Cannabinol derivatives: binding to cannabinoid receptors and inhibition of adenylylcyclase

Several derivatives of cannabinol and the 1,1-dimethylheptyl homolog (DMH) of cannabinol were prepared and assayed for binding to the brain and the peripheral cannabinoid receptors (CB₁ and CB₂), as well as for activation of CB₁– and CB₂-mediated inhibition of adenylylcyclase. The DMH derivatives were much more potent than the pentyl (i.e., cannabinol) derivatives. 11-Hydroxycannabinol (4a) was found to bind potently to both CB₁ and CB₂ (K_i values of 38.0 ± 7.2 and 26.6 ± 5.5 nM, respectively) and to inhibit CB₁-mediated adenylylcyclase with an EC₅₀ of 58.1 ± 6.2 nM but to cause only 20% inhibition of CB₂-mediated adenylylcyclase at 10 μM. It behaves as a specific, though not potent, CB₂ antagonist. 11-Hydroxycannabinol-DMH (4b) is a very potent agonist for both CB₁ and CB₂ (K_i values of 100 ± 50 and 200 ± 40 pM; EC₅₀of adenylylcyclase inhibition 56.2 ± 4.2 and 207.5 ± 27.8 pM, respectively).

Cannabinol Derivatives: Binding to Cannabinoid Receptors and Inhibition of Adenylylcyclase

Man-Hee Rhee,† Zvi Vogel,† Jacob Barg,† Michael Bayewitch,† Rivka Levy,† Lumir Hanusˇ,‡ Aviva Breuer,‡ and Raphael Mechoulam*,‡
Department of Neurobiology, The Weizmann Institute of Science, Rehovot 76100, Israel, and Department of Natural Products, Medical Faculty, Hebrew University, Ein Kerem Campus, Jerusalem 91120, Israel

Received March 3, 1997

Introduction

Two cannabinoid receptors have been described to date. Both are proteins with seven transmembranespanning domains. These receptors were originally found in rat brain and spleen, respectively, and are generally known as the central cannabinoid receptor, CB1, and the peripheral cannabinoid receptor, CB2.(1),(2) CB1 is mainly expressed in the central nervous system (CNS). It is also found, to a lesser extent, outside the CNS, in numerous other tissues such as vas deferens, adrenal gland, heart, lung, prostate, uterus, ovary, testis, bone marrow, thymus, and tonsils.(3),(4) The CB2 gene is not expressed in the brain, being found mostly in the immune system. In certain tissues, such as spleen, the mRNA content of CB2 is particularly high. In blood cell subpopulations, it is found particularly in B-cells.(3),(4) It has been suggested that “cannabinoids may exert specific receptor-mediated action on the immune system through the CB2 receptor”.(4) While a considerable amount of work has been done on structure–activity relationships (SAR) as regards binding to CB1,(5)–(7) very little is known on binding to CB2.(8) In the original paper on the identification of CB2, Munro et al.(9) showed that cannabinol (CBN) (1a), which only binds feebly to CB1, much less so than ¢9-tetrahydrocannabinol (¢9-THC) (2a), binds to CB2 at the same level of potency as ¢9-THC. The low binding to CB1 is compatible with the low THC-type in vivo activity recorded for CBN when compared to ¢9-THC.(10) The binding of CBN to CB2, with a potency equivalent to that of ¢9-THC, suggests that cannabinol derivatives could serve as a novel starting point for SAR analysis in the cannabinoid series and could possibly lead to selectivity as regards binding to the two cannabinoid receptors. We present here data on several cannabinol derivatives which have been tested for binding to CB1 and CB2.

(7a)(22) and of the DMH homolog of (–)-¢8-THC-11-oic acid (6a)23 into the DMH homolog of cannabinol-11-oic acid (7c). The lactones 8a,b were prepared following a route reported by the groups of Adams and Todd about 50 years ago.(24) Von Pechmann condensation of either olivetol (9a) or 3-(1,1-dimethylheptyl)resorcinol (9b) with ethyl 5-methyl-1-oxocyclohexane-2-carboxylate (10) led to the known tetrahydrodibenzopyrone 11a or 11b, respectively.(24) Dehydrogenation of 11a or 11b as described above gave 8a24 or 8b, respectively.

Biological Results

All compounds were tested for binding to CB1 using African green monkey kidney (COS-7) cells transfected with the cDNA of rat CB1 12,25 as well as rat brain synaptosomal membrane preparations.14,26 Binding to CB2 was performed using transfected COS-7 and Chinese hamster ovary (CHO) cells transfected with the cDNA of human CB2.9,11,12 The Ki values were determined by displacement of [3H]HU-243, a probe for both cannabinoid receptors.(26),(27) For consistency, the results discussed below are those from transfected cells except when specifically indicated. All compounds were also assayed for inhibition of CB1– and CB2-mediated adenylylcyclase.(11),(12) The results are summarized in Table 1. First, CBN (1a) was compared to ¢9-THC (2a). Under our experimental conditions, CBN was about 2–4 times less potent on binding to either CB1 or CB2, in contrast to the data reported by Munro et al.,(9) who found that CBN and ¢9-THC are equipotent on CB2. These differences in recorded potencies are not unexpected, as binding data are very sensitive to experimental conditions. The CB1-mediated adenylylcyclase inhibition data seem to be closer to in vivo observations, where, as mentioned above, ¢9-THC is considerably more potent than CBN. We found that CBN is about 13 times less potent than ¢9-THC. With CB2 we see a different profile. While the EC50 of CBN inhibition of CB2– mediated cyclase is 261.2 ( 46.4 nM, ¢9-THC causes only 21% inhibition at 1 íM. We have previously shown that ¢9-THC is a weak agonist for CB2 and can actually antagonize the CB2-mediated inhibition of adenylylcyclase caused by more potent agonists.(12) The next derivative to be examined was the DMH homolog of CBN (1b). We observed essentially no difference between its binding affinity to CB1 or CB2. On both receptors, 1b was about 100 times more potent than CBN. It was also 100 times more potent than CBN when tested for binding to rat brain membranes.

This compound also potently inhibited adenylylcyclase via both CB1 and CB2: the Ki of inhibition by 1b of CB1– mediated cyclase was ca. 660-fold lower, and the CB2– mediated cyclase Ki was ca. 300-fold lower than that of CBN. This result demonstrates that the exchange of the pentyl group with the DMH group (at position 3) dramatically increases the affinity of the ligand to both CB1 and CB2 and makes it a much more potent agonist. A similar result was obtained with other cannabinol derivatives (see below). Neither CBN (1a) nor its DMH homolog 1b express any particular selectivity with regard to the inhibition of adenylyl cylase via the two receptors. However, a major metabolite of CBN, namely, 11-hydroxy-CBN (4a),(19),(22),(28) had a different profile. It binds to both receptors with Ki values ca. 5-fold lower than the Ki values recorded for CBN. However, while  Morever ¢9-THC antagonizes the agonist-induced inhibition of adenylylcyclase mediated by CB2. Therefore, we concluded that ¢9-THC constitutes a weak antagonist for the CB2 receptor.(12) It was thus of interest to find out whether CBN derivatives could selectively activate one of the cannabinoid receptors. Therefore, all compounds were tested for CB1– as well as CB2– mediated inhibition of adenylylcyclase. The present paper is apparently the first one which examines the SAR of cannabinol derivatives for binding to the cannabinoid receptors as well as for inhibition of adenylylcyclase. Chemistry Cannabinol (1a), ¢9-THC (2a), and cannabidiol (3a) were extracted from hashish.(15),(16) The 1,1-dimethylheptyl (DMH) homolog (2c) (at position 3) of ¢9-THC was prepared by Lewis acid-catalyzed ring cyclization of the DMH homolog of cannabidiol (3b),(17)as previously described for the parallel conversion of cannabidiol (3a) into ¢9-THC (2a).(16) The DMH homolog of cannabinol (1b) was prepared by dehydrogenation with sulfur at 240 °C of 2d (synthesized by acetylation of 2c), leading to 1c, followed by removal of the acetate group with ethanolic sodium hydroxide. The isomeric 1,2-dimethylheptyl homolog of CBN has previously been prepared and shown to be a potent cannabimimetic.(18) 11-Hydroxycannabinol (4a) was prepared, as previously described, by selenium dioxide oxidation and dehydrogenation of ¢9-THC acetate (2b) followed by removal of the ester grouping.(19) 11-Hydroxy-¢8-THC (5c) and its DMH homolog (5a) were prepared as previously described.(19),(20) Dehydrogenation of the DMH homolog of 11-hydroxy-¢8-THC diacetate (5b)(20) with sulfur at 240 °C led to the DMH homolog of 11-hydroxycannabinoldiacetate (4c), which was converted to the diol 4b. The same procedure was followed for the conversion of ¢8-THC-7-oic acid (6c)(21) into cannabinol-7-oic acid  the EC50 for the CB1-mediated inhibition of cyclase was 58.1 ( 6.2 nM, the inhibition of the CB2-mediated cyclase was negligible: 20% inhibition at 10 íM. This result indicates that 4a could serve as a functional antagonist of CB2 as it binds, but does not activate, CB2.

Indeed, it reduces the CB2-mediated adenylylcyclase inhibitory activity of the potent agonists 5a (HU- 210)5,8b,20 and 7c (Figure 1). However while the adenylylcyclase inhibition by 7c is very strongly reduced by 4a, that of 5a is considerably less. This difference may well be due to the higher inhibitory potency of 5a, compared to 7c. It should be pointed out that 4a, a major metabolite of CBN, in its differential activity between CB1 and CB2, resembles ¢9-THC rather than CBN (see above):12 both ¢9-THC and 4a bind to CB1 and CB2 and inhibit CB1-mediated cyclase in the nanomolar range but cause only negligible inhibition of CB2– mediated cyclase up to micromolar concentrations. The DMH homolog of 11-hydroxycannabinol (4b) is the most potent cannabinoid in the present series. It binds to CB1 and CB2 with Ki values of 100 ( 50 and 200 ( 40 pM, respectively. Compound 4b also inhibits adenylylcyclase via both CB1 and CB2, at very low concentrations: 56.2 ( 4.2 and 207.5 ( 27.8 pM, respectively. Both the binding constants and EC50 values for CB1 and CB2 are comparable to the values observed for HU-210 (5a).

Cannabinol-11-oic acid (7a),(22) as well as its acetate (7b), were essentially inactive on binding to either CB1 or CB2, as well as in CB1– or CB2-mediated cyclase inhibition. However, the DMH homologs, compounds 7c,d, bind to both CB1 and CB2 and inhibit adenylylcyclase via the two receptors with high potency. The nonacetylated derivative 7c binds with Ki values of 6.1 ( 1.1 and 4.8 ( 1.9 nM to CB1 and CB2, respectively, and inhibits cyclase CB1– and CB2-mediated activity with EC50 values of 2.4 ( 0.2 and 5.4 ( 0.3 nM, respectively. These results were unexpected as THCtype 11-oic cannabinoid acids (including DMH homologs) such as 6a,c have not been reported to cause psychotropic effects,(10),(21),(23) which are presumably mediated by CB1. The acetate 7d is also active in all these assays, although its activity is 4–7 times lower than the nonacetylated acid (see Table 1). This difference was even larger (90 times) when the acetate 7d was compared to the nonacetylated acid 7c with regards to  binding to CB1 on synaptosomal membranes. It is not clear whether the weaker activity of the acetate is due to its own lower intrinsic activity or to a certain amount of nonacetylated 7c formed by hydrolysis during the assays. In view of these results (see Discussion), we looked again into the activity of the DMH homolog of ¢8-THC- 7-oic acid (6a).(23) This compound, synthesized in our laboratory several years ago, was found then to reduce paw edema and leukocyte adhesion to culture dishes and was considered to be antiinflammatory. It did not cause catalepsy in mice at doses up to 1 mg/kg.23 We now find that 6a binds to CB1 with a rather low Ki of 32.3 ( 3.7 nM but is only a relatively weak inhibitor of CB1-mediated cyclase (EC50 ) 927.0 ( 39.6 nM). It binds to the CB2 receptor with a Ki of 170.5 ( 7.8 nM and inhibits CB2-mediated cyclase (EC50 ) 116.2 ( 74.7 nM).

This material thus affects signaling via CB2 more effectively than via CB1. The last series to be examined were cannabinol derivatives in which the pyran ring was modified; instead of the two methyl groups on ring B, a ketone was introduced, thus forming a lactone. Compound 8a (with a pentyl side chain)(24) did not bind to either CB1 or CB2 and did not inhibit adenylylcyclase, while the DMH derivative 8b was highly potent, binding to both CB1 and CB2 with Ki values of 32.7 ( 4.9 and 17.3 ( 5.6 nM, respectively. It inhibited both CB1– and CB2– mediated adenylylcyclase with EC50 values of 9.6 ( 0.6 and 4.6 ( 0.3 nM, respectively.

Discussion

For reasons discussed in the Introduction, we assumed that CBN derivatives may show differential binding for CB1 compared to CB2. The results obtaineddid not fulfill these expectations. Compounds binding poorly to CB1 (such as 7a, 8a) had the same profile with CB2, and this was the case also for the more potent compounds (such as 1b, 4a,b, 7c,d, 8b). However, our results proved to be of interest in another direction: we found that binding to CB2 (see compound 4a) does not necessarily imply a parallel level of inhibition of CB2– mediated adenylylcyclase. As mentioned above, the major CBN metabolite, 11-hydroxy-CBN (4a), binds well to both CB1 and CB2 (Ki ) 38 ( 7.2 and 26.6 ( 5.5 nM, respectively). However, while 4a inhibits adenylylcyclase strongly via CB1 (EC50 ) 58.1 ( 6.2 nM), its inhibition of adenylylcyclase via CB2 is negligible (Table1). Further experiments (see Biological Results and Figure 1) indeed showed that 4a is a weak antagonist to CB2. Hence, this CBN metabolite may represent a useful target for future research and possibly a tool in CB2 investigations. As 4a is formed in vivo from CBN, its presence in the body may have physiological consequences associated with CB2-promoted activities, possibly in the immune system.

The replacement of the pentyl group at position 3 with the DMH group increased affinity to CB1 as well as to CB2 in all compounds tested. The dramatic increase of pharmacological activity associated with such a structural change was first noted about 50 years ago.24c,d As indicated above, the DMH homolog of 11-hydroxy-CBN (4b) is the most active compound in the present series as regards binding to CB1 and CB2, as well as inhibition of adenylylcyclase (see Table 1). It may serve alongside HU-210 (5a) as a potent tool in cannabinoid research.

Indeed, 4b shows binding values and cyclase inhibition levels very similar to those of HU-210 (5a) (see Table 1). Oxidation of the 11-CH3 group in ¢9-THC (2a) or ¢8– THC (12) to a carboxyl group forming THC-7-oic acids (13 or 6c, respectively) leads to inactivation, as seen in behavioral assays.10,21 Indeed, this route is the major inactivation pathway of THC metabolism. The DMH homolog of ¢8-THC, 11-oic acid 6a, has also been reported to lack THC-type activity,(23) although the reported test range was limited. Unexpectedly, we now find that although 6a binds to both CB1 and CB2 in the 30–170 nM range, the EC50 of adenylylcyclase inhibition via CB1 is in the range of 927 ( 39.6 nM. The CB2– mediated adenylylcyclase is in the intermediate range (116.2 ( 74.7 nM), i.e., a certain separation between CB1– and CB2-mediated activation is noted. In the CBN-11-oic acid series, we observed a different profile: while in the pentyl series (7a,b), as expected, we recorded no binding or inhibition of cyclase, with DMHCBN- 11-oic acid (7c), we found potent binding to both CB1 and CB2 and cyclase inhibition (see Table 1). Compounds 6a and 7c differ in the conformation of ring A: a planar aromatic ring in 7c versus a half-chair one in 6a. This conformational difference may represent the molecular basis for the different activities of the two cannabinoids.

A comparison of the binding values of 6a with those of its CB1– and CB2-mediated cyclase inhibition (see Table 1) may explain some of the previously reported properties of 6a. The relatively high levels of 6a needed to inhibit adenylylcyclase via CB1 may explain the absence of catalepsy (within the limited dose range tested), in spite of its considerable binding potency, while the ca. 10 times lower levels needed to inhibit adenylyl cylcase via CB2 may be the basis of its reported antiinflammatory activity. These observations may serve to open new leads toward the development of antiinflammatory agents in which wider separations of activity can be achieved than these reported now. As mentioned above, all compounds presented in Table 1 were tested for binding both in transfected cells and in rat brain synaptosomal preparations. While the general potency trend is comparable in the two assays, some individual differences are striking. Compounds 6b, 7b, and 8a are inactive in both assays; 7a is inactive in transfected cells and poorly active in the membrane assay. The highly potent 1b, 5a, and 7c have the same profile in both assays. However 11-hydroxy CBN-DMH (4b) has a Ki of 100.0 ( 50.0 pM in transfected cells and nearly 20 times higher Ki (i.e., lower potency) in brain membranes. In the compounds within the intermediate range, the Ki values differ widely: from about 2 times (i.e., essentially equipotent) in 1a, 2a, and 8b to about 10 times or more in 7d and 6a. The reasons for these differences are not clear, but they should be taken into account when binding data in the cannabinoid series are compared. In summary, a comparison of the binding potency of several cannabinol derivatives to CB1 and CB2 with  their capacity to inhibit adenylylcyclase has led to the discovery of a CBN metabolite (11-hydroxy-CBN, 4a) as a specific, though not potent, CB2 antagonist and to a new very potent agonist for both CB1 and CB2 (11- hydroxy-CBN-DMH, 4b).

Experimental Section

Chemistry. 1H NMR spectra were measured on a Varian VXR-300S spectrophotometer using TMS as the internal standard. All chemical shifts are reported in ppm. Specific rotations were detected with a Perkin-Elmer 141 polarimeter. Melting points (uncorrected) were determined on a Buchi 530 apparatus. Column chromatography was performed with ICN silica gel 60A. Organic solutions were dried over anhydrous magnesium sulfate. Elemental analyses were obtained for all new compounds (or their acetates) and were (0.4% of the theoretical values. The analyses were performed at the Elemental Analysis Laboratory of the Hebrew University.

¢9-Tetrahydrocannabinol-DMH (2c). Boron trifluoride etherate (5.5 mL) was added to cannabidiol-DMH (3b)(17) (5.7 g, 15.4 mmol) in dry dichloromethane (150 mL) containing magnesium sulfate (1 g), under a nitrogen atmosphere. The reaction mixture was stirred at room temperature for 20 min. A saturated solution of sodium bicarbonate was added until the red color observed during the reaction faded. The reaction mixture was washed with water, separated, dried, and evaporated. The oil obtained was chromatographed on a silica gel column. ¢9-Tetrahydrocannabinol-DMH (2c) (3 g, 53%) was eluted with 4% ether in petroleum ether: 1H NMR (CDCl3) ä 6.38, 6.26 (s, 2H), 6.3 (s 1H), 3.16 (d, 1H, J ) 11.1 Hz), 2.14 (2H), 0.8–1.8 (m); IR (neat) 3300, 2950, 2850, 1620, 1570 cm–1. Anal. (C25H38O2) C,H. ¢9-THC-DMH (2c) is not stable. At room temperature it rapidly becomes violet; on TLC after 0.5 h, numerous new spots are observed.

¢9-Tetrahydrocannabinol-DMH acetate (2d): 1H NMR (CDCl3) ä 6.67, 6.50 (s, 2H), 6.0 (s, 1H), 3.2 (d, 1H), 2.28 (s, 3H), 2.14 (2H), 1.9–0.8 (m); IR (neat) 2900, 1760, 1620, 1560 cm–1. Dehydrogenations of 2d, 5b, 6b,d, and 11b. The dehydrogenations were carried out by heating each compound with sulfur at 238–240 °C, under a nitrogen atmosphere, for ca. 4 h. Each mixture was extracted with ether and evaporated. The residue was chromatographed on a silica gel column using variable concentrations of ether in petroleum ether as eluent. Compound 2d led to cannabinol-DMH acetate (1c); compound 5b gave 4c; compound 6b gave 7d; compound 11b gave 8b. For yields and spectroscopic data, see below.

Cannabinol-DMH acetate (1c): obtained in 24% yield, mp 84–86 °C; 1H NMR (CDCl3) ä 7.8, 7.25, 7.13 (s, 3H), 6.85 (d, 1H, J ) 2.1 Hz), 6.67 (d, 1H, J ) 2.1 Hz), 2.36 (s, 3H), 2.32 (s, 3H), 1.6–0.8 (m); IR (neat) 2930, 1780, 1620, 1560 cm–1. Anal. (C27H36O3) C,H. 11-Hydroxycannabinol-DMH acetate (4c): obtained in 44% yield from 5b; 1H NMR (CDCl3) ä 8.00, 7.25 (s, 3H) 6.85 (d, 1H, J ) 1.5 Hz), 6.68 (d, 1H, J ) 1.5 Hz), 5.10 (s, 2H), 2.34 (s, 3H), 2.10 (s, 3H), 1.6–0.85 (m); IR (neat) 2930, 1780, 1740, 1620, 1540 cm–1.

Cannabinol-11-oic acid acetate (7b): obtained in 22% yield from 6d, mp 149–151 °C; 1H NMR (CDCl3) ä 8.81 (s, 1H), 8.02 (d, 1H, J ) 8.1 Hz), 7.38 (d, 1H, J ) 8.4 Hz), 6.78, 6.64 (s, 2H), 2.42 (s, 3H), 1.7–0.8 (m); IR (neat) 2960, 1760, 1670, 1610, 1570 cm–1. Anal. (C23H26O5) C, H. Cannabinol-11-oic acid-DMH acetate (7d): obtained in 20% yield from 6b; 1H NMR (CDCl3) ä 8.8 (s, 1H), 8.02 (d, 1H, J ) 8.1 Hz), 7.38 (d, 1H, J ) 8.4 Hz), 6.86, 6.75 (s, 2H), 2.42 (s, 3H), 1.7–0.8 (m); IR (neat) 2960, 1780, 1680, 1620, 1570cm–1. Anal (C27H34O5) C,H.

Cannabinol-DMH (1b): Hydrolysis of 1c in ethanolic sodium hydroxide solution gave compound 1b, mp 95–98 °C; 1H NMR (CDCl3) ä 8.11, 7.22, 7.08, 6.36, 6.4 (s, 5H), 2.33 (s, 3H), 1.5–0.7 (m). 11-Hydroxycannabinol-DMH (4b): Compound 4c (526 mg, 1.13 mmol) in dry ether (10 mL) was added to lithium aluminum hydride (123 mg) in dry ether (8 mL). The mixture was boiled under reflux for 2 h. The oil obtained after workup  was chromatographed on a silica gel column (60 g). Elution with 20% ether in petroleum ether gave, after crystallization from pentane, compound 4b (340 mg, 78%): mp 128–130 °C; 1H NMR (CDCl3) ä 8.45, 7.255, 7.253, 6.60, 6.40 (s, 5H), 4.75 (s, 2H), 1.65–0.8 (m, 2H). Anal. (C25H34O3) C,H.

Cannabinol-11-oic acid-DMH (7c). Compound 7d (60 mg) was disolved in 0.6 mL of ethanol. A solution of 60 mg of sodium hydroxide in 0.4 mL of water was added under nitrogen. The solution was stirred at room temperature for 30 min. The solution was acidified with 10% HCl (5 mL); the mixture was extracted with ether, dried with MgSO4, and evaporated. The material obtained was separated on a preparative TLC plate (elution with ether:petroleum ether, 7:3), leading to 41 mg of 7c, 76% yield: 1H NMR 9.2 (s, 1H), 8.0 (d, 1H), 7.38 (d, 1H), 6.61 (s, 1H), 6.45 (s, 1H), 1.7–0.8 (m).

Cannabinol-11-oic acid (7a): 7b was hydrolyzed in ethanolic sodium hydroxide solution as described for compound 7c to yield 7a, 82% yield; 1H NMR ä 9.2 (s, 1H), 8.0 (d, 1H), 7.38 (d, 1H), 6.62 (s, 1H), 6.47 (s, 1H), 1.8–0.8 (m). Compound 11b. A solution of 3-DMH-resorcinol (9b) (3.9 g, 16.5 mmol), ethyl 4-methyl-2-oxocyclohexanecarboxylate (10) (4 g, 21.7 mmol), and POCl3 (3.06 mL) in dry benzene (15 mL) was boiled under reflux, under a nitrogen atmosphere, for 3 h. The solution was washed with NaHCO3 followed by water. After drying and evaporation the oil was chromatographed on a silica gel column. Compound 11b24c (40%) was eluted with 10% ether in petroleum ether: mp 158–160 °C (from pentane); 1H NMR (CDCl3) ä 6.83, 6.63 (s, 2H), 3.4 (dd, 2H), 2.8 (dd, 2H), 2.7–0.8 (m); IR (Nujol) 1680, 1610sh, 1580 cm–1; MS M+ 356.

Compound 8b: dehydrogenation of 11b with sulfur as described above gave compound 8b (72%), mp 184–185 °C (from pentane); 1H NMR (CDCl3) ä 8.82 (s, 1H), 8.3 (d, 1H), 7.38 (d, 1H), 6.94, 6.72 (s, 2H), 2.54 (s, 3H), 1.6–0.8 (m, 19H). Anal. (C23H28O3) C,H. Receptor Binding Assay. a. Binding to Synaptosomal Membranes. Synaptosomal brain preparations were made from whole rat brain as described previously.26,27 Binding of [3H]HU-243 (50.4 Ci/mmol) was assayed in triplicate as described.26,27 In brief, each reaction mixture of 1 mL in siliconized Eppendorf tubes contained 2.4–3.8 íg of synaptosomal membrane protein, 28–48 fmol of [3H]HU-243, and various concentrations of competing unlabeled cannabinoids.  Tubes were incubated at 30 °C for 90 min and centrifuged at 13 000 rpm, and the tips of the tubes containing the pelleted membranes were cut and counted for their radioactivity. b.

Binding to Transfected COS-7 Cells. Two days after transfection (with 5 íg/100 mm of dish plasmids encoding CB1 or CB2) the cells were washed with phosphate-buffered saline, scraped, pelleted, and stored at –80 °C. Cell pellets were homogenized in binding buffer (50 mM Tris-HCl, 5 mM MgCl2, and 2.5 mM EDTA, pH 7.4), and 50 íg protein aliquots were assayed for binding of [3H]HU-243 as described above, except that the final concentration of [3H]HU-243 was 300 pM. For more detailed information see refs(11) and(12). Specific binding was defined as the difference between the amount of radioactivity bound to the pelleted membranes in the absence and presence of 50 nM unlabeled HU-243 (for a) or 1 íM unlabeled 5a (HU-210) (for b) and was typically 70– 80% of the total bound. The Ki values for the various cannabinoids and related compounds were calculated from the competition data according to the formula: Ki ) IC50/1 + ([3H]- HU-243/Kd).29 The Kd values for HU-243 binding were 4526 and 61 pM11 for CB1 and CB2, respectively.

Cell Cultures. Cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 8% fetal calf serum, 2 mM glutamine, nonessential amino acids, 100 units/mL penicillin, and 100 íg/mL streptomycin in a humidified atmosphere consisting of 5% CO2 and 95% air at 37 °C. CHO cells stably transfected with the cDNA of human CB2 receptor9 were described earlier.(11) COS-7 cells in 100 mm dishes were transiently transfected(12) with plasmids encoding rat CB1 25 orhuman CB2 9 (5 íg each) and, when indicated, with 2 íg of the plasmid containing the cDNA of adenylylcyclase type V.12 Twenty-four hours later, the cells were trypsinized and cultured in 24-well plates. After an additional 24 h, the cells  were assayed for adenylylcyclase activity. Transfection efficiency, determined by transfection with the cDNA for â-galactosidase, was 40–80%. Adenylylcyclase Assay. The assay were performed as described.(11),(12),(27) Cells cultured in 24-well plates were incubated for 3 h with 0.25 mL/well fresh growth medium containing 5 íCi/mL [3H]adenine. This medium was replaced with Dulbecco’s modified Eagle’s medium containing 20 mM Hepes (pH 7.4), 1 mg/mL fatty acid-free bovine serum albumin, 0.5 mM 1-methyl-3-isobutylxanthine, and 0.5 mM RO-20-1724. Cannabinoids and forskolin (1 íM) were added, and the cells were incubated at 37 °C for 10 min. The reaction was terminated with 1 mL of 2.5% perchloric acid containing 0.1 mM unlabeled cAMP. Aliquots of 0.9 mL were neutralized with 100 íL of 3.8 M KOH and 0.16 M K2CO3 and applied to a two-step column separation procedure.(12),(27) The [3H]cAMP was eluted into scintillation vials and counted. Background levels (cAMP accumulation in the absence of forskolin) were subtracted from all values and represented less than 10% of forskolin-stimulated cAMP accumulation. Statistical Analysis. Data were anlyzed using the Student’s t-test. Inhibition curves were generated with the Sigma Plot 4.11 program, and the EC50 values were determined using an equation from the ALLFIT program.(30)

Acknowledgment – This work was supported by NIDA grants DA 9789 and DA 8169 (to R.M.) and a grant by the Israel Science Foundation (to Z.V. and R.M.), the Forchheimer Center for Molecular Genetics (to Z.V.), and the Henry and Anne Reich Research Fundfor Mental Health (to Z.V.). M.B. was supported by a fellowship from the Israeli Ministry of Science and Technology. Z.V. is the incumbent of the Ruth and Leonard Shimon Chair of Cancer Research. R.M. is the Lionel Jacobson Professor Medicinal Chemistry. The authors thank Drs. T. Bonner (NIMH, NIH, Bethesda, MD), S. Munro (Cambridge, U.K.), and T. Pfeuffer (Dusseldorf, Germany) for their kind donations of plasmids encoding CB1, CB2, and adenylylcyclase type V, respectively.