptor from rat brain synaptosomes (prepared as described below; ref. 11) on the basis of displacement of the potent labeled agonist [3H]HU-243 purchased from Tocris (Bristol, U.K.). The Abbreviations: 2-AG, 2-arachidonoyl glycerol; 2D, two-dimensional. ʈ material eluted in fractions 45–54 (10-ml each) was found to bind To whom reprint requests should be addressed at: Hebrew University, Department of Medicinal Chemistry and Natural Products, Faculty of Medicine, Ein Kerem Campus, to the receptor. These fractions were combined and purified Jerusalem 91120, Israel. E-mail: [email protected]. further by HPLC (see below). A polar-active compound devel- ͞ The publication costs of this article were defrayed in part by page charge payment. This oped on a TLC plate (silica gel 60 F254, Merck) in a hexane article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. ϭ acetone (4:1) solvent system gave a single spot (Rf 0.13). From §1734 solely to indicate this fact. 3662–3665 ͉ PNAS ͉ March 27, 2001 ͉ vol. 98 ͉ no. 7 www.pnas.org͞cgi͞doi͞10.1073͞pnas.061029898 Downloaded by guest on October 2, 2021 Fig. 1. Synthesis of noladin ether: step a, LiAlH4͞ether, 25°C; step b, CH3SO2Cl͞pyridine, 25°C; step c, benzylideneglycerol-KOH͞benzene, 25°C; step d, HCl͞CH3OH, 25°C. Percentages indicate yields of reactions. sample was spiked with natural noladin ether and injected into 0.25 mm ϫ 0.25 ␮m. Experimental conditions were: inlet, 250°C; the HPLC, one single peak was eluted at 22.88 min. Waters detector, 280°C; splitless injection͞purge time, 1.0 min; initial MILLENNIUM software purity calculations confirmed the exis- temperature, 70°C; initial time, 3.0 min; rate, 30°C͞min; final tence of only one component under this peak. temperature, 280°C. The MS spectrum is presented in Fig. 4. MS. A Hewlett Packard G 1800B GCD system that has a HP-5971 NMR Spectroscopy. NMR spectra were acquired on a Bruker GC with electron ionization detector was used. The software used (Billerica, MA) Advanced DMX 600-MHz spectrometer. Proton was GCD PLUS CHEMSTATION. The column used was HP-5 chemical shifts at 295.0 K were referenced to chloroform at 7.26 (crosslinked 5% PH ME siloxane) with a film thickness of 28 m ϫ and 77.0 ppm for proton and carbon signals, respectively. One- and two-dimensional (2D) homo- and heteronuclear data were collected at natural abundance by using 2D-correlated spectros- copy, 2D double-quantum filtered correlated spectroscopy, 2D total-correlation spectroscopy, and 2D heteronuclear multiple- quantum correlation experiments. All NMR data were processed by using XWINNMR 2.6 software (Bruker) running on Silicon Graphics (Mountain View, CA) INDY workstations. Zero-filling and suitable apodization func- tions were applied to enhance the spectral resolution. The spectra were analyzed by using Aurelia 2.7.5 software (Bruker). The one-dimensional proton spectra and the 2D-correlated spectroscopy and total-correlation spectroscopy spectra of nat- ural and synthetic noladin were compared and found to be identical under the acquisition conditions. The chemically iden- tical alkene methine protons 6, 7, 9, 10, 12, 13, 15, and 16 are overlapped, as are the methylene protons 8, 11, and 14 that have double bonds at both sides. The alkane protons 2 and 3 are CHEMISTRY Fig. 2. Competitive inhibition of [3H]HU-243 binding by natural noladin Fig. 3. Overlay of HPLC chromatograms of natural and synthetic noladin ether. For details see Materials and Methods. ether. For details see Materials and Methods. Hanusˇ et al. PNAS ͉ March 27, 2001 ͉ vol. 98 ͉ no. 7 ͉ 3663 Downloaded by guest on October 2, 2021 Fig. 4. Electron impact͞mass spectrum of di(trimethylsilyl) derivative of natural noladin ether.