- Email: [email protected]
Rat brain cannabinoid receptors are N-linked glycosylated proteins
Life Sciences, Vol. 36, Noa 23/2A, pp. 19831989,199S Copyright 0 1995 Ekvier Science Ltd Printed in the USA. All tights reserved oo?A-3205/9s $950 t .al
Pergamon 0024-3205(95)001794
RAT BRAIN CANNABINOID RECEPTORS ARE N-LINKED GLYCOSYLATED PROTEINS Chao Song and Allyn C. Howlett
Department of Pharmacological and Physiological Science, Saint Louis University School of Medicine, 1402 South Grand Blvd., St. Louis, MO 63104 USA
Summary To study the N-linked glycosylation properties of the CBl receptor, rat brain membranes were treated with exo- and endoglycosidases. For visualizing CBl receptors, an antipeptide antibody was raised against the N-terminal 14 amino acids and used to specifically detect the protein by Western blotting. We found that the apparent molecular weight of mature CBl receptors was 64 kDa. Treatment of membranes with endoglycosidase F shifted the 64 kDa band to the 59 kDa and 53 The latter is consistent with the calculated molecular weight of kDa bands. deglycosylated CBl receptors. Treatment of membranes with endoglycosidase H and a-mannosidase partially shifted the 64 kDa band to 53 kDa band, indicating a portion of the oligosaccharides was of the high marmose type. These data confirmed that the CBl receptors in brain are N-linked glycoproteins with heterogeneous carbohydrate composition. Among three potential N-linked glycosylation sites on the N-terminus of the CBl receptor, only two sites are actually glycosylated. Key Words: endoglycosidases, SDS-PAGE, CBl cannabinoid receptor antibodies, G protein coupled receptors Rat brain CBl cannabinoid receptors may be glycoproteins since the consensus sequence (N-X-S/T) for three potential sites of N-linked glycosylation exists on the extracellular N-terminus of the receptors (1). However, whether these sites are actually glycosylated and the nature of the carbohydrate moieties on CBl receptors have not been studied. To study N-linked glycosylation on the CBl receptor, we treated rat brain membranes with exo- and endoglycosidases. Such enzyme treatments cut specific oligosaccharide moieties from the protein and result in a decrease in the molecular weight reflected by a band shift on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting as has been done for other G protein coupled receptors (2, 3, 4). We have developed an antipeptide antiserum against the N-terminus of the CBl receptor to detect the band shifts of CBl receptors in the immunoblot caused by the deglycosylation. We now report that the CBl receptors are predominantly N-linked glycoproteins with carbohydrate moieties composed of high mannose, hybrid and bi-, tri- and tetra- antennary chain complex carbohydrates. Materials
and Methods
Development of an antipeptide antibody_ A peptide consisting of the first 14 N-terminal amino acids of the CBl receptor was synthesized and purified by high-performance liquid chromatography. Polyclonal antibodies were produced in rabbits and for some experiments,
1984
Glycmylated Camabiioid Receptors
Vol. 56, No.s 23/24,1995
the fraction of IgG in the polyclonal antiserum was purified using affinity chromatography immobilized peptide as described elsewhere (Song and Howlett, in preparation).
on
Rat brain membrane prepamtion. Rat brain membranes were prepared as previously described (5) except that a protease inhibitor cocktail (15 pg/ml benzamidine, 5 pg/ml leupeptin, 50 pg/ml soy bean trypsin inhibitor, 1mM phenylmethanesulfonylfluoride and 0.7 pg/ml pepstatin) was included in the preparation. Treatment with exo- and endoglycosidases. Membrane samples were treated with exo- or endoglycosidases (Boehringer Mannheim Biochemical) in the presence of the protease inhibitor cocktail at the indicated doses and times. Controls were incubated under the same conditions without the enzymes. For studies with endoglycosidase F (N-glycanase F-free), samples were treated with 0.5 U/ml to 5 U/ml endoglycosidase F in phosphate buffered saline at pH 7 at the indicated temperature and time (4). For studies with endoglycosidase H, membrane protein was incubated with 100 mu/ml in 100 mM sodium acetate buffer, pH 5.5 at 37 “C for 2, 6 and 18 hours (6, 7, 8). For treatment with a-mannosidase, an aliquot of membranes was incubated with cy-mannosidase (5 U/ml) in 50 mM sodium citrate, pH 4.5 at 25 “C for 24 h. Immunab~ots. Rat brain membrane proteins were separated on 10% polyacrylamide gels by SDS-PAGE (9). Proteins on the gels were electrophoretically transferred to nitrocellulose membranes (Fisher) overnight at 4 ‘C using a CAPS, (3-[cyclohexylaminol-1-propanesulfonic acid), transfer buffer containing 0.01% SDS. The nitrocellulose membranes were washed with a buffer containing 20 mM Tris-HCl, pH 7.5, 137 mM NaCl and 0.1% Tween 20 (TBST). Then, the membranes were incubated with blocking buffer containing 1% BSA in TBST, at ambient temperature for 1 h. Membranes were washed with TBST buffer twice and then incubated with the antiserum or affinity purified IgG (diluted in blocking buffer) for 1 h at ambient temperature. After incubation, the membranes were washed with TBST buffer three times and subsequently incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (l/10,000 dilution in blocking buffer) at ambient temperature for 1 h. The membranes were extensively washed with TBST buffer. The peroxidase activity was detected using a Band densities were quantitated using a BioRad luminescence reagent kit (Amersham). Standard molecular weight markers densitometer and the 1-D Analyst computer program. were biotinylated proteins (Bio-Rad). Results Our immunoblot assay identified a 64 kDa band, a putative glycosylated CBl receptor. This band represents a mature glycosylated CBl receptor since the specificity of the antibody has been confirmed by different methods (Song and Howlett, manuscript in preparation). Both 64 kDa and 53 kDa bands, corresponding to supposedly glycosylated and non-glycosylated cannabinoid receptors, respectively, can be detected in control samples by affinity purified antibody as shown in Fig. 1. with arrows in lane 2 in panel A and B. Another faint band of about 59 kDa was observed, indicating a population of partially glycosylated cannabinoid receptors. Densitometry scanning of these three bands (lane 2 of panel B) showed that optical densities (in relative units) were 6.93 (66.1%), 1.75 (16.7%) and 1.80 (17.2%) for the 64 kDa, 59 kDa and 53 kDa bands, respectively.
Vol. 56, No.s 23/24,1995
Glycosylated Cannabiioid Receptors
1985
Fig. 1. CBl cannabinoid receptors are glycoproteins as determined by endoglycosidase Fsensitivity. Rat brain membranes were treated with endoglycosidase F (5 U/ml. in PBS buffer at 37 ‘C for 6 and 18 hours or with pro tease inhibitors alone (control), andprocessedb y SDS-PAGE and Western blotting. The specific immunoreactivity was detected with affinity purified IgG (3.5 pg/ml). Treated (lane 1) and control (lane 2) membranes are shown after 6 hours (panel A) and 18 hours (panel Bl of incubation. Upper and lower arrows indicate 64 kDa and 53 kDa bands, respectively. Apparent molecular weight markers are indicated at 40 kDa and 58 kDa. This figure represents one of three separate experiments.
Endoglycosidase F cleaves the carbohydrate chain between the di-N-acetylglucosamine core linkage of N-linked glycans. Glycoproteins possessing biantennary complex, hybrid or high mannose chains are susceptible to endoglycosidase F hydrolysis (2, 7). After rat brain membranes were incubated with endoglycosidase F for 18 hours, the 64 kDa band was nearly completely shifted to lower molecular weight positions (Fig. 1, lane 1 of panel B). Densitometry scanning demonstrated that optical densities were 0.14 (1.8%), 2.3 (28.9 %) and 5.52 (69.3 %) for the 64 kDa, 59 kDa and 53 kDa bands, respectively. The sensitivity of oligosaccharides of the CBl receptor to endoglycosidase F indicated that these receptors were glycoproteins and the oligosaccharide moieties of the receptors consisted of high mannose, hybrid and biantennary complex types. Whereas the 64 kDa band was completely shifted to the 59 kDa or 53 kDa positions (lane 1 of panel B) after treated with endoglycosidase F (5 U/ml) for 18 h, the band was not shifted for endoglycosidase F treatment for 6 h (lane 1 of panel A), indicating that deglycosylation was time-dependent. The effect of endoglycosidase F was also dose-dependent since lower doses (0.5 U/ml) for an 18 h incubation did not completely shift the 64 kDa band (data not shown). Endoglycosidase H trims the carbohydrate chain between the di-N-acetylglucosamine core linkage of N-linked glycans having high mannose or hybrid carbohydrate chains (7, 8). To determine whether the oligosaccharide of CBl receptors is also sensitive to endoglycosidase H, the membrane preparation was treated with 100 mu/ml at 37 “C for 2, 6 and 18 h, respectively. The densities of 64 kDa, 58 kDa and 53 kDa bands in the control sample were 1.87 (42.4%), 1.0 (22.7%) and 1.54 (34.9 %), respectively (Fig. 2, lane 1). Endoglycosidase
1986
Glycosylated
Camabiioid
Receptors
Vol. 56, No.s 23/24, 1995
H treatment for 18 h partially shifted the 64 kDa band to the 53 kDa relatively dense band shown in lane 2, suggesting that at least a portion of the oligosaccharide of the receptors is high mannose. The optical density for the 64 kDa, 58 kLIa and 53 kDa bands was 1.25 (36.4%), 0.15 (4.4%) and 2.03 (59.2%), respectively (lane 2). The unshifted portion of 64 kDa band may represent biantennary complex type oligosaccharides since the existence of this type of carbohydrate 1
2
Fig. 2 Endoglycosidase H treatment of rat brain membranes. The membranes were incubated with protease inhibitors alone (lane l! or with 100 mtJ/ml endoglycosidase H (lane 21 at 37 ‘C for 18 h, and processed as described in the text. Rabbit antiserum at a 7:500 dilution was used in the immunoblo t. This figure presents data from one representative experiment from three separate experiments. 1
2
Fig. 3 a-Mannosidase treatment of ra r brain membranes. a-Mannosidase treatment (5 U/ml) was carried out at 25 ‘-‘C for 24 h and membranes were processed as described. The immunoblot was detected by the antiserum at a 1:400 dilution. Controland enzyme-treated samples are shown in lanes I and 2, respectively. The data are one representative from four individual experiments.
Vol. 56, No.s U/24,1995
Glycosylated Camabiioid
Receptors
1987
moiety was sensitive to endoglycosidase F but not to endoglycosidase H. Endoglycosidase H treatment was also time-dependent because incubation of the membranes with endoglycosidase H for 2 and 6 h did not shift the 64 kDa band (data not shown). The membrane preparation was treated with cw-mannosidase, which trims terminal mannose residues from high mannose and hybrid carbohydrate chains of the glycoprotein (10). The immunoblot showed that both 64 kDa and 53 kDa bands were detected with the antiserum in the control sample (Fig. 3, lane 1). cy-Mannosidase treatment shifted the 64 kDa band to a smear at lower molecular weight positions between 64 kDa and 53 kDa (lane 2). This sensitivity indicated that carbohydrate moieties of the receptor contained high mannose type or hybrid carbohydrate moieties with terminal mannose residues. The existence of unshifted proteins suggest that the receptors may contain complex chains without terminal mannose residues, as seen in fl-adrenergic receptors (2, 4).
Our immunoblot assay has demonstrated that immunoreactivity to the N-terminaus of the CBl receptor can appear at 64 kDa, 59 kDa, and 53 kDa bands. The difference of apparent molecular weight is about 5 or 6 kDa between bands. This may suggest that only two of three potential N-glycosylation sites are actually glycosylated. Stiles et al. (11) reported a molecular weight shift of mammalian fl-adrenergic receptors from 62 kDa to 49 kDa after endoglycosidase F treatment. They interpreted this to mean that the ,&adrenergic receptors possess two N-linked oligosaccharides. Although we know that there may be two of three Nlinked glycosylation sites occupied with oligosaccharides, we still have no knowledge about which two of the three these are. A proteolytic study of the purified receptors or a sitedirected mutation study of the CBl receptor could elucidate this question. Rands et al. (12) used site-directed mutations to determine the location of two N-liked glycosylation sites of the @-adrenergic receptors. Our research has demonstrated that the majority of the CBl receptors are mature or glycosylated receptors. The densitometry analysis of the bands showed that glycosylated CBl receptors are found in about 65 to 80% abundance, whereas nonglycosylated CBl receptors at 53 kDa account for only about 20 to 35 % . The immunoreactivity that appears at 53 kDa may represent newly synthesized receptor that may have escaped cotranslational glycosylation. Alternatively, it may represent a fraction of the receptors that have been trimmed of their high mannose carbohydrates in the golgi but not further modified with complex carbohydrate chains. Pettit et al. (13) infected Sf9 insect cells with a CBl receptor recombinant baculovirus, isolated the mRNA and subjected it to in vitro translation. A single band was observed to migrate at 55 kDa on SDS-PAGE as determined by Western immunoblotting with antisera to a fusion protein possessing the CBl receptor N-terminal amino acids 26 to 108. However, proteins harvested from CBl receptor recombinant baculovirus-infected Sfp cells resulted in multiple immunoreactive bands appearing in the 32 kDa to 79 kDa relative molecular weight range (13). These investigators hypothesized that the 55 kDa band represented the unglycosylated protein and that bands at 73 kDa and 79 kDa represented glycosylated species. It is not likely that the carbohydrate processing in these infected insect cells would be identical to that occuring by normal synthetic steps in rat brain cells. This would account for the differrences in relative molecular weight between the two glycosylated forms observed by Pettit et al. (13) compared with those reported here. It should be noted that an amino-terminal splice variant of the CBl receptor has been predicted from cDNA lacking a 167-base pair sequence within
1988
Glycosylated
Cannabiioid
Receptors
Vol. 56, No.s 23124, 1995
the region (14). However, the amino acid sequence of the N-terminus of the splice variant fails to possess any homology to the first 14-amino acid sequence of the N-terminus of the CBl receptor that we used to produce antisera. Therefore, the splice variant CBl receptor would not be detected in the present investigation. Our research suggests that CBl receptors in rat brain have heterogeneous oligosaccharide moieties. After endoglycosidase H treatment for 18 h, about 40% of the glycosylated CBl receptors were trimmed. This might represent the fraction of receptors possessing oligosaccharides of the high marmose and/or hybrid forms. Because N-linked glycosylation of the high mannose form occurs cotranslationally, this fraction would include newly-synthesized receptors present in the endoplasmic reticulum. It could also include mature receptors if the high mannose oligosaccharide chains are not processed as a part of the maturation process in the golgi. The latter would include the hybrids of high mannose and complex chains. Endoglycosidase F treatment completely shifted the 64 kDa band to either 59 kDa or 53 kDa bands. This would be consistent with the cleavage of complex biantennary chains and hybrids in addition to high mannose chains. About 30% remained in the 59 kDa band. If this represents one of two N-linked glycosylation sites, then one would assume that one site Some endoglycosidase F preparations are mixtures of remains refractory to this cleavage. endoglycosidase F and N-glycanase F. Because the endoglycosidase F used in these experiments was free of N-glycanase F, which cleaves the larger tri- or tetraantennary chain complex oligosaccharides, one glycosylation site may possess tri- or tetra-antennary moieties. It has been shown that many G protein-coupled receptors and receptor channels are Nlinked glycoproteins with heterogeneous oligosaccharides, including D2 dopamine receptors (15), &adrenergic receptors (4), muscarinic acetylcholine receptors (16), AZadenosine receptors (2), and others. In some G protein-coupled receptors, the glycosylation is important for functions related to the expression and processing of the receptor, binding of the ligand or second messenger coupling of these receptors (4, 8, 10, 14, 17). We previously reported that prevention of glycosylation of CBl receptors in N18TG2 cells with tunicamycin did not appear to affect the cannabinoid-induced inhibition of cyclic AMP accumulation in cultured neuroblastoma cells (18). However, it was not known whether the receptors that would have been synthesized prior to the tunicamycin block might be sufficient to produce the full response. Studies to determine directly if glycosylation plays a role in ligand binding and signal transduction are currently underway. Acknowledeements We thank Rusoun Belue for preparation of the rat brain membranes and Dr. Qianjing Liu for help with the immunoblots. This work was supported by NIDA grants DA-03690 and DA-06312. References 1. 2. 3.
L.A. MATSUDA, S.J. LOLAIT, M.J. BROWNSTEIN, A.C. YOUNG, and T.I. BONNER, Nature, 356 561-564 (1990). W.W. BARRINGTON, K.A. JACOBSON, and G.L. STILES, Mol. Pharmacol. 38 177-183 (1990). K.R. JARVIE, H.B. NIZNIK, and P. SEEMAN, Mol. Pharmacol. 4 91-97 (1988).
Vol. 56, No.s 23/B, 1995
4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
15. 16. 17. 18.
Glycosylated Cannabiioid Receptors
1989
G.L. STILES, J.L. BENOVIC, M.G. CARON, and R.J. LEFKOWITZ, J.Biol. Chem. 259 8655-8663 (1984). W.A. DEVANE, F.A. DYSARZ III, M.R. JOHNSON, L.S. MELVIN, and A.C. HOWLETI’, Mol. Pharmacol. 34 605-613 (1988). A.L. TARENTINO, Methods of Enzymology 50 50-74 (1978). A.L. TARENTINO, C.M. GOMEZ, and T.H. PLUMMER JR, B&hem. 24 46654671 (1985). S. RENS-DOMIANO, and T. REISINE, J. Biological Chem. 266 20094-20102 (1991). U.K. LAEMMLI, Nature (London) 227, 680-685 (1970). D.G. SAWUTZ, S.M. LANIER, C.D. WARREN and R.M. GRAHAM, Mol. Pharmacol. 32 566-571 (1987). G.L. STILES, Arch. B&hem. Biophys. 237 65-71 (1985). E. RANDS, M.R. CANDELORE, A.H. CHEUNG, W.S. HILL, C.D. STRADER, and R.A.F. DIXON, J. Biol. Chem. 265 10759-10764 (1990). D.A. PETTIT, V.M. SHOWALTER, M.E. ABOOD AND G.A. CABRAL, B&hem. Pharmacol. 48 1231-1243 (1994). D. SHIRE, C. CARILLON, M. KAGHAD, B. CALANDRA, M. RINALDICARMONAS, G. LE FUR, D. CAPUT AND P. FERRARA, J. Biol. Chem. 270 3726-373 1 (1995). M.N. LEONARD, R.A. WILLIAMSON and P.G. STRANGE, B&hem. J. 255 877-883 (1988). G.S. HERRON and M.I. SCHIMERLIK, J. Neurochem. 4-l 1414-1420 (1983). S.T. GEORGE, A.E. RUOHO and C.C. MALBON, J. Biological Chem. 261 1655916564 (1986). A.C. HOWLETT, T.M. CHAMPION-DOROW, L.L. MCMAHON and T.M. WESTLAKE, Pharmacol. B&hem. & Behavior 40 565-569 (1991).
Pergamon 0024-3205(95)001794
RAT BRAIN CANNABINOID RECEPTORS ARE N-LINKED GLYCOSYLATED PROTEINS Chao Song and Allyn C. Howlett
Department of Pharmacological and Physiological Science, Saint Louis University School of Medicine, 1402 South Grand Blvd., St. Louis, MO 63104 USA
Summary To study the N-linked glycosylation properties of the CBl receptor, rat brain membranes were treated with exo- and endoglycosidases. For visualizing CBl receptors, an antipeptide antibody was raised against the N-terminal 14 amino acids and used to specifically detect the protein by Western blotting. We found that the apparent molecular weight of mature CBl receptors was 64 kDa. Treatment of membranes with endoglycosidase F shifted the 64 kDa band to the 59 kDa and 53 The latter is consistent with the calculated molecular weight of kDa bands. deglycosylated CBl receptors. Treatment of membranes with endoglycosidase H and a-mannosidase partially shifted the 64 kDa band to 53 kDa band, indicating a portion of the oligosaccharides was of the high marmose type. These data confirmed that the CBl receptors in brain are N-linked glycoproteins with heterogeneous carbohydrate composition. Among three potential N-linked glycosylation sites on the N-terminus of the CBl receptor, only two sites are actually glycosylated. Key Words: endoglycosidases, SDS-PAGE, CBl cannabinoid receptor antibodies, G protein coupled receptors Rat brain CBl cannabinoid receptors may be glycoproteins since the consensus sequence (N-X-S/T) for three potential sites of N-linked glycosylation exists on the extracellular N-terminus of the receptors (1). However, whether these sites are actually glycosylated and the nature of the carbohydrate moieties on CBl receptors have not been studied. To study N-linked glycosylation on the CBl receptor, we treated rat brain membranes with exo- and endoglycosidases. Such enzyme treatments cut specific oligosaccharide moieties from the protein and result in a decrease in the molecular weight reflected by a band shift on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting as has been done for other G protein coupled receptors (2, 3, 4). We have developed an antipeptide antiserum against the N-terminus of the CBl receptor to detect the band shifts of CBl receptors in the immunoblot caused by the deglycosylation. We now report that the CBl receptors are predominantly N-linked glycoproteins with carbohydrate moieties composed of high mannose, hybrid and bi-, tri- and tetra- antennary chain complex carbohydrates. Materials
and Methods
Development of an antipeptide antibody_ A peptide consisting of the first 14 N-terminal amino acids of the CBl receptor was synthesized and purified by high-performance liquid chromatography. Polyclonal antibodies were produced in rabbits and for some experiments,
1984
Glycmylated Camabiioid Receptors
Vol. 56, No.s 23/24,1995
the fraction of IgG in the polyclonal antiserum was purified using affinity chromatography immobilized peptide as described elsewhere (Song and Howlett, in preparation).
on
Rat brain membrane prepamtion. Rat brain membranes were prepared as previously described (5) except that a protease inhibitor cocktail (15 pg/ml benzamidine, 5 pg/ml leupeptin, 50 pg/ml soy bean trypsin inhibitor, 1mM phenylmethanesulfonylfluoride and 0.7 pg/ml pepstatin) was included in the preparation. Treatment with exo- and endoglycosidases. Membrane samples were treated with exo- or endoglycosidases (Boehringer Mannheim Biochemical) in the presence of the protease inhibitor cocktail at the indicated doses and times. Controls were incubated under the same conditions without the enzymes. For studies with endoglycosidase F (N-glycanase F-free), samples were treated with 0.5 U/ml to 5 U/ml endoglycosidase F in phosphate buffered saline at pH 7 at the indicated temperature and time (4). For studies with endoglycosidase H, membrane protein was incubated with 100 mu/ml in 100 mM sodium acetate buffer, pH 5.5 at 37 “C for 2, 6 and 18 hours (6, 7, 8). For treatment with a-mannosidase, an aliquot of membranes was incubated with cy-mannosidase (5 U/ml) in 50 mM sodium citrate, pH 4.5 at 25 “C for 24 h. Immunab~ots. Rat brain membrane proteins were separated on 10% polyacrylamide gels by SDS-PAGE (9). Proteins on the gels were electrophoretically transferred to nitrocellulose membranes (Fisher) overnight at 4 ‘C using a CAPS, (3-[cyclohexylaminol-1-propanesulfonic acid), transfer buffer containing 0.01% SDS. The nitrocellulose membranes were washed with a buffer containing 20 mM Tris-HCl, pH 7.5, 137 mM NaCl and 0.1% Tween 20 (TBST). Then, the membranes were incubated with blocking buffer containing 1% BSA in TBST, at ambient temperature for 1 h. Membranes were washed with TBST buffer twice and then incubated with the antiserum or affinity purified IgG (diluted in blocking buffer) for 1 h at ambient temperature. After incubation, the membranes were washed with TBST buffer three times and subsequently incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (l/10,000 dilution in blocking buffer) at ambient temperature for 1 h. The membranes were extensively washed with TBST buffer. The peroxidase activity was detected using a Band densities were quantitated using a BioRad luminescence reagent kit (Amersham). Standard molecular weight markers densitometer and the 1-D Analyst computer program. were biotinylated proteins (Bio-Rad). Results Our immunoblot assay identified a 64 kDa band, a putative glycosylated CBl receptor. This band represents a mature glycosylated CBl receptor since the specificity of the antibody has been confirmed by different methods (Song and Howlett, manuscript in preparation). Both 64 kDa and 53 kDa bands, corresponding to supposedly glycosylated and non-glycosylated cannabinoid receptors, respectively, can be detected in control samples by affinity purified antibody as shown in Fig. 1. with arrows in lane 2 in panel A and B. Another faint band of about 59 kDa was observed, indicating a population of partially glycosylated cannabinoid receptors. Densitometry scanning of these three bands (lane 2 of panel B) showed that optical densities (in relative units) were 6.93 (66.1%), 1.75 (16.7%) and 1.80 (17.2%) for the 64 kDa, 59 kDa and 53 kDa bands, respectively.
Vol. 56, No.s 23/24,1995
Glycosylated Cannabiioid Receptors
1985
Fig. 1. CBl cannabinoid receptors are glycoproteins as determined by endoglycosidase Fsensitivity. Rat brain membranes were treated with endoglycosidase F (5 U/ml. in PBS buffer at 37 ‘C for 6 and 18 hours or with pro tease inhibitors alone (control), andprocessedb y SDS-PAGE and Western blotting. The specific immunoreactivity was detected with affinity purified IgG (3.5 pg/ml). Treated (lane 1) and control (lane 2) membranes are shown after 6 hours (panel A) and 18 hours (panel Bl of incubation. Upper and lower arrows indicate 64 kDa and 53 kDa bands, respectively. Apparent molecular weight markers are indicated at 40 kDa and 58 kDa. This figure represents one of three separate experiments.
Endoglycosidase F cleaves the carbohydrate chain between the di-N-acetylglucosamine core linkage of N-linked glycans. Glycoproteins possessing biantennary complex, hybrid or high mannose chains are susceptible to endoglycosidase F hydrolysis (2, 7). After rat brain membranes were incubated with endoglycosidase F for 18 hours, the 64 kDa band was nearly completely shifted to lower molecular weight positions (Fig. 1, lane 1 of panel B). Densitometry scanning demonstrated that optical densities were 0.14 (1.8%), 2.3 (28.9 %) and 5.52 (69.3 %) for the 64 kDa, 59 kDa and 53 kDa bands, respectively. The sensitivity of oligosaccharides of the CBl receptor to endoglycosidase F indicated that these receptors were glycoproteins and the oligosaccharide moieties of the receptors consisted of high mannose, hybrid and biantennary complex types. Whereas the 64 kDa band was completely shifted to the 59 kDa or 53 kDa positions (lane 1 of panel B) after treated with endoglycosidase F (5 U/ml) for 18 h, the band was not shifted for endoglycosidase F treatment for 6 h (lane 1 of panel A), indicating that deglycosylation was time-dependent. The effect of endoglycosidase F was also dose-dependent since lower doses (0.5 U/ml) for an 18 h incubation did not completely shift the 64 kDa band (data not shown). Endoglycosidase H trims the carbohydrate chain between the di-N-acetylglucosamine core linkage of N-linked glycans having high mannose or hybrid carbohydrate chains (7, 8). To determine whether the oligosaccharide of CBl receptors is also sensitive to endoglycosidase H, the membrane preparation was treated with 100 mu/ml at 37 “C for 2, 6 and 18 h, respectively. The densities of 64 kDa, 58 kDa and 53 kDa bands in the control sample were 1.87 (42.4%), 1.0 (22.7%) and 1.54 (34.9 %), respectively (Fig. 2, lane 1). Endoglycosidase
1986
Glycosylated
Camabiioid
Receptors
Vol. 56, No.s 23/24, 1995
H treatment for 18 h partially shifted the 64 kDa band to the 53 kDa relatively dense band shown in lane 2, suggesting that at least a portion of the oligosaccharide of the receptors is high mannose. The optical density for the 64 kDa, 58 kLIa and 53 kDa bands was 1.25 (36.4%), 0.15 (4.4%) and 2.03 (59.2%), respectively (lane 2). The unshifted portion of 64 kDa band may represent biantennary complex type oligosaccharides since the existence of this type of carbohydrate 1
2
Fig. 2 Endoglycosidase H treatment of rat brain membranes. The membranes were incubated with protease inhibitors alone (lane l! or with 100 mtJ/ml endoglycosidase H (lane 21 at 37 ‘C for 18 h, and processed as described in the text. Rabbit antiserum at a 7:500 dilution was used in the immunoblo t. This figure presents data from one representative experiment from three separate experiments. 1
2
Fig. 3 a-Mannosidase treatment of ra r brain membranes. a-Mannosidase treatment (5 U/ml) was carried out at 25 ‘-‘C for 24 h and membranes were processed as described. The immunoblot was detected by the antiserum at a 1:400 dilution. Controland enzyme-treated samples are shown in lanes I and 2, respectively. The data are one representative from four individual experiments.
Vol. 56, No.s U/24,1995
Glycosylated Camabiioid
Receptors
1987
moiety was sensitive to endoglycosidase F but not to endoglycosidase H. Endoglycosidase H treatment was also time-dependent because incubation of the membranes with endoglycosidase H for 2 and 6 h did not shift the 64 kDa band (data not shown). The membrane preparation was treated with cw-mannosidase, which trims terminal mannose residues from high mannose and hybrid carbohydrate chains of the glycoprotein (10). The immunoblot showed that both 64 kDa and 53 kDa bands were detected with the antiserum in the control sample (Fig. 3, lane 1). cy-Mannosidase treatment shifted the 64 kDa band to a smear at lower molecular weight positions between 64 kDa and 53 kDa (lane 2). This sensitivity indicated that carbohydrate moieties of the receptor contained high mannose type or hybrid carbohydrate moieties with terminal mannose residues. The existence of unshifted proteins suggest that the receptors may contain complex chains without terminal mannose residues, as seen in fl-adrenergic receptors (2, 4).
Our immunoblot assay has demonstrated that immunoreactivity to the N-terminaus of the CBl receptor can appear at 64 kDa, 59 kDa, and 53 kDa bands. The difference of apparent molecular weight is about 5 or 6 kDa between bands. This may suggest that only two of three potential N-glycosylation sites are actually glycosylated. Stiles et al. (11) reported a molecular weight shift of mammalian fl-adrenergic receptors from 62 kDa to 49 kDa after endoglycosidase F treatment. They interpreted this to mean that the ,&adrenergic receptors possess two N-linked oligosaccharides. Although we know that there may be two of three Nlinked glycosylation sites occupied with oligosaccharides, we still have no knowledge about which two of the three these are. A proteolytic study of the purified receptors or a sitedirected mutation study of the CBl receptor could elucidate this question. Rands et al. (12) used site-directed mutations to determine the location of two N-liked glycosylation sites of the @-adrenergic receptors. Our research has demonstrated that the majority of the CBl receptors are mature or glycosylated receptors. The densitometry analysis of the bands showed that glycosylated CBl receptors are found in about 65 to 80% abundance, whereas nonglycosylated CBl receptors at 53 kDa account for only about 20 to 35 % . The immunoreactivity that appears at 53 kDa may represent newly synthesized receptor that may have escaped cotranslational glycosylation. Alternatively, it may represent a fraction of the receptors that have been trimmed of their high mannose carbohydrates in the golgi but not further modified with complex carbohydrate chains. Pettit et al. (13) infected Sf9 insect cells with a CBl receptor recombinant baculovirus, isolated the mRNA and subjected it to in vitro translation. A single band was observed to migrate at 55 kDa on SDS-PAGE as determined by Western immunoblotting with antisera to a fusion protein possessing the CBl receptor N-terminal amino acids 26 to 108. However, proteins harvested from CBl receptor recombinant baculovirus-infected Sfp cells resulted in multiple immunoreactive bands appearing in the 32 kDa to 79 kDa relative molecular weight range (13). These investigators hypothesized that the 55 kDa band represented the unglycosylated protein and that bands at 73 kDa and 79 kDa represented glycosylated species. It is not likely that the carbohydrate processing in these infected insect cells would be identical to that occuring by normal synthetic steps in rat brain cells. This would account for the differrences in relative molecular weight between the two glycosylated forms observed by Pettit et al. (13) compared with those reported here. It should be noted that an amino-terminal splice variant of the CBl receptor has been predicted from cDNA lacking a 167-base pair sequence within
1988
Glycosylated
Cannabiioid
Receptors
Vol. 56, No.s 23124, 1995
the region (14). However, the amino acid sequence of the N-terminus of the splice variant fails to possess any homology to the first 14-amino acid sequence of the N-terminus of the CBl receptor that we used to produce antisera. Therefore, the splice variant CBl receptor would not be detected in the present investigation. Our research suggests that CBl receptors in rat brain have heterogeneous oligosaccharide moieties. After endoglycosidase H treatment for 18 h, about 40% of the glycosylated CBl receptors were trimmed. This might represent the fraction of receptors possessing oligosaccharides of the high marmose and/or hybrid forms. Because N-linked glycosylation of the high mannose form occurs cotranslationally, this fraction would include newly-synthesized receptors present in the endoplasmic reticulum. It could also include mature receptors if the high mannose oligosaccharide chains are not processed as a part of the maturation process in the golgi. The latter would include the hybrids of high mannose and complex chains. Endoglycosidase F treatment completely shifted the 64 kDa band to either 59 kDa or 53 kDa bands. This would be consistent with the cleavage of complex biantennary chains and hybrids in addition to high mannose chains. About 30% remained in the 59 kDa band. If this represents one of two N-linked glycosylation sites, then one would assume that one site Some endoglycosidase F preparations are mixtures of remains refractory to this cleavage. endoglycosidase F and N-glycanase F. Because the endoglycosidase F used in these experiments was free of N-glycanase F, which cleaves the larger tri- or tetraantennary chain complex oligosaccharides, one glycosylation site may possess tri- or tetra-antennary moieties. It has been shown that many G protein-coupled receptors and receptor channels are Nlinked glycoproteins with heterogeneous oligosaccharides, including D2 dopamine receptors (15), &adrenergic receptors (4), muscarinic acetylcholine receptors (16), AZadenosine receptors (2), and others. In some G protein-coupled receptors, the glycosylation is important for functions related to the expression and processing of the receptor, binding of the ligand or second messenger coupling of these receptors (4, 8, 10, 14, 17). We previously reported that prevention of glycosylation of CBl receptors in N18TG2 cells with tunicamycin did not appear to affect the cannabinoid-induced inhibition of cyclic AMP accumulation in cultured neuroblastoma cells (18). However, it was not known whether the receptors that would have been synthesized prior to the tunicamycin block might be sufficient to produce the full response. Studies to determine directly if glycosylation plays a role in ligand binding and signal transduction are currently underway. Acknowledeements We thank Rusoun Belue for preparation of the rat brain membranes and Dr. Qianjing Liu for help with the immunoblots. This work was supported by NIDA grants DA-03690 and DA-06312. References 1. 2. 3.
L.A. MATSUDA, S.J. LOLAIT, M.J. BROWNSTEIN, A.C. YOUNG, and T.I. BONNER, Nature, 356 561-564 (1990). W.W. BARRINGTON, K.A. JACOBSON, and G.L. STILES, Mol. Pharmacol. 38 177-183 (1990). K.R. JARVIE, H.B. NIZNIK, and P. SEEMAN, Mol. Pharmacol. 4 91-97 (1988).
Vol. 56, No.s 23/B, 1995
4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
15. 16. 17. 18.
Glycosylated Cannabiioid Receptors
1989
G.L. STILES, J.L. BENOVIC, M.G. CARON, and R.J. LEFKOWITZ, J.Biol. Chem. 259 8655-8663 (1984). W.A. DEVANE, F.A. DYSARZ III, M.R. JOHNSON, L.S. MELVIN, and A.C. HOWLETI’, Mol. Pharmacol. 34 605-613 (1988). A.L. TARENTINO, Methods of Enzymology 50 50-74 (1978). A.L. TARENTINO, C.M. GOMEZ, and T.H. PLUMMER JR, B&hem. 24 46654671 (1985). S. RENS-DOMIANO, and T. REISINE, J. Biological Chem. 266 20094-20102 (1991). U.K. LAEMMLI, Nature (London) 227, 680-685 (1970). D.G. SAWUTZ, S.M. LANIER, C.D. WARREN and R.M. GRAHAM, Mol. Pharmacol. 32 566-571 (1987). G.L. STILES, Arch. B&hem. Biophys. 237 65-71 (1985). E. RANDS, M.R. CANDELORE, A.H. CHEUNG, W.S. HILL, C.D. STRADER, and R.A.F. DIXON, J. Biol. Chem. 265 10759-10764 (1990). D.A. PETTIT, V.M. SHOWALTER, M.E. ABOOD AND G.A. CABRAL, B&hem. Pharmacol. 48 1231-1243 (1994). D. SHIRE, C. CARILLON, M. KAGHAD, B. CALANDRA, M. RINALDICARMONAS, G. LE FUR, D. CAPUT AND P. FERRARA, J. Biol. Chem. 270 3726-373 1 (1995). M.N. LEONARD, R.A. WILLIAMSON and P.G. STRANGE, B&hem. J. 255 877-883 (1988). G.S. HERRON and M.I. SCHIMERLIK, J. Neurochem. 4-l 1414-1420 (1983). S.T. GEORGE, A.E. RUOHO and C.C. MALBON, J. Biological Chem. 261 1655916564 (1986). A.C. HOWLETT, T.M. CHAMPION-DOROW, L.L. MCMAHON and T.M. WESTLAKE, Pharmacol. B&hem. & Behavior 40 565-569 (1991).