Introduction

Top Summary Introduction Materials and Methods Results Discussion References

The cannabinoid CB₁ receptor is a member of the seven-membrane-spanning receptor family (Matsuda et al., 1990) that mediates the actions of cannabinoids in the central nervous system (Pertwee, 1997). The CB₁ receptor activates heterotrimeric pertussis toxin (PTX)-sensitive G proteins (G_i/G_o) in many tissues, which results in a variety of actions within the cell. These actions include activation of an inwardly rectifying potassium channel (KIR; Mackie et al., 1995), inhibition of voltage-gated Ca²⁺ channels (Mackie et al., 1995), inhibition of cAMP production (Howlett, 1985; Howlett et al., 1986), and activation of the MAP kinase cascade (Bouaboula et al., 1995). The seven-membrane-spanning receptor family has numerous highly conserved amino acids thought to be responsible for universally critical actions of the receptor family (Savarese and Fraser, 1992). An aspartate residue in the second membrane-spanning helix (D164 in the rat CB₁ receptor) is among these conserved amino acids. Mutation of this residue in various receptors has a variety of effects, including decreasing agonist affinity (Strader et al., 1987; Chung et al., 1988; Ho et al., 1992; Chanda et al., 1993; Brodbeck et al., 1995; Parent et al., 1996; Chakrabarti et al., 1997), blocking coupling to KIR current (Surprenant et al., 1992), decreasing phosphoinositide hydrolysis (Brodbeck et al., 1995; Jagerschmidt et al., 1995; Sealfon et al., 1995), blocking inhibition of cAMP production (Chakrabarti et al., 1997), and eliminating allosteric receptor modulation by sodium ions (Horstman et al., 1990; Kong et al., 1993; Parent et al., 1996). This conserved aspartate in the second transmembrane region is thought to interact with an asparagine in the seventh transmembrane helix (Zhou et al., 1994), and many of the actions resulting from mutation of the aspartate are thought to result from disruption of this interaction. Consistent with this idea, a double point mutation that exchanges these conserved aspartate and asparagine residues in the 5-HT_2A receptor restores coupling of this receptor to phosphoinositide hydrolysis, which is lost by mutation of the aspartate residue (Sealfon et al., 1995).

To determine the role of this highly conserved aspartate residue in cannabinoid CB₁ receptors, we mutated this amino acid and tested the ability of the mutant CB₁ receptor to couple to several intracellular signaling cascades which normally couple to the wild-type CB₁ receptors in AtT20 cells. We also examined whether any disruptive effects of D164N were a consequence of loss of an interaction between D164 and N394 in the seventh transmembrane region. For this purpose, we compared the coupling of D164N to that of D164N/N394D. The results reveal that D164 has a role, independent of an interaction with N394, in receptor internalization and G protein-mediated modulation of KIR.

	Materials and Methods

Top Summary Introduction Materials and Methods Results Discussion References

Radioligand Binding and Adenylyl Cyclase Activity. For binding studies, AtT20 cells transfected with pcDNA3 (control), CB₁, or CB₁-D164N were grown to confluency in 500-cm² plates. The monolayers were washed twice with TEM (25 mM Tris-HCl, 6 mM MgCl₂, 1 mM EDTA, 10 µM phenylmethylsulfonyl fluoride, and 1 µg/ml leupeptin, pH 7.4) and homogenized in TEM (1 ml/plate). The homogenate was centrifuged at 800g for 10 min; the resulting pellet was homogenized in TEM and spun at 800g for 10 min. The supernatants were combined and centrifuged at 100,000g for 1 h; then, the pellets were homogenized in TEM at a final protein concentration of 1 to 5 mg/ml. Membranes (50 µg) were incubated in 20 mM HEPES and 1 mg/ml bovine serum albumin (BSA), pH 7.5 (final volume 200 µl) in 0.325 to 6 nM [³H]WIN 55,212-2 [{R-(+)-(2,3-dihydro-5-methyl-3-[{4-morpholinyl}methyl]pyrol [1,2,3-de]-1,4-benzoxazin-6-yl)(1-naphtalenyl)methanone monomethanesulfonate}]. After 90 min at 30°C, the membranes were filtered on a Tomtec Harvester 96 (Tomtec, Orange, CT) programmed to collect and wash captured membranes rapidly three times with 5 ml HEPES/BSA (Kuster et al., 1993). Specific binding was defined as the fraction of binding displaced by 1 µM unlabeled WIN 55,212-2. Saturation studies were analyzed by computer with the INPLOT program (Graphpad, San Diego, CA). For adenylyl cyclase studies, CB₁-expressing cells grown on 24-well plates were rinsed twice with PBS and preincubated first with Dulbecco's modified Eagle's medium for 60 min, then with Dulbecco's modified Eagle's medium, 1 mM isobutylmethylxanthine, and 3 µM BSA for 10 min. When appropriate, 100 nM WIN 55,212-2 or WIN 55,212-3 was added, and the sample was incubated for 5 min. The samples were next incubated for 10 min at 37°C with or without 10 µM forskolin and lysed with 0.1 N HCl; cAMP was measured by a scintillation proximity assay kit (Amersham Corp., Arlington Heights, IL).

Internalization. AtT20 cells stably expressing either wild-type CB₁ or D164N were treated as indicated in the figure legends. At the end of the experiment, cells were washed with phosphate buffer, fixed with 4% paraformaldehyde, permeabilized in 10% nonfat dry milk/0.1% saponin in PBS, and incubated with rabbit polyclonal antibody directed against the amino terminus of the CB₁ receptor (Twitchell et al., 1997). The cells were then washed extensively with PBS, and bound primary antibody was detected by FITC-conjugated anti-rabbit IgG secondary antibody (Zymed, San Francisco, CA). After washes in PBS, phosphate buffer, and water, coverslips were dried, mounted with Vectashield (Vector Laboratories, Burlingame, CA), and viewed with a Bio-Rad MRC600 confocal microscope (Bio-Rad Laboratories, Hercules, CA). Images were processed with Adobe Photoshop and Canvas. Additional experiments were performed with cells that were incubated for 24 h in serum-free medium containing 10 or 100 nM SR141716A (SR; [N-9-piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide hydrochloride]) plus 1 mg/ml BSA, which was exchanged with fresh serum-free medium containing SR and BSA 2 h before the experiment to maintain treatment consistent with that used for the MAP kinase assay with identical results.

MAP Kinase Assay. Stably transfected AtT20 cells were plated on six-well plates. Twenty-four hours before the experiment, the cells were incubated in serum-free medium containing either 10 or 100 nM SR plus 1 mg/ml BSA to reduce basal levels of MAP kinase activity. The medium was exchanged with fresh serum-free medium containing SR and BSA 2 h before the experiment. Cells were then incubated with medium appropriate for the desired experiment for 15 min. The experiment was stopped with ice-cold PBS, and the cells were lysed with buffer containing SDS, bromophenol blue, and -mercaptoethanol. The plates were then scraped, and the resulting samples were then normalized for cell protein by using a BCA protein assay kit (Pierce Chemical Co., Rockford, IL). The samples were sonicated for ~15 s, heated for 5 min, and centrifuged for 5 min; then, the supernatants were resolved by SDS-10% polyacrylamide gel electrophoresis. The gels were transferred to nitrocellulose filters, blocked with 5% milk Tris-buffered saline, probed with a primary antibody specific for the dually phosphorylated form of p44/42 MAP kinase (New England Biolabs, Beverly, MA), probed with an anti-rabbit peroxidase-labeled secondary antibody, and detected by enhanced chemiluminescence (Amersham).

Electrophysiological Recording. AtT20 cells stably transfected with rat CB₁ (Mackie et al., 1995) or D164N were plated on polylysine-coated coverslips and grown in Dulbecco's modified Eagle's medium + 10% heat-inactivated horse serum + 1:200 penicillin/streptomycin + 400 µg/ml G418 in a humidified environment with 5% CO₂ at 37°C. Cells were passaged with 0.05 mg/ml trypsin in PBS and used within 15 passages after the initial clones were isolated. Currents were recorded by using the whole-cell voltage clamp technique (Hamill et al., 1980). Pipettes were pulled from microhematocrit glass (VWR Scientific, Seattle, WA) and fire polished. For recording, a coverslip containing cells was transferred to a 200-µl chamber that was constantly perfused (1-2 ml/min) with the appropriate external solution. Solution reservoirs were selected by means of a series of solenoid valves, and solution changes were accomplished in <30 s. Voltage protocols were generated and currents were digitized, recorded, and analyzed with the Pulse program (HEKA Elektronik, Lambrecht, Germany). Liquid junction potentials are uncorrected.

	Results

Top Summary Introduction Materials and Methods Results Discussion References

A D164N Point Mutation of the CB₁ Receptor Does Not Affect Agonist Binding to the Receptor or Receptor-Mediated Inhibition of Forskolin-Induced cAMP Production. In other members of the seven-membrane-spanning receptor family, mutation of aspartate residues homologous to D164 disrupts agonist binding to the receptor (Strader et al., 1987; Chung et al., 1988; Ho et al., 1992; Chanda et al., 1993; Brodbeck et al., 1995; Parent et al., 1996; Chakrabarti et al., 1997), leading to the hypothesis that this conserved aspartate is part of the ligand binding pocket. To test whether D164 contributed to agonist binding to the CB₁ receptor, we measured specific binding of [³H]WIN 55,212-2 to membranes containing D164N and compared the values to those for wt CB₁. The specific binding of the synthetic CB₁ agonist [³H]WIN 55,212-2 to AtT20 cell membranes that expressed D164N receptor (K_d = 5.5 ± 1.3 nM, Fig. 1, A and B) was comparable to binding to membranes expressing wt CB₁ receptors (K_d = 4.3 ± 2.8 nM; Fig. 1B). The B_max for [³H]WIN 55,212-2 binding was also similar for D164N (390 ± 40 fmol/mg, Fig. 1C) when compared with wt CB₁ (450 ± 220 fmol/mg; Fig. 1C). These data indicate that expression levels as well as affinity for the synthetic agonist [³H]WIN 55,212-2 are similar for D164N and the wt CB₁ receptor for the cell lines studied.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   A D164N point mutation of the CB1 receptor does not affect agonist binding to the receptor, nor does it affect receptor-mediated inhibition of forskolin-induced cAMP production. A, specific binding of [3H]WIN 55212-2 to membranes prepared from AtT20 cells stably expressing D164N. Specific binding was 50% of total binding at the Kd. B, forskolin induced cAMP production in the absence () and presence () of CB1 agonist WIN (200 nM) for AtT20 cells stably expressing either wt CB1 (left) or D164N (right) (n = 4-6).

D164 Is Not Necessary for CB₁ Receptor-Mediated Inhibition of Ca²⁺ Channels. The wild-type CB₁ receptor has been shown previously to inhibit high voltage-activated N- and P/Q-type Ca²⁺ channels in AtT20 cells (Mackie et al., 1995). The inhibition is voltage dependent, a widely observed type of G protein-mediated inhibition characterized by: 1) slowed channel activation kinetics, and 2) reversal of inhibition by large depolarizing prepulses (Marchetti et al., 1986; Bean, 1989). We found that the CB₁ agonist WIN caused an inhibition of I_Ba in cells expressing D164N (Fig. 2A). The inhibition was dose-dependent, with an IC₅₀ of 6.8 ± 1.3 nM (Fig. 2B). A nearly identical IC₅₀ was found for cells expressing the wt CB₁ receptor (8.0 ± 1.6 nM; Fig. 2B). The maximal inhibition of peak current amplitude by 200 nM WIN was also similar for cells expressing either wt CB₁ or D164N-CB₁ receptors (Fig. 2C). The WIN-induced inhibition of I_Ba was blocked by overnight incubation with PTX in cells expressing D164N as well as in cells expressing wt CB₁ (Fig. 2C), implicating a G protein from the G_i/G_o family in the inhibition, via D164N, as well as wt CB₁. Thus, the D164N mutation does not alter the ability of the CB₁ receptor to inhibit Ca²⁺ channels or the nature of the inhibition.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   D164 is not necessary for CB1 receptor-mediated inhibition of Ca2+ channels. A, whole cell current recordings from AtT20 cells stably expressing the wt CB1 (left) or the CB1 receptor with the D164N point mutation (right). Shown are the currents elicited with a voltage step to +10 mV from a holding potential of 80 mV with no prior treatment (Control), after application of 100 nM WIN, and after application of 100 µM Cd2+. A 10 mM concentration of Ba2+ was the charge carrier. B, dose-response curves showing inhibition of barium current as a function of increasing WIN concentration. Shown are the results obtained for wild-type CB1 receptor (; n = 3-9) and D164N (; n = 2-4). C, mean inhibition (± S.E.M.) of barium current after application of 200 nM WIN () in AtT20 cells stably expressing either wt CB1 (n = 9) or D164N (n = 4). Also shown is the mean inhibition of barium current after overnight incubation with pertussis toxin () for cells expressing wt CB1 (n = 3) or D164N (n = 3).

D164 Is Necessary for CB₁ Receptor-Mediated Potentiation of KIR Currents. Activation of the wt CB₁ receptor also increases the KIR current in AtT20 cells (Fig. 3, A and B) by coupling to a PTX-sensitive G protein (Mackie et al., 1995). This potassium current also can be activated by somatostatin receptor agonists via receptors endogenous to the AtT20 cell line (Fig. 3, A and B) (Pennefather et al., 1988). A mutation of an aspartate residue of the ₂-adrenergic receptor that is homologous to the D164N mutation of the CB₁ receptor has previously been shown to block coupling of the ₂-receptor to activation of the KIR current in these cells (Surprenant et al., 1992). We tested whether cells expressing the D164N receptor were still capable of activating KIR current in response to a CB₁ agonist. We found that, unlike wt CB₁, the D164N receptor could not activate KIR (Fig. 3). However, the KIR current could still be activated in the D164N-expressing cells by activation of the endogenous somatostatin receptor. The mean potentiation of KIR current by 200 nM WIN was 4.0 ± 0.6 pA/pF for wt CB₁-expressing cells but only 0.3 ± 0.2 pA/pF for cells expressing D164N. In contrast, the mean potentiation of KIR current produced by somatostatin application was not significantly different for these two cell lines (Fig. 3C). At all WIN concentrations tested, we found very little potentiation of KIR current in D164N cells (Fig. 3D). Thus, the D164N mutation largely abolishes the ability of the CB₁ receptor to activate KIR.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   D164 is necessary for CB1 receptor-mediated potentiation KIR currents. A, whole cell current recordings from AtT20 cells stably expressing the wt CB1 (left) or the CB1 receptor with the D164N point mutation (right). Shown are the currents elicited with a voltage step to 100 mV from a holding potential of 45 mV with no prior treatment (Control), after application of 100 nM WIN, after application of somatostatin (SOM), and after application of 2 mM Ba2+. B, the time course of KIR current potentiation for both wt CB1 and D164N. KIR current amplitude is plotted versus time. Each circle represents the average current amplitude attained at a voltage step to 100 mV, stepped at 5-s intervals. Bars represent the period of time in which the various agents were applied to the bath. C, the mean potentiation (± S.E.M.) of KIR current after application of 200 nM WIN () in cells expressing either wt CB1 (left; n = 8) or the D164N mutant (right; n = 7). Also shown is the potentiation of KIR via an endogenous somatostatin receptor () in the cells expressing wt CB1 (left; n = 6) or D164N (right; n = 14). D, dose-response curve showing potentiation of KIR current as a function of increasing concentrations of the CB1 agonist WIN. Shown are the results obtained for both wt CB1 (; n = 5-8) and D164N (; n = 2-7). The EC50 for the WIN-induced KIR activation of wt CB1 was 8.0 ± 4.9 nM. The D164N dose-response curve could not be fitted.

Inability of the Mutant Receptor to Potentiate KIR Currents Is Not Agonist Specific. Previous mutations in other seven-transmembrane receptors of aspartate residues homologous to D164N in the CB₁ receptor have shown effects on agonist binding. Therefore, we wanted to test a structurally dissimilar CB₁ agonist, CP-55,940 (CP; {1-2-(R)-5-(1,1dimethylheptyl)-2-[5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-phenol}), to preclude the possibility that the failure of WIN to activate the KIR current in D164N cells was specific to this agonist. Application of CP to cells expressing D164N did not activate the KIR current, whereas application of somatostatin caused a robust activation of KIR current in the same cells (Fig. 4, A-C). However, CP inhibited Ca²⁺ channels when applied to D164N- expressing cells (Fig. 4, D and E), similar to the results seen with application of WIN (Fig. 3). Thus, the loss of activation of the KIR current after the D164N mutation is not specific to the agonist WIN.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Inability of the D164N mutant receptor to potentiate KIR currents is not agonist specific. A, whole cell KIR current recordings from AtT20 cells stably expressing the CB1 receptor with the D164N point mutation. Shown are the currents elicited with a voltage step to 100 mV from a holding potential of 45 mV with no prior treatment (Control), after application of 100 nM CP-55,940 (CP), after application of somatostatin (SOM), and after application of 2 mM Ba2+. B, the time course of KIR current potentiation for cells expressing D164N. KIR current amplitude is plotted versus time, each circle representing the average current amplitude attained at a voltage step to 100 mV from a holding potential of 45 mV at 5-s intervals. Bars represent the period of time during which the various agents were applied to the bath. C, the mean potentiation of KIR current via the agonists CP (solid; n = 5) and SOM (hatched; n = 3) in AtT20 cells stably expressing D164N. D, whole cell barium current recordings from AtT20 cells stably expressing D164N. Shown are the currents elicited with a voltage step to +10 mV from a holding potential of 80 mV with no prior treatment (Control), after application of 100 nM CP, and after application of 100 µM Cd2+. E, the mean inhibition (± S.E.M.) of barium current after application of 200 nM CP (; n = 6) or the muscarinic agonist oxotremorine (; n = 2) in AtT20 cells stably expressing D164N.

The Mutated Receptor Does Not Internalize on Agonist Exposure But Does Activate p42/44 MAP Kinase. Agonist-induced internalization of seven-membrane-spanning receptors on agonist application is a widely studied phenomenon and is thought to be involved in receptor down-regulation and/or resensitization (Zhang et al., 1997; Lefkowitz, 1998). When the wt CB₁ receptor is exposed to the agonist WIN, internalization occurs rapidly and is nearly complete after 15 min (Hsieh et al., 1999). Internalization is thought to result from agonist binding the receptor, activation of heterotrimeric G proteins, and subsequent activation of a number of proteins, including G protein receptor kinases and -arrestins (Zhang et al., 1997; Lefkowitz, 1998). These proteins, particularly -arrestin, are thought to form a complex with the receptor that aids in directing the receptor to clathrin-coated pits. These clathrin-coated pits are then endocytosed, and the receptor is either targeted to lysosomes or returned to the cell surface. Because we have documented above a selective disruption in signaling via the D164N mutant, we wanted to know whether this mutation affected receptor internalization. We exposed cells expressing mutant or wt receptors to 100 nM WIN for 15 min and compared the amount of receptor internalization. We found that the mutant showed no evidence of internalization after a 15-min WIN exposure (n = 6), an exposure that caused almost complete internalization of the wt receptor (Fig. 5A). Longer exposure (up to 1 h) also failed to cause any internalization of the mutant receptor (data not shown).

View larger version (74K): [in this window] [in a new window]   Fig. 5.   The D164N mutation blocks the receptor's ability to internalize on agonist exposure but does not block MAP kinase activation. A, a confocal image of AtT20 cells stably expressing either wt CB1 (top) or D164N (bottom) stained with antibody specific for the N terminus of the CB1 receptor. Cells were incubated for 15 min either untreated (0 WIN) or in the presence of 100 nM WIN (100 nM WIN). Arrows highlight regions that clearly demonstrate the location of the cell surface antibody-stained receptor, whereas arrowheads highlight antibody-stained internalized receptor. The scale bar indicates a length of 10 µm. B, Western blots probed with antibody specific for the dually phosphorylated form of p42/44 MAP kinase. AtT20 cells stably expressing either wt CB1 or D164N were incubated in serum-free media for 15 min at 37°C with the CB1 antagonist SR (10 nM) in addition to the various concentrations of WIN indicated above each lane as well as the PKC activator PMA (1 µM) and the MAP kinase-kinase inhibitor PD (100 µM). The dash indicates that no additional drugs were added. SR was applied to all cells in order to prevent an increase in phospho-p42/44 resulting from increased receptor activity on removal of antagonist.

A Reciprocal Mutation (D164N,N394D) Does Not Restore the Ability to Potentiate KIR Currents or Internalize on Agonist Exposure. The aspartate residue in the second transmembrane region is highly conserved through most members of the seven-membrane-spanning receptor family. In the GnRH receptor, however, an asparagine replaces the aspartate at this position and a highly conserved asparagine in the seventh transmembrane region is replaced by an aspartate. Molecular modeling indicates that the second and seventh transmembrane regions come into close proximity, suggesting that the highly conserved aspartate and asparagine residues in the second and seventh transmembrane regions interact with each other (Zhou et al., 1994). Indeed, when the residues at these two sites were exchanged in the 5-hydroxytryptamine (serotonin) 5-HT_2A receptor, the receptor functioned normally (Sealfon et al., 1995). However, mutating the second-transmembrane aspartate alone disrupted coupling of this receptor to phosphoinositide turnover. Here we tested whether a similar double mutation could restore functional coupling of the D164N receptor.

View larger version (58K): [in this window] [in a new window]   Fig. 6.   A reciprocal mutation (D164N/N394D) does not restore the receptor's ability to potentiate KIR currents or internalize on agonist exposure. A, the time course of KIR current potentiation for D164N/N394D. KIR current amplitude is plotted versus time. Each circle represents the average current amplitude attained at a voltage step to 100 mV. Bars represent the period of time during which the various agents were applied to the bath. B, the mean potentiation (± S.E.M.) of KIR current after application of 200 nM WIN in cells expressing either wt CB1 (left; ; n = 4), D164N (middle; ; n = 7), or D164N/N394D (right; ; n = 6). C, the time course of IBa current inhibition for D164N/N394D. IBa current amplitude is plotted versus time. Each circle represents the average current amplitude attained at a voltage step to +10 mV from a holding potential of 80 mV. Bars represent the period of time in which the various agents were applied to the bath. D, the mean inhibition (± S.E.M.) of IBa current after application of 200 nM WIN in cells expressing either wt CB1 (left; n = 9), D164N (middle; n = 4), or D164N/N394D (right; n = 8). E, confocal images of AtT20 cells stably expressing either wt CB1 (top) or the double point mutant D164N/N394D (bottom) stained with antibody specific for the N terminus of the CB1 receptor. Cells were incubated for 1 h either untreated (control) or in the presence of 100 nM WIN. Arrows highlight regions that clearly demonstrate the location of the cell surface antibody-stained receptor; arrowheads highlight antibody-stained internalized receptor. Scale bar indicates a length of 10 µm.

	Discussion

Top Summary Introduction Materials and Methods Results Discussion References

We have shown that mutation of a highly conserved aspartate residue in the second transmembrane region of the CB₁ cannabinoid receptor disrupts certain G protein signaling cascades without affecting others. The mutation did not affect binding of the CB₁ agonist WIN 55,212-2. These results suggest that D164 is required for the correct coupling of agonist-activated receptor to the stimulation of certain G proteins. The mechanism underlying the selectivity is unclear. Possibilities include: 1) selectively altering coupling of specific families of G proteins to the activated receptor; 2) reducing the activation efficiency of all coupled G proteins with effectors differentially affected by this change; and 3) altering the localization of the receptor in relation to the various effectors.

Selectively altering the coupling of specific G protein families is possible but seems unlikely. Both voltage-dependent inhibition of Ca²⁺ channels and activation of KIR are thought to result from direct binding of G protein subunits. Because these two actions are differentially affected by the D164N mutation, after the D164N mutation, the CB₁ receptor would have to specifically couple to a PTX-sensitive G protein with a subunit that is capable of modulating I_Ba but not KIR. Although a G protein subunit (G) with such specificity may exist, it appears unlikely, as at this time no G is known to have such a profile (Yan and Gautam, 1997; Garcia et al., 1998).

A reduced coupling efficiency of several G protein families has been documented previously after a similar mutation to the ₂-adrenergic receptor (Chabre et al., 1994), suggesting that this mutation may affect coupling of the receptor to all G proteins but the decreased efficiency only blocks coupling to some effectors. Differing effector sensitivity may reflect whether these effectors are modulated directly by G protein subunits or, instead, by second messenger cascades that could potentially amplify even a very small signal. Because voltage-dependent inhibition of Ca²⁺ channels and activation of KIR are both thought to result from direct binding of Gs, this possibility is unlikely. Alternatively, differing effector sensitivity may reflect the number of G protein subunits required to modulate the different effectors. Inhibition of Ca²⁺ channels and activation of KIR may have different stoichiometric requirements for Gs and thus would be differentially affected by a reduction in G concentrations. However, a comparison of the dose-response curves for WIN-induced modulation of Ba²⁺ current and KIR current via wt CB₁ receptors reveals almost identical sensitivity (Figs. 2B and 3D), making the possibility of different G requirements less likely.

Both activation of inwardly rectifying potassium currents and inhibition of Ca²⁺ channels are membrane-delimited forms of G protein inhibition, and as such they are dependent on close localization of the activated receptor to the channel. Our results could be explained if the D164N mutation interfered with colocalization of the CB₁ receptor with KIR channels but not with Ca²⁺ channels. Such an effect would selectively alter signaling to these channel types in agreement with our data. Localization of receptor is also important for internalization, because the receptors are first clustered in clathrin-coated pits on the cell membrane and subsequently internalized. The D164N mutation likely prevents internalization at a step before clustering of the receptor in clathrin-coated pits, inasmuch as the antibody staining remains evenly distributed over the entire membrane. We cannot differentiate from these data, however, whether the lack of clustered receptor is a failure in activation of the G protein signaling cascade or a failure in a localization signal requiring D164 occurring subsequent to G protein activation.

Mutations of this conserved aspartate residue in many different G protein-coupled receptors affect binding of agonist to the receptor (Strader et al., 1987; Chung et al., 1988; Ho et al., 1992; Chanda et al., 1993; Brodbeck et al., 1995; Parent et al., 1996; Chakrabarti et al., 1997; Tao and Abood, 1998). We have shown here, however, that mutation of D164 had no effect on agonist (WIN 55,212-2) binding to the CB₁ receptor, indicating that this residue was not a necessary component of the WIN binding pocket, and that any conformational change caused by the D164N mutation does not impact binding of agonist to this pocket.

The conserved aspartate in the second transmembrane region is thought to interact with a conserved asparagine in the seventh transmembrane region (Zhou et al., 1994), and several studies show that exchanging these two residues results in a receptor that functions identically to the wild type (Zhou et al., 1994; Sealfon et al., 1995). However, when we made this double mutation in the CB₁ receptor, the mutated receptor behaved similarly to the D164N mutant rather than the wild-type CB₁ receptor. These data indicate that, contrary to the results seen for other receptors, an aspartate in the 164th position of the CB₁ receptor seems to be critical for functional coupling rather than for interaction with the asparagine in the seventh transmembrane region. This aspartate could interact with another region of the protein.

Our results indicate that the D164N mutation had no effect on inhibition of cAMP production or binding of the CB₁ agonist WIN 55,212-2. These results are in contrast to a previous report that found that a similar point mutation in the human CB₁ receptor (D163N) blocked coupling of the receptor to inhibition of cAMP production and significantly reduced the receptor's affinity for WIN (Tao and Abood, 1998). Interestingly, a similar disparity was seen for the analogous mutation in the ₂-adrenergic receptor (D79N). In AtT20 cells, the D79N mutation had no effect on the coupling of this receptor to the inhibition of adenylyl cyclase (Surprenant et al., 1992), whereas in Chinese hamster ovary cells, the D79N mutation disrupted coupling of the receptor to inhibition of adenylyl cyclase (Wang et al., 1991). Thus, the disparity between our results and those of Tao and Abood (1998) is likely a product of the cell type used for expression of the receptor. The cell types used may provide different ratios of G protein families that couple to the CB₁ receptor, provide different forms of adenylyl cyclase, or localize the receptors in different manners. Comparison of receptor localization and specific G protein families activated by the CB₁ receptor in these two cell types may offer clues to the mechanism underlying the selective disruption of G protein signaling after mutation of D164.

We have found that the CB₁-D164N mutant, a GPCR that does not internalize, is capable of activating p42/44 MAP kinase, indicating that receptor internalization is not necessary for activation of this cascade. Previous experiments demonstrating that receptor internalization is necessary for p42/44 MAP kinase activation were carried out with dominant negative dynamin mutants to block internalization of -adrenergic receptors (Daaka et al., 1998). These results, taken in light of our data, suggest either that internalization is necessary only for certain receptor types to activate MAP kinase or that the dynamin mutants block activation of p42/44 MAP kinase via a mechanism other than blockade of receptor internalization. A less likely explanation is that D164N may have a low level of internalization, undetectable through the use of specific antibodies, that is sufficient to fully activate the MAP kinase cascade. Even more surprising was the finding that the D164N mutant activated p42/44 MAP kinase better than wt receptor, suggesting that the mutated receptor coupling to G proteins involved in activation of the p42/44 MAP kinase cascade may be enhanced. Alternatively, the enhanced activation of P42/44 MAP kinase may simply reflect the lack of internalization of this mutant receptor as the receptor would be exposed to agonist at the cell surface for longer periods of time. This result makes a general decrease in receptor/G protein coupling unlikely and favors either altering the specific G protein families coupled to the receptor or altered localization of the receptor.

In summary, a highly conserved aspartate residue in the CB₁ receptor is responsible for correctly coupling receptor activation to some, but not all, intracellular signaling cascades. Future studies will determine whether the loss of coupling to specific effectors results from: 1) the loss of activation of specific subtypes of G proteins; 2) a general reduction in G protein activation with different effector sensitivity to the reduced G protein activation; or 3) altered localization of the receptor with respect to certain effectors.