C-terminal truncated cannabinoid receptor 1 coexpressed with G protein trimer in Sf9 cells exists in a precoupled state and shows constitutive activity Chandramouli Reddy Chillakuri, Christoph Reinhart and Hartmut Michel Department of Molecular Membrane Biology, Max-Planck-Institute for Biophysics, Frankfurt ⁄ Main, Germany Cannabinoid receptors belong to the seven transmem- brane G protein coupled receptor (GPCR) family. Two different subtypes have been reported in humans, namely cannabinoid receptor 1 (CB1) [1] and cannabi- noid receptor 2 (CB2) [2]. A splice variant of CB1, called CB1a, has also been identified to be expressed in low levels in rodent brain [3]. Cannabinoid receptors form the site of action for the active ingredients of marijuana (D 9 -tetrahydrocannabinol, D 9 THC), ananda- mide being the important endocannabinoid. CB1 is present predominantly in the nerve axons of the cen- tral nervous system, is known to be neuroprotective in its function and thus forms an important target in the pharmaceutical industry [4]. CB1 is coupled to the G i ⁄ o family of heterotrimeric G proteins. Activation of the CB1-mediated G i ⁄ o proteins inhibits adenylyl cyclases to reduce cAMP production, inhibit calcium channels and increase inwardly rectifying potassium currents. The modulation of ion channels has been shown to be independent of cAMP production, indicating that G proteins, especially Gbc, may inter- act directly with the effector molecules [5,6]. An earlier concept states that agonist-bound GPCR alone can couple to G proteins to transduce the signal. However, this rather historical hypothesis is opposed today because of several reports confirming Keywords cannabinoid receptor; G protein coupled receptor; G proteins; membrane proteins; signal transduction Correspondence H. Michel, Department of Molecular Membrane Biology, Max-Planck-Institute for Biophysics, Max-von-Laue Str.3, D-60438 Frankfurt ⁄ Main, Germany Fax: +49 69 6303 1002 Tel: +49 69 6303 1001 E-mail: Hartmut.Michel@mpibp-frankfurt. mpg.de (Received 23 July 2007, revised 15 Septem- ber 2007, accepted 8 October 2007) doi:10.1111/j.1742-4658.2007.06132.x We have investigated the existence of a precoupled form of the distal C-ter- minal truncated cannabinoid receptor 1 (CB1-417) and heterotrimeric G proteins in a heterologous insect cell expression system. CB1-417 showed higher production levels than the full-length receptor. The production lev- els obtained in our expression system were double the values reported in the literature. We also observed that at least the distal C-terminus of the receptor was not involved in receptor dimerization, as was predicted in the literature. Using fluorescence resonance energy transfer, we found that CB1-417 and Ga i1 b 1 c 2 proteins were colocalized in the cells. GTPcS bind- ing assays with the Sf9 cell membranes containing CB1-417 and the G pro- tein trimer showed that the receptor could constitutively activate the Ga i1 protein in the absence of agonists. A CB1-specific antagonist (SR 141716A) inhibited this constitutive activity of the truncated receptor. We found that the CB1-417 ⁄ Ga i1 b 1 c 2 complex could be solubilized from Sf9 cell mem- branes and coimmunoprecipitated. In this study, we have proven that the receptor and G proteins can be coexpressed in higher yields using Sf9 cells, and that the protein complex is stable in detergent solution. Thus, our sys- tem can be used to produce sufficient quantities of the protein complex to start structural studies. Abbreviations B max , maximum binding capacity; CB1, cannabinoid receptor 1; CB2, cannabinoid receptor 2; CFP, cyan fluorescent protein; CHS, cholesterol hemisuccinate; DM, decylmaltoside; FRET, fluorescence resonance energy transfer; GPCR, G protein coupled receptor; K d , equilibrium dissociation constant; m.o.i., multiplicity of infection; D 9 THC, D 9 -tetrahydrocannabinol; YFP, yellow fluorescent protein. 6106 FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS the precoupled form of GPCR and G proteins. In a recent study, using fluorescence resonance energy trans- fer (FRET), several receptors were identified, such as the muscarinic receptor M4, the adrenergic receptor a2A, the adenosine receptor A1 and the dopamine receptor D2, in a precoupled form to the G protein tri- mer (Ga o b 1 c 2 ) [7]. It was suggested that GPCR dimers and the G protein heterotrimer are present in cell mem- branes in a resting state as a pentameric complex in the absence of agonists. However, it may not be true that this whole complex is resting and inactive, and can only be activated by agonists. Reports constantly support the constitutive activity of GPCRs, indicating that GPCRs have some basal activity in the cell even in the absence of agonists. Inverse agonists (antagonists) to several GPCRs have shown that these ligands can reverse this basal activation of GPCRs, demonstrating the existence of constitutively active receptors [8]. Solubilization of the cannabinoid receptor from rat brain and GTP binding studies initiated the concept of the presence of a stable CB1 ⁄ G protein complex [9]. The development of the CB1-specific antagonist SR 141716A led to the identification of constitutively active receptors [10]. The existence of such an active form was tested, and it was found that CB1 can sequester G proteins from a common pool, making them unavailable to other GPCRs present in the cell [11]. Meanwhile, it was found that a CB1 C-terminal peptide (CB1-401–417) was able to activate the specific G proteins in brain [12]. Further, a C-terminal trun- cated CB1 (CB1-417) was found to have enhanced ability to sequester G proteins and exhibited increased constitutive activity [13]. Recently, it has been reported that different subtypes of Ga i coimmunoprecipitate with CB1 from N18TG2 cell membranes in the absence of exogenously added agonists [14]. Most of the results discussed above were either performed on native tissue or mammalian cells. In this work, for the first time, we investigated the existence of a precoupled form of truncated CB1 in a heterologous Sf9 insect cell expression system. Insect cell expression systems are cheaper than mammalian expression systems, and are therefore more feasible for the large-scale production of receptor required for structural determination. Results Functional production of CB1 The C-terminal truncated CB1 (CB1-417) was pro- duced in a functional form in Sf9 insect cells (the gene constructs used in this study are shown in Fig. 1). Analysis of the protein produced, by immunoblotting with antiflag M2 IgG, showed (Fig. 2) a monomeric band at 47 kDa and an oligomeric band specific to the cannabinoid receptor at a size of approximately Strep-tagIICB1 HisFlag P PH Melittin Tev 1 to 472 aa Strep-tagIICB1 HisFlag P PH Melittin Tev 1 to 417 aa YFPCB1 HisFlag P PH Melittin Tev 1 to 417 aa G?i1CFPG?i1 P PH Strep-tagIICB1 HisFlag P PH Melittin Tev 1 to 472 aa Strep-tagIICB1 HisFlag P PH Melittin Tev 1 to 417 aa YFPCB1 HisFlag P PH Melittin Tev 1 to 417 aa CFPG α i1 G α i1 P PH Fig. 1. Gene constructs used for the expression in insect cells. The four gene constructs were prepared using the basic pVL baculovirus transfer vector. CFP, cyan fluorescent protein; Flag, flag epitope used for immunoblotting; His, decahistidine tag; P PH , polyhedrin promoter; Strep-tagII, used for the purification of the receptor; YFP, yellow fluorescent protein. The names of each construct mentioned in this work are given before each graphic representation. C. R. Chillakuri et al. Precoupled form of human CB1 and G protein trimer FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS 6107 160 kDa. A similar higher oligomeric band was reported for full-length CB1 between 160 and 200 kDa. Wager–Miller et al. [15] took this oligomeric form of the cannabinoid receptor as an example to explain the dimerization of GPCRs. Radioligand binding assay showed saturation binding on the cell membranes (Fig. 3). It was found that the use of 10 multiplicity of infection (m.o.i.) virus and incubation for 72 h resulted in the best production levels (data not shown). The production levels of both constructs of this receptor were much higher than the values reported so far in the literature. The full-length receptor (FHTCB1StII) gave a maximum binding capacity (B max ) of 39.7 ± 1.3 pmolÆmg )1 , and the C-terminal truncated version [FHTCB1(417)StII] gave a B max value of 52 ± 3.09 pmolÆmg )1 . The equilibrium dissociation constants (K d ) observed were 2.5 and 3.6 nm, which fall within the range of the values reported earlier. The best value reported so far is 24.5 pmolÆmg )1 for N-terminal histi- dine-tagged CB1 in the Sf21 cell line [16]. Therefore, the gene constructs used in this report with a StrepII tag on the C-terminus gave a two-fold increase in pro- duction levels compared with the constructs used in the literature. The truncated receptor showed a 30% higher production than the full-length receptor. Determining the colocalization of the CB1 ⁄ G protein complex by FRET The cannabinoid receptor exists in a precoupled form to G proteins. The receptor R can exist in an RG GDP 175 83 62 47.5 32.5 KDa M 1 2 Fig. 2. Immunoblotting of full-length and truncated CB1. The full- length and distal C-terminal truncated CB1 were produced in Sf9 cells. Thirty micrograms of cell membrane were used to run SDS ⁄ PAGE, and were immunoblotted using antiflag M2 IgG. The lower band is the monomeric band and the upper band is the oligomeric band (presumably tetrameric). Lane 1, marker in kDa; lane 2, full-length CB1; lane 3, truncated construct of CB1 (CB1- 417). 0 0.5x10 4 1x10 4 1.5x10 4 2.0x10 4 2.5x10 4 CB1 [SR 141716A] pmol/mg 40 30 20 10 0 0 0.2x10 4 0.6x10 4 1.0x10 4 1.4x10 4 CB1(417) [SR 141716A] pmol/mg 0 0.5x10 4 1x10 4 1.5x10 4 2.0x10 4 2.5x10 4 40 35 30 25 20 15 10 5 0 CB1 [SR 141716A] pmol/mg 50 0 0.2x10 4 0.6x10 4 1.0x10 4 1.4x10 4 CB1(417) [SR 141716A] pmol/mg Fig. 3. Saturation binding curves of full-length and truncated CB1. One microgram of cell membrane containing either full-length or truncated CB1 was used to determine the production level of the protein. Eight concentrations (200 p M to 25 nM) of the radioactive cannabinoid ligand SR 141716A were used. Each point is the specific binding calculated from the mean of triplicates of the positive reaction and duplicates of the negative reaction. The x-axis represents the concentration of the radioactive ligand in picomoles. The y-axis represents the specific bind- ing in pmolÆmg )1 of total cell membranes. The full-length receptor gave a B max value of 39.7 ± 1.3 pmolÆmg )1 and K d value of 2.5 nM. The truncated receptor gave a B max value of 52 ± 3.09 pmolÆmg )1 and K d value of 3.6 nM. Precoupled form of human CB1 and G protein trimer C. R. Chillakuri et al. 6108 FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS (GDP-bound form) or RG – (nucleotide-lacking form) in addition to the free R form [17]. These G protein- bound forms of the receptor are believed to be respon- sible for the sequestration and constitutive activity mechanisms of the receptor. In this study, we investi- gated the presence of such a precoupled complex in heterologous Sf9 cells using the FRET technique. FRET occurs when a donor (cyan fluorescent protein, CFP) transfers its energy obtained on excitation to an acceptor (yellow fluorescent protein, YFP) by dipole– dipole interactions. This phenomenon occurs when FRET partners specifically interact and are present within a close proximity of less than 100 A ˚ . In this study, we used FHTCB1(417)-YFP (accep- tor) and G i1 -CFP (donor) fusion proteins. Sf9 cells coexpressing FHTCB1(417)-YFP, G i1 -CFP and b 1 c 2 were imaged using a laser scanning confocal micro- scope. No cannabinoid ligands were included in the cell cultures or buffers. Nevertheless, both proteins were found to be colocalized in the cell membrane (Fig. 4). The fluorescence energy transfer between the proteins was investigated by acceptor bleaching and donor dequenching experiments. The conditions used for bleaching, with a 515 nm laser, were optimal for > 90% acceptor bleaching and minimal donor bleaching. When the acceptor protein was bleached, there was an increase in donor fluorescence. The increase in donor fluorescence calculated for 10 cells was 7% ± 2%. As a negative control CB1-CFP and histamine1 receptor-YFP were coexpressed to almost equal levels and the experiment was repeated in the same way as above. These two GPCRs have not been reported to form a dimer complex. An increase of less than 2% was detected, which could be a result of the random collision of fluorescent molecules in the membrane. This was considered as the background signal. Constitutive activity of the truncated cannabinoid receptor The cannabinoid receptor has been shown to exhibit constitutive activity, thereby transducing the biological signal, even in the absence of ligand [18]. Truncation of the distal C-terminal tail has been shown to enhance the constitutive activity and sequestration ability of the cannabinoid receptor [13]. Using the patch clamp tech- nique on neurones, it has been shown that the C-termi- nal distal tail constrains the receptor from interacting with G proteins. In this study, we investigated the con- stitutive activity of the truncated cannabinoid receptor in heterologous Sf9 cell membranes. We used fluores- cent and radioactive GTPcS binding assays to observe the constitutive activity. In the fluorescence experi- ment, the cell membranes containing CB1-417 were incubated with purified Ga i1 or Ga sL proteins. CB1- 417 enhanced the fluorescent GTPcS binding to the Ga i1 protein, whereas no effect was seen when Ga sL protein was used, which is not a physiological partner to the cannabinoid receptor (Fig. 5A). This increase in GTPcS binding was observed even in the absence of agonist. Wild-type Sf9 cell membranes were included in the reactions to monitor the basal activity of the Ga proteins used. Purified Ga i1 protein showed an increased activity in the presence of cell membranes, although the reason was unclear. However, this increase in GTPcS binding was not additive with increasing wild-type membrane concentration, unlike the CB1-417 membrane, which showed an additive effect on GTPcS binding (Fig. 5B). GDP (10 lm) was CFP YFP Overlap Post-bleach Pre-bleach Fig. 4. Confocal images showing the colo- calization of the receptor and G protein. Sf9 cells producing the CB1-417-YFP fusion pro- tein and G i1 -CFP fusion protein were imaged using a laser scanning confocal microscope. CFP was excited with a 458 nm laser and YFP with a 515 nm laser. Images were col- lected using the filters 475–525 nm (CFP) and > 530 nm (YFP). Overlap images show the colocalized receptor and the G protein. YFP was bleached using a 515 nm laser in a donor dequenching experiment. Donor de- quenching gave a 7% increase in acceptor fluorescence. The white rectangle in the images shows the area bleached using the laser. C. R. Chillakuri et al. Precoupled form of human CB1 and G protein trimer FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS 6109 used in the reactions in order to reduce the basal activ- ity of the purified G proteins. Sf9 cell membranes containing heterotrimeric G pro- teins alone or together with the cannabinoid receptor were used for the radioactive GTPcS binding assay. The production of all subunits was confirmed by immunoblotting against each subunit. There was a sig- nificant increase in GTP binding when the receptor was coexpressed together with the G proteins (Fig. 6). This increase in the absence of ligand shows the constitutive activity of the receptor. The antagonist AM251 inhibited this GTP binding to G proteins. The agonist WIN 55,212–2 increased GTP binding to a lesser extent. These ligand-dependent effects were not seen in the membranes lacking the receptor. Similar results have been reported in [19], where cannabinoid receptors produced in Sf9 cell membranes and G pro- teins purified from brain were reconstituted. In the present study, a defined receptor and G protein complex was used rather than a whole pool of Ga subunits. Coimmunoprecipitation of CB1-417 and the G protein complex Sf9 cell membranes containing the Ga i1 b 1 c 2 protein complex together with FHTCB1(417)StII were used for coimmunoprecipitation (Fig. 7). The membranes Time (min) 0312 45 Time (min) 0312 45 Intensity (AU)Intensity (AU) 24 A B 23 21 19 17 15 13 11 34 32 28 24 20 16 12 Fig. 5. Fluorescent Bodipy GTPcS binding assay. (A) Specific increase in GTPcS binding to the G i1 (r) protein and not the G sL (j) protein in the presence of membranes containing the truncated cannabinoid receptor. Symbols s and d represent GTPcS binding to pure G i1 and G sL in the absence of cell membranes. Symbols e and h represent GTPcS binding to G i1 and G sL in the presence of Sf9 cell membranes. (B) Increase in GTPcS binding to G i1 is addi- tive because of CB1 and not just because of the cell membranes. Doubling the concentration of cell membranes with CB1 (r) adds to the GTPcS binding, whereas Sf9 cell membranes (e) do not show a similar effect. Symbols n and m represent the binding of GTPcStoG i1 protein in the presence of a 1· concentration of Sf9 membranes or CB1 membranes, respectively. (The bullets used in this figure are not data points, but are used to distinguish between the different spectra.) 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 G CB1 + G CB1 + G + A CB1 + G + IA Dpm (GTP [S 35 ]) Fig. 6. Radioactive GTPcS binding assay. Thirty micrograms of cell membrane containing G protein trimer only (G) or G protein trimer coexpressed with truncated CB1 (R) were used to estimate GTPcS binding. Coexpression of the receptor together with G proteins increased GTPcS binding. The cannabinoid receptor agonist WIN 55,212–2 (A) further increased GTPcS binding. This binding was inhibited by the cannabinoid selective antagonist AM251 (IA). Precoupled form of human CB1 and G protein trimer C. R. Chillakuri et al. 6110 FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS containing only G proteins or receptor + G proteins were solubilized using a 1% decylmaltoside (DM) + 0.2% cholesterol hemisuccinate (CHS) mix- ture for 1 h. The presence of CHS during the solubili- zation and purification of GPCRs has been demonstrated to be crucial in retaining the functional- ity of the receptor [20]. We have had a similar experi- ence with other GPCRs in our laboratory (C. R. Chillakuri et al., unpublished data). In a recent report, CHS was used in the purification of CB2 [21]. Mukho- padhyay & Howlett [14] used Chaps detergent to solu- bilize the CB1 ⁄ G protein complex from N18TG2 neuroblastoma cell membranes. Immunoprecipitation of the receptor ⁄ G protein complex was performed as described in Experimental procedures. The eluted pro- tein from the antiflag M2 agarose matrix was analysed using different antibodies. The immunoblot with antiflag M2 IgG showed the CB1 band at 47 kDa. Anti-histidine tag immunoblot showed the receptor band as well as the bc dimer (c subunit has the histidine tag on the N-terminus) at  34 kDa. The immunoprecipitated sample from the membranes containing only G proteins did not show any specific band in either immunoblot. Anti-G i1 immunoblot showed a faint band in the negative control, indicating a nonspecific interaction of Ga i1 with the matrix. How- ever, the signal in the positive control was higher, and therefore it was concluded that the Ga i1 protein was specifically bound to the receptor. The presence of bc subunits only in the positive reaction supports this conclusion. Discussion One of the prime limiting factors for structural studies of GPCRs is the availability of material. Obtaining sufficient quantities of pure and active receptor is in itself a challenge for many GPCRs, including CB1. In this study, we focused on the overproduction of CB1 and the investigation of the precoupled form of this receptor with G proteins. Structural studies of the receptor ⁄ G protein complex are needed to understand the mechanism of interaction between the partners. Instead of producing the subunits separately and using them for cocrystallization, it may be worthwhile to isolate the ternary complex for structural studies. Another proposal behind the choice of the recep- tor ⁄ G protein complex for three-dimensional crystalli- zation is that the G protein trimer increases the hydrophilic portion of the complex, and thus enhances the chances of crystallization of GPCR, an integral membrane protein, which has been a challenge for crystallographers. We used a heterologous insect cell expression system for the overproduction of CB1. We obtained two-fold higher production levels for this receptor than those reported in the literature [16]. Truncation of the distal C-terminal tail of CB1 has been reported to increase the constitutive activity and sequestration tendency of the receptor [13]. In this study, we observed that this truncation also increases the production levels of the receptor in Sf9 insect cells. The truncated receptor was produced in insect cell culture at up to 500 lgÆL )1 ( 52 pmolÆmg )1 of 47 kDa protein). These moder- ately higher production levels provide better scope for producing more protein required for structural studies. Truncation of the receptor did not hinder ligand bind- ing to the receptor, indicating that the receptor was functional. An important observation from the immunoblot of the truncated receptor was that this truncated receptor also exists as an oligomer, as does full-length CB1. Wager-Miller et al. [15] reported that the C-terminal tail may be important in the assembly of the oligomer. Our results show that at least the distal C-terminal tail (418–472) is not involved in CB1(417) G +G Anti-Flag M2 IgG Anti-His tag IgG Anti-His tag IgG Anti-G i1 /G i2 IgG Receptor Receptor G G i1 Fig. 7. Coimmunoprecipitation. Sf9 cell membranes containing CB1-417 and the Ga i1 b 1 c 2 trimer complex were solubilized using a mixture of DM and CHS. The complex was immunoprecipitated using antiflag M2 IgG agarose matrix. The matrix was washed thrice and the bound protein was eluted by denaturation with SDS gel loading buffer. Immunoprecipitated samples: lane 1, cell membrane containing G protein trimer only; lane 2, cell membrane containing both receptor and G protein trimer. The antibodies used to identify the different subunits of the complex are denoted on the left side of the image. The subunit that was identified is men- tioned on the right side of the immunoblot image. The anti-G i1 ⁄ G i2 IgG immunoblot showed that the Ga subunit exhibits nonspecific binding to the matrix. However, the intensity of the Ga subunit was much higher when the receptor was present in the solubilizate. C. R. Chillakuri et al. Precoupled form of human CB1 and G protein trimer FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS 6111 oligomerization, as the truncated protein was observed to oligomerize. According to the classical hypothesis, the presence of agonist is necessary for receptor ⁄ G protein complex formation and activation. In contradiction to this hypothesis, some receptors, such as the j-opioid recep- tor [22], dopamine receptor D2, adrenergic receptor a2A, muscarinic receptor M4 and adenosine receptor A1, have been found to exist as complexes with their corresponding G proteins prior to ligand activation [7]. CB1 was found to exist in a precoupled form in N18TG2 cells, even in the absence of ligands [14]. In this study, we investigated the existence of a precou- pled form of truncated CB1 in Sf9 insect cells. A FRET experiment on Sf9 cells producing the truncated cannabinoid receptor and the G protein heterotrimer showed that these proteins were colocalized. The fluo- rescent and radioactive GTPcS binding experiments showed that the truncated receptor produced in Sf9 cells retained the ability to activate G proteins in the absence of ligands. The fluorescent GTPcS binding experiment showed the specificity of the cannabinoid receptor to Ga i1 protein and not Ga sL protein. The radioactive GTPcS binding experiment showed that the receptor was constitutively active and the antago- nist AM251 inhibited the basal activity of the receptor. Similar results have been reported previously by Glass & Northup [19] using the G proteins purified from bovine brain. The relatively small increase in GTPcS binding to G protein on addition of the agonist WIN 55,212–2 in our experiments can be explained by the presence of low levels of the receptor conformational state recognized by the agonist. We observed that the maximum binding of antagonist (AM251) to the recep- tor was five times higher than the maximum binding of agonist (CP-55 940) (data not shown). This shows that most of the receptor produced in insect cells is in a conformation not recognized by the full agonist. A similar result was reported by Xu et al. [16]. Another reason may be that the agonist-activated conforma- tional state found in Sf9 cell membranes prefers other subtypes of Ga i ⁄ o protein than the Ga i1 used. In this study, we also investigated the possibility of the solubi- lization and isolation of the whole receptor ⁄ G protein complex produced in insect cells. The coimmunopre- cipitation experiment showed that the complex could be solubilized using the mild nonionic detergent DM in combination with CHS. These mild conditions are necessary to retain the function of the receptor and the G protein complex. Taken together, our results confirm that the distal C-terminal truncated CB1 can be produced in a func- tional form in Sf9 insect cells in higher yields than those obtained previously. This receptor exhibits con- stitutive activity, and it is also possible to coimmuno- precipitate the whole receptor ⁄ G protein complex in the absence of any ligands. This paves the way for fur- ther investigations to determine the possibility of stabi- lizing and purifying this complex to homogeneity, so that it can be used for crystallization, in order to obtain a better understanding of the interaction between the partners, and the mechanism of signal transduction. Experimental procedures Chemicals and reagents The general laboratory chemicals used were of analytical grade and were purchased from Roth (Carl Roth & Co. KG, Karlsruhe, Germany), Merck (Merck KGaA, Darms- tadt, Germany) and Fluka ⁄ Sigma (Sigma-Aldrich Chemie GmbH, Diesenhofen, Germany). Bodipy FL-GTPcS was purchased from Molecular Probes (Eugene, OR, USA). Radioactive cannabinoid agonist [ 3 H]CP-55 940 was obtained from Perkin Elmer LAS, (Deutschland) GmbH ⁄ Rodgan-Ju ¨ gesheim, Germany, and antagonist [ 3 H]SR 141716A was purchased from Amersham Biosciences. GTPc[S 35 ] was obtained from Perkin Elmer Life Sciences. Unlabelled cannabinoid ligands WIN 55,212–2 mesylate and AM251 were purchased from Tocris (Bristol, UK). DM was purchased from Glycon Biochemicals, Luckenwal- de, Germany. The protease inhibitors used were obtained from Biomol Feinchemikalien GmbH, Hamburg, Germany. Antiflag M2 IgG conjugated to alkaline phosphatase and anti-polyhistidine antibody conjugated to alkaline phospha- tase were obtained from Sigma-Aldrich Chemie GmbH (Munich, Germany). Anti-Ga i1 ⁄ Ga i2 IgG was purchased from Calbiochem (Merck KGaA, Darmstadt, Germany). Anti-flag M2 IgG agarose was obtained from Sigma- Aldrich Chemie GmbH. Cloning The gene encoding CB1 was cloned into modified pVL1393 baculovirus transfer vector with the mellitin signal sequence. The forward primer for the CB1 gene with the BamHI restriction site was 5¢-GC G GAT CC G ACC ATG GCG AAG TCG ATC CTA GAT GGC-3¢. The reverse primer for the full-length CB1 gene, with EcoRI and NotI restriction sites, was 5¢-GAA T GC GGC CGC TCA CTT TTC GAA TTG AGG GTG CGA CCA GAA TTC AGC CTC GGC AGA CGT GTC TGT GGA-3¢, which contains the StrepII tag between the EcoRI and NotI sites. The reverse primer for the truncated receptor with the EcoRI site was 5¢-CCA GAA TTC GCC TTC ACA AGA GGG AAA CAT-3¢. The full-length receptor gene (1–472 amino Precoupled form of human CB1 and G protein trimer C. R. Chillakuri et al. 6112 FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS acids) was cloned between the BamHI and NotI sites of the vector to prepare pVLMelFHTevCB1StrepII. The truncated receptor (1–417 amino acids) was cloned into the BamHI and EcoRI sites of the above vector to prepare pVLMelFHTevCB1(417)StrepII. In the CB1-417-YFP con- struct, the YFP gene was cloned between the EcoRI and NotI sites of the truncated construct. G i1 -CFP was prepared by introducing CFP between the NcoI and XbaI sites. These restriction sites (ACC ATG GTG TCT AGA) were generated in the G i1 protein between amino acids Ser62 (TCA) and Glu63 (GAA) by overlap PCR. The G i1 -CFP gene was cloned into the pVL vector (no mellitin sequence) using BamHI and EcoRI. Digestion of DNA and ligation were performed according to the protocols in the New England Biolabs GmbH, Frankfurt am Main, Germany. The DH5a strain of Escherichia coli was used to amplify and clone the DNA constructs. Ga i1 and Ga sL genes were cloned into the pDEST14 vector using Gateway cloning technology (Invitrogen), for protein production in E. coli. Recombinant virus production and selection Recombinant virus for the DNA constructs was prepared according to the protocols given in Invitrogen’s baculovirus expression system catalogue. Three to five positive plaques (according to X-gal selection) for each gene were selected, and virus was produced by infecting Sf9 cells. The virus from these clones was used to infect the 2 · 10 6 cells in a tissue culture plate. After 4 days of incubation, the cells were harvested by spinning in a centrifuge; 2 · 10 5 cells from each clone were lysed using 1% SDS in a buffer con- taining protease inhibitors and DNase. This lysate was used for analysis on SDS ⁄ PAGE. The gel was immunoblotted using anti-flag M2 IgG to confirm recombinant protein production. The clone showing the best expression profile was selected, and the corresponding virus was amplified and stored at 4 °C. The titre for the virus was calculated using the 96-well plate end point dilution assay [23]. Virus for human Ga i1 and bovine Gb 1 c 2 was obtained from for- mer laboratory members [24]. Cell culture and cell membrane preparation Sf9 cells were grown in TNMFH medium (C.C. pro GmbH, Germany) containing 5% fetal bovine serum (PAA Laboratories GmbH, Germany). The cells were maintained in a tissue culture flask. For suspension culture, cells from the tissue culture flask were added to the medium contain- ing 0.1% Pluronic F68 at a density of 1 · 10 5 cellsÆmL )1 . The conical flask was incubated in a shaker at 27 °C and 125 r.p.m. Cells were grown to a density of 2 · 10 6 cell- sÆmL )1 , and harvested using sterile centrifuge tubes by cen- trifugation at 1000 g. Old medium was removed and an equal volume of fresh medium was added to the cells and resuspended gently using a pipette. Virus was used at 10 m.o.i. for protein production. Cells were incubated for 3 days before harvesting. Sf9 cells were harvested by centri- fugation. The pellet was resuspended in ice-cold lysis buffer (20 mm Tris ⁄ HCl pH 8.0, 100 mm NaCl, 5 mm MgCl 2 , 250 mm sucrose) containing protease inhibitors (1 lm E64, 5 lgÆmL )1 leupeptin, 2 lgÆ mL )1 pepstatin A, 10 l g ÆmL )1 aprotonin). The cells were broken in a Parr Bomb for 1 h at 35 kg ⁄ cm 2 . The suspension was collected and centrifuged at 2000 g to remove the unbroken cells. The turbid super- natant was ultracentrifuged at 100 000 g to sediment the cell membranes. The membrane pellet was resuspended in resuspension buffer (20 mm Tris ⁄ HCl pH 8.0, 100 mm NaCl, 5 mm MgCl 2 , 10% glycerol) and homogenized using a potter. The homogenate was aliquoted and stored at ) 80 °C in a freezer. Radioactive cannabinoid ligand binding The total protein concentration was estimated by a bicinch- oninic acid protein assay kit (Pierce, Rockford, IL, USA); 1 lg of Sf9 cell membranes was used for each reaction in a saturation binding assay. Membranes were added to bind- ing assay buffer A (20 mm Tris ⁄ HCl, 5 mm MgCl 2 ,1mm EDTA, 1% BSA). Eight concentrations of [ 3 H]SR 141716A were used in the saturation binding assay. Triplicates of positive reactions (radioactive ligand only) and duplicates of negative reactions (radioactive ligand + 10 lm cold ligand AM251) were set up in 1.5 mL Eppendorf tubes for each radioactive ligand concentration. The reactions (250 lL) were incubated for 1 h at 30 °C and filtered over glass fibre filters (GF-B) from Whatmann GmbH (Dassel, Germany). The filters were washed three times with warm (30 °C) binding assay buffer. The filters were collected in 5 mL radioactivity counting tubes and 4.5 mL of scintillant (Roth) was added. The radioactivity was measured in terms of disintegrations per minute (d.p.m.). The specific d.p.m. (mean positive reactions ) mean negative reactions) was used to calculate the receptor concentration in pmolÆmg )1 . A Kaleida graph (Synergy software) was used to plot the receptor binding sites versus radioactive ligand concentra- tion in a nonlinear regression curve using the following for- mula: specific binding Y ¼ (M 1 · M 0 ) ⁄ (M 2 + M 0 ); M 1 ¼ 1; M 2 ¼ 1. M 1 is B max and M 2 is K d . Determination of the colocalization of receptor and G protein by FRET Cells were allowed to attach to the glass surface before the images were taken. An LSM-510 Meta confocal microscope (Carl-Zeiss AG, Oberkochen, Germany), fitted with a · 60 oil objective, was used to collect the images. CFP was excited with a 458 nm laser and images were obtained using a bandpass filter from 475 to 525 nm. YFP was excited with a 515 nm laser and images were obtained using a longpass filter above 530 nm. The acceptor bleaching C. R. Chillakuri et al. Precoupled form of human CB1 and G protein trimer FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS 6113 experiment was performed by sequential scanning. A region of the cell along the surface was selected manually using the software provided with the microscope, and was bleached using a 515 nm laser at 80% laser power. Pre- bleach and postbleach images were taken for both proteins using their respective filters. The difference in the donor flu- orescence was calculated from the background subtracted images. The percentage FRET was calculated using the fol- lowing formula: (intensity of postbleach image ) intensity of prebleach image) ⁄ intensity of prebleach image. Ten cells were observed and the mean value was calculated. Cross- talk between the channels was corrected using the cells pro- ducing only CFP or only YFP. The laser power, pinhole size and detector gain were chosen optimally to avoid satu- ration of the images. Fluorescent Bodipy FL-GTPcS binding assay To measure the basal G protein activation by the cannabi- noid receptor, a fluorescent Bodipy FL-GTPcS binding assay was performed. Sf9 cell membranes coexpressing CB1- 417 and Gb 1 c 2 were used. Ga i1 and Ga sL were produced in E. coli and purified. The cell membrane concentration was chosen to contain 20 nm receptor (as estimated by radioli- gand binding) in the final reaction. Purified G protein in buffer B (20 mm Tris ⁄ HCl pH 8.0, 100 mm NaCl, 5 mm MgCl 2 ,10lm GDP, 0.25 mm dithiothreitol) was used at a five-fold molar excess (100 nm) to the receptor. A 20 lL membrane ⁄ G protein reaction mixture (10·) was made and incubated at 25 °C for 45 min. This reaction mixture was diluted in assay buffer (buffer B + 500 nm Bodipy FL- GTPcS), and the fluorescence was monitored immediately using a time drive program for 5 min in an LSM-50 lumi- nescence spectrophotometer (Perkin Elmer Life Sciences). The fluorescent ligand was excited at 485 nm and the emis- sion was monitored at 520 nn. In a negative control experi- ment, wild-type Sf9 cell membranes were used. Radioactive GTPc[S 35 ] binding assay Radioactive binding assay was used as a complementary experiment to observe the constitutive activity of the recep- tor. Sf9 cell membranes containing CB1-417 and Ga i1 b 1 c 2 were used. Cell membranes containing only G protein sub- units were used as a negative control. Thirty micrograms of total cell membrane were used for each reaction. Cell mem- branes were diluted in buffer C (buffer B containing 0.5% BSA) to prepare a reaction volume of 200 lL. Agonist or antagonist (4 lm) dissolved in dimethylsulfoxide was used as required in the reactions containing ligands. Radioactive GTPcS was diluted in buffer C and added to each reaction to obtain a final concentration of 4 nm. The reactions were incubated at 30 °C for 1 h. The reactions were filtered under vacuum, over glass fibre filters wetted with buffer C. The filters were washed thrice with buffer C. The radio- activity on these filters was counted using a b-counter calibrated with the 14 C isotope. Coimmunoprecipitation Cell membranes containing the receptor and G proteins were solubilized using a mixture of 1% DM and 0.2% CHS at 4 °C for 1 h. The supernatant was clarified by ul- tracentrifugation and incubated with 50 lL of anti-flag M2 IgG agarose (rinsed with buffer B) at 4 °C for 1 h. The supernatant was removed and the antibody matrix was washed three times with buffer B containing 0.2% DM and 0.04% CHS. The matrix was resuspended in SDS gel load- ing buffer to elute the protein bound to the antibody by denaturation. This eluate was used to run an SDS gel and analysed by immunoblotting. Anti-M2 IgG was used to identify the receptor. Anti-polyhistidine tag IgG was used to detect both the receptor and the Gbc dimer (histidine tag on the C-terminus of the c subunit). Ga i1 ⁄ i2 antibody was used to detect Ga i1 protein. Acknowledgements We would like to thank the UMR cDNA resource centre for the kind donation of the cDNA for CB1. We thank Heinz Schewe (Bio-zentrum, University of Frankfurt, Germany) for his help in using the laser scanning confocal microscope. This work was funded by the Max-Planck-Gesselschaft and Sanofi-Aventis. References 1 Matsuda LA, Lolait SJ, Brownstein MJ, Young AC & Bonner TI (1990) Structure of a cannabinoid receptor and functional expression of the cloned cDNA. Nature 346, 561–564. 2 Munro S, Thomas KL & Abu-Shaar M (1993) Molecu- lar characterization of a peripheral receptor for cannabi- noids. Nature 365, 61–65. 3 Rinaldi-Carmona M, Calandra B, Shire D, Bouaboula M, Oustric D, Barth F, Casellas P, Ferrara P & Le Fur G (1996) Characterization of two cloned human CB1 cannabinoid receptor isoforms. J Pharmacol Exp Ther 278, 871–878. 4 Szabo B & Schlicker E (2005) Effects of cannabinoids on neurotransmission. Handbook Exp Pharmacol 168, 327–365. 5 Mackie K, Lai Y, Westernbroek R & Mitchell R (1995) Cannabinoids activate an inwardly rectifying potassium conductance and inhibit Q type calcium currents in AtT20 cells transfected with rat brain cannabinoid receptor. J Neurosci 15, 6552–6561. 6 Guo J & Ikeda SR (2004) Endocannabinoids modulate N-type calcium channels and G protein coupled Precoupled form of human CB1 and G protein trimer C. R. Chillakuri et al. 6114 FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS inwardly rectifying potassium channels via CB1 cannab- inoid receptors heterologously expressed in mammalian neurons. Mol Pharmacol 65, 665–674. 7 Nobles M, Benians A & Tinker A (2005) Heterotrimeric G proteins precouple with G protein-coupled receptors in living cells. Proc Natl Acad Sci USA 102, 18706– 18711. 8 de Light RA, Kourounakis AP & Ijzerman AP (2000) Inverse agonism at G protein-coupled receptors: (patho)physiological relevance and implications for drug discovery. Br J Pharmacol 130, 1–12. 9 Houston DB & Howlett AC (1992) Solubilization of the cannabinoid receptor from rat brain and its functional interaction with guanine nucleotide-binding protein. Mol Pharmacol 43, 17–22. 10 Rinaldi-Carmona M, Barth F, Heaulme M, Shire D, Calandra B, Congy C, Martinez S, Maruani J, Neliat G & Caput D (1994) SR141716A, a potent and selective antagonist of the brain cannabinoid receptor. FEBS Lett 350, 240–244. 11 Vasquez C & Lewis DL (1999) The CB1 cannabinoid receptor can sequester G proteins, making them unavailable to couple to other receptors. J Neurosci 19, 9271–9280. 12 Mukhopadhyay S, Mcintosh HH, Houston DB & Howlett AC (1999) The CB1 cannabinoid receptor juxta- membrane C-terminal peptide confers activation to spe- cific G proteins in brain. Mol Pharmacol 57, 162–170. 13 Nie J & Lewis DL (2001) Structural domains of the CB1 cannabinoid receptor that contribute to constitu- tive activity and G-protein sequestration. J Neurosci 21, 8758–8764. 14 Mukhopadhyay S & Howlett AC (2005) Chemically dis- tinct ligands promote differential CB1 cannabinoid receptor–Gi protein interactions. Mol Pharmacol 67, 2016–2024. 15 Wager-Miller J, Westenbroek R & Mackie K (2002) Dimerization of G protein coupled receptors: CB1 can- nabinoid receptors as an example. Chem Phys Lipids 121, 83–89. 16 Xu W, Filppula SA, Mercier R, Yaddanapudi S, Pavlopoulos S, Cai J, Pierce WM & Makriyannis A (2005) Purification and mass spectroscopic analysis of human CB1 cannabinoid receptor functionally expressed using the baculovirus system. J Pept Res 66, 138–150. 17 Howlett AC (2004) Efficacy in CB1 receptor-mediated signal transduction. Nature 142, 1209–1218. 18 Bouaboula M, Perrachon S, Milligan L, Canat X, Rinaldi-Carmona M, Portier M, Barth F, Calandra B, Pecceu F, Lupker J, et al. (1997) A selective inverse agonist for central cannabinoid receptor inhibits mitogen-activated protein kinase activation stimulation by insulin or insulin-like growth factor 1. Evidence for a new model of receptor ⁄ ligand interactions. J Biol Chem 272, 22330–22339. 19 Glass M & Northup JK (1999) Agonist selective regula- tion of G proteins by cannabinoid CB1 and CB2 recep- tors. Mol Pharmacol 56, 1362–1369. 20 Weiss HM & Grisshammer R (2002) Purification and characterization of the human adenosine A(2a) receptor functionally expressed in Escherichia coli. Eur J Biochem 269, 82–92. 21 Yeliseev AA, Wong KK, Soubias O & Gawrisch K (2005) Expression of human peripheral cannabinoid receptor for structural studies. Prot Sci 14, 2638– 2653. 22 Frances B, Puget A, Moisand C & Meunier JC (1990) Apparent precoupling of j- but not l-opioid receptors with a G protein in the absence of agonist. Eur J Phar- macol 189, 1–9. 23 Summers MD & Smith GE (1987) A manual of methods for baculovirus vectors and insect cell culture procedures. Texas Agric Exp Station Bull 1555. 24 Gru ¨ newald S, Reila ¨ nder H & Michel H (1996) In vivo reconstitution of dopamine D2S receptor-mediated G protein activation in baculovirus infected cells: preferred coupling to Gil versus Gi2. Biochemistry 35, 15162– 15173. C. R. Chillakuri et al. Precoupled form of human CB1 and G protein trimer FEBS Journal 274 (2007) 6106–6115 ª 2007 The Authors Journal compilation ª 2007 FEBS 6115 . full-length CB1 gene, with EcoRI and NotI restriction sites, was 5¢-GAA T GC GGC CGC TCA CTT TTC GAA TTG AGG GTG CGA CCA GAA TTC AGC CTC GGC AGA CGT GTC TGT GGA-3¢,. the distal C-terminal tail ( 418 –472) is not involved in CB1( 417 ) G +G Anti-Flag M2 IgG Anti-His tag IgG Anti-His tag IgG Anti -G i1 /G i2 IgG Receptor Receptor G G i1 Fig.