Address correspondence and reprint requests to Eamonn Kelly, Department of Pharmacology, School of Medical Sciences, University of Bristol, Bristol BS8 1TD, UK. E-mail: E.Kelly@bristol.ac.uk
Three group I mGluR antagonists CPCCOEt, LY367385 and BAY36-7620, were analyzed for their effect on cell surface expression of metabotropic glutamate receptor 1a and 1b. All three antagonists inhibited glutamate-induced internalization of mGluR1a and mGluR1b. However, when added alone, either LY367385 or BAY36-7620 increased the cell surface expression of mGluR1a but not mGluR1b. Both LY367385 and BAY36-7620 displayed inverse agonist activity as judged by their ability to inhibit basal inositol phosphate accumulation in cells expressing the constitutively active mGluR1a. Interestingly, mGluR1a but not mGluR1b was constitutively internalized in HEK293 cells and both LY367385 and BAY36-7620 inhibited the constitutive internalization of this splice variant. Furthermore, coexpression of dominant negative mutant constructs of arrestin-2 [arrestin-2-(319–418)] or Eps15 [Eps15(EΔ95-295)] increased cell surface expression of mGluR1a and blocked constitutive receptor internalization. In the presence of these dominant negative mutants, incubation of cells with LY367385 and BAY36-7620 produced no further increase in cell surface expression of mGluR1a. Taken together, these results suggest that the constitutive activity of mGluR1a triggers the internalization of the receptor through an arrestin- and clathrin-dependent pathway, and that inverse agonists increase the cell surface expression of mGluR1a by promoting an inactive form of mGluR1a, which does not undergo constitutive internalization.
sodium dodecyl sulfate–polyacrylamide gel electrophoresis
Glutamate is the major excitatory neurotransmitter in the central nervous system and exerts its actions via both ionotropic and metabotropic receptors (Nakanishi 1994; Conn and Pin 1997). Metabotropic glutamate receptors (mGluRs) are G protein-coupled receptors (GPCRs) and have been classified into three groups, with mGluR1 and mGluR5 constituting group I (Conn and Pin 1997). Five splice variants for mGluR1 exist, all of which differ in the length of their C-terminal tail, but all are coupled to the phospholipase C pathway via Gq/11 proteins (Pin et al. 1992; Conn and Pin 1997). Activation of mGluRs typically initiates processes involved in the negative feedback regulation of the receptor, like the desensitization of receptor responsiveness and receptor internalization (Dale et al. 2002). GPCR desensitization is often a consequence of G protein uncoupling in response to receptor phosphorylation by second messenger-dependent protein kinases (e.g. PKC) or G protein-coupled receptor kinases (GRKs), and the phosphorylation of mGluR1a by PKC (Francesconi and Duvoisin 2000) and GRKs (Dale et al. 2000; Sallese et al. 2000) has been reported. The phosphorylation of the receptor by GRKs promotes the binding of non-visual arrestins, which physically uncouple mGluR1 from G proteins and target the receptors for internalization in clathrin-coated pits (Mundell et al. 2001).
GPCR activation of downstream effectors in the absence of agonist, first reported for δ-opioid receptors (Costa and Herz 1989), has now been observed with a number of GPCRs. Interestingly, the long splice variant mGluR1a shows constitutive activity in the absence of agonists, while the short splice variant mGluR1b does not (Prezeau et al. 1996). In the case of GPCR constitutively active mutants (CAMs), constitutive receptor activity is often accompanied by constitutive internalization, also called agonist-independent or tonic internalization (for a review see Parnot et al. 2002). Constitutive internalization has also been observed for some wild type GPCRs, including thromboxane A2β receptors (Parent et al. 2001) and PAR1 protease-activated receptors (Shapiro and Coughlin 1998). Interestingly, agonist-independent internalization of the constitutively active mGluR1a (Dale et al. 2001) and mGluR5a (Fourgeaud et al. 2003) receptors has recently been described.
In this study we have analyzed the effect of different group I mGluR antagonists on surface expression and trafficking of the long and short mGluR1 splice variants, mGluR1a and mGluR1b, transiently expressed in HEK293 cells. Both splice variants were tagged in the N-terminus with a hemagglutinin epitope, and cell surface expression of the resultant constructs assessed by a combination of ELISA, immunofluorescence and immunoblotting with specific antibodies. The results indicate that antagonists with inverse agonist activity increase the cell surface expression of mGluR1a, and that this occurs as a result of the ability of these inverse agonists to inhibit the arrestin- and clathrin-dependent constitutive internalization of mGluR1a.
Materials and methods
Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum, Lipofectamine 2000 and poly-L-lysine were purchased from Life Technologies Inc. (Paisley, UK). Sodium pyruvate, glutamate pyruvate transaminase (GPT enzyme), glutamate, cycloheximide, the anti-β-tubulin antibody and the secondary antibody utilized for ELISA (goat anti-mouse conjugated with alkaline phosphatase) were obtained from the Sigma-Aldrich Company (Poole, Dorset, UK). The primary antibody (anti-HA monoclonal; HA-11) used in ELISA assays, the rhodamine-conjugated mouse monoclonal anti-HA antibody (12CA5) used in immunofluorescence experiments, the anti-HA high affinity antibody 3F10 used in western blot experiments and the Mini Complete protease inhibitor cocktail were from Roche Applied Science (Lewes, East Sussex, UK). EZ-Link Sulfo-NHS-SS-Biotin and MagnaBind Streptavidin beads were obtained from Pierce (Rockford, IL, USA), while colorimetric alkaline phosphatase substrate and AG 1-X8 resin columns were obtained from Bio-Rad (Hercules, CA, USA). myo-[3H]inositol, horseradish peroxidase-linked secondary antibodies and the enhanced chemiluminescence system (ECL) used in western blot experiments were supplied by Amersham Life Science (Little Chalfont, Buckinghamshire, UK). CPCCOEt and LY367385 were purchased from Tocris-Cookson (Avonmouth, Avon, UK). BAY36-7620 was kindly provided by Bayer (Wuppertal, Germany), pcDNA3-DNM-arrestin by Professor J.L. Benovic (Thomas Jefferson University, Philadelphia, USA) and EGFP-C2-Eps15(EΔ95-295) by Dr A. Benmerah (Institut Cochin, Paris, France).
Production of epitope-tagged mGluR1a and mGluR1b receptors
The construction of rat mGluR1a and mGluR1b receptors tagged with a HA-epitope between amino acids 57 and 58 of the N-terminus and their subcloning into the pcDNA3 vector have been described in full previously (Mundell et al. 2001; Mundell et al. 2002).
Cell culture and transfection
HEK293 cells were maintained in DMEM supplemented with 10% fetal bovine serum, 100 units/mL penicillin G, and 100 µg/mL streptomycin sulfate (from now on referred to as ‘culture medium’) at 37°C in a humidified atmosphere of 95% air, 5% CO2. For transient transfections, HEK293 cells were grown in 100 mm dishes to 80–90% confluence and transfected with recombinant plasmid DNA (pcDNA3-mGluR1a or pcDNA3-mGluR1b ± pcDNA3-DNM-arrestin-2 [arrestin-2 (319–418)] or EGFP-C2-Eps15(EΔ95-295) using Lipofectamine 2000 following the manufacturer's instructions. In experiments with cells cotransfected with mGluR1a and a dominant negative mutant plasmid (pcDNA3-DNM-arrestin-2 or EGFP-C2-Eps15(EΔ95-295)), control cells were also transfected with the receptor DNA plus the empty pcDNA3 or EGFP-C2 vectors. After 24 h of incubation in serum-free DMEM containing a DNA/Lipofectamine 2000 mixture, the cells were cultured in normal culture medium for a further 24 h before experiments.
ELISA analysis of mGluR1a and mGluR1b cell surface expression in HEK293 cells
The cell surface expression of mGluR1a and mGluR1b was assessed by ELISA as described previously (Mundell et al. 2001). Briefly, cells plated at a density of around 1 × 106 cells per 100-mm dish were transiently transfected as described above and 24 hours post-transfection, cells were split into 24-well tissue culture dishes coated with 0.1 mg/mL poly l-lysine and maintained in culture medium. A further 24 h later, culture medium was replaced with fresh culture medium containing different drug treatments and cells were incubated for a further 30 min (unless otherwise specified) at 37°C in a humidified atmosphere of 95% air, 5% CO2. For experiments in the presence of glutamate pyruvate transaminase (GPT), the cell incubation was performed in culture medium supplemented with 5 mm sodium pyruvate and 3 units/mL of GPT. Reactions were stopped by removing the medium and fixing the cells with 3.7% formaldehyde in TBS (20 mm Tris, pH 7.5, 150 mm NaCl, 20 mm CaCl2) for 5 min at room temperature. Cells were then washed three times with TBS, incubated for 45 min with TBS containing 1% BSA (TBS/BSA), and then incubated with a primary antibody (anti-HA monoclonal HA-11, 1 : 1000 dilution in TBS/BSA) for 1 h at room temperature. Cells were then washed three times with TBS, reblocked with TBS/BSA for 15 min at room temperature, and then incubated with secondary antibody (goat anti-mouse conjugated with alkaline phosphatase, 1 : 1000 dilution in TBS/BSA) for 1 h at room temperature. Cells were then washed a further three times with TBS, and a colorimetric alkaline phosphatase substrate was added. After 20–30 min, 100 µL of sample was added to 100 µL of 0.4 m NaOH to terminate the reaction, and the samples were read at 405 nm using a microplate reader. Results are expressed as percentage change in cell surface receptor expression compared to untreated mGluR1a- or mGluR1b-expressing cells (A0), with the background signal from cells transfected with pcDNA3 alone (AHEK293) subtracted from all receptor-transfected values (AX). The formula is:
Analysis of constitutive internalization by ELISA
Cells were plated and transfected as described above. For experimentation, culture medium was replaced with fresh culture medium containing primary antibody (anti-HA monoclonal HA-11; dilution 1 : 1000) and incubated (first incubation) in the presence of the primary antibody for 45 min at 37°C in a humidified atmosphere of 95% air, 5% CO2. Cells were then washed once with TBS and incubated (second incubation) at 37°C for a further time period of between 0 and 120 min in fresh culture medium containing 5 mm sodium pyruvate, 3 units/mL GPT, various drug treatments but no primary antibody. The loss of cell surface receptor was then quantified by ELISA (see section above). Briefly, cells were treated with formaldehyde, incubated with secondary antibody, then washed three times with TBS. The same colorimetric procedure described above for cell surface expression was then applied. The background signal from pcDNA3-transfected controls (AHEK293) was subtracted from all receptor-transfected values. The amount of constitutively internalized mGluR1a was then calculated by the following formula:
where Ax is the ELISA reading for mGluR1a-expressing cells following different times and conditions of second incubation, and A0 is the ELISA reading for mGluR1a-expressing cells in the absence of treatments and with no second incubation.
Cell imaging by immunofluorescence confocal microscopy
Cell distribution of mGluR1a and mGluR1b was analyzed by confocal microscopy, as previously described (Mundell et al. 2001). Briefly, HEK293 cells, grown on poly l-lysine coated coverslips in six-well plates, were transiently transfected as described above and 48 hours post-transfection, culture medium was changed for TBS/BSA containing rhodamine-conjugated mouse monoclonal anti-HA antibody (12CA5; diluted 1 : 200). Cells were incubated at 4°C for 30 min, fresh culture medium containing different drug treatments was then substituted for the antibody incubation buffer, and cells were incubated for a further 45 min at 37°C in a humidified atmosphere of 95% air, 5% CO2. Reactions were stopped by removing the medium and fixing the cells with 3.7% formaldehyde in TBS for 5 min at room temperature. Coverslips were then mounted using Slow-Fade mounting medium and examined by microscopy on an upright Leica TCS-NT confocal laser scanning microscope attached to a Leica DM IRBE epifluorescence microscope (Leica, Wetzlar, Germany) with phase-contrast and a Plan-Apo 40 × 1.40 NA oil immersion objective.
Membrane protein preparation following surface biotinylation
Cells plated at a density of around 106 cells per 100 mm dish were transiently transfected as described above and 48 h later the medium was changed for fresh culture medium containing 5 mm sodium pyruvate, 3 units/mL GPT enzyme and different drug treatments. Cells were then incubated for 45 min at 37°C in a humidified atmosphere of 95% air, 5% CO2. After one wash in PBS, cells were incubated for 10 min at room temperature in PBS containing 0.5 mg/mL sulfo-NHS-SS-biotin, with occasional shaking. After rapidly washing three times in PBS containing 100 mm glycine to remove the excess of biotin, cells were solubilized in 400 µL of ice-cold Triton X-100 lysis buffer [20 mm HEPES, 200 mm NaCl, 10 mm EDTA, 1% Triton X-100 (v/v), 1 tablet/5 mL Mini Complete protease inhibitor cocktail] and the cell lysates were clarified by centrifugation (10 000 × g for 10 min). After protein quantification, 500 µg of protein per sample was mixed with 2.5 mg of Magnabind Streptavidin beads and resuspended in 500 μL of PBS containing Mini Complete protease inhibitor cocktail. After incubation for 30 min at room temperature with continuous agitation, beads were magnetically separated from the supernatant, resuspended in 20 µL of PBS containing Mini Complete protease inhibitor cocktail, snap-frozen in liquid nitrogen and stored at − 80°C until used for western blotting.
Total cell preparation
Cells were transiently transfected as described above and 48 h later the medium was changed with fresh culture medium containing 5 mm sodium pyruvate, 3 units/mL GPT enzyme and different drug treatments. Cells were then incubated for 45 min at 37°C in a humidified atmosphere of 95% air, 5% CO2. After one wash with ice-cold PBS, cells were solubilized in ice-cold Triton X-100 lysis buffer. The cell lysates were clarified by centrifugation (10 000 × g for 10 min) and cell samples were then snap-frozen in liquid nitrogen.
Five times loading buffer [315 mm Tris-HCl, pH 7.5, 0.5 m dithiothreitol, 5% SDS, 58% glycerol (v/v), 0.1% bromophenol blue (v/v)] was added to protein samples from total cell or cell membrane preparations. After heating for 5 min at 90°C, 20 µg of protein per lane was separated by 1% SDS–PAGE, blotted onto nitrocellulose and probed using an anti-HA antibody (high affinity 3F10; 1 : 1000) or an anti-β-tubulin antibody (1 : 1000). The immunoreactive bands were visualized using horseradish peroxidase-linked secondary antibody (1: 2000) and enhanced chemiluminescence (ECL) detection system.
Inositol phosphate determination
This technique was undertaken as described previously (Mundell and Benovic 2000). Briefly, cells were plated and transfected as described above. Twenty-four hours later, cells were split into 24-well tissue culture dishes coated with 0.1 mg/mL poly l-lysine. The following day cells were labelled for 18–24 h with myo-[3H]inositol (4 µCi/mL) in DMEM (high glucose, without inositol). After labelling and pretreatment with antagonists (e.g. Fig. 4), cells were washed once in PBS and incubated in prewarmed DMEM containing 20 mm LiCl and eventual mGluR1 antagonists. To investigate the constitutive activity of the receptor; cells were then incubated a further 30 min in the presence of 5 mm sodium pyruvate and 3 units/mL GPT. In order to study the agonist-induced activity of the receptor, glutamate to a final concentration of 30 µm was added and the cells incubated for a further 15–30 min. Reactions were terminated by removing the stimulation medium and adding 0.8 mL of ice-cold 0.4 m perchloric acid solution. Samples were harvested in Eppendorf tubes, and 0.4 mL of 0.72 m KOH, 0.6 m KHCO3 was added. Tubes were vortexed and centrifuged for 5 min at 20 000 × g in a microcentrifuge. Inositol phosphates (IPs) were separated on AG 1-X8 resin columns as described previously (Mundell and Benovic 2000). Total labelled IPs were then determined by liquid scintillation counting.
Experimental design and statistics
Log concentration-effect curves were fitted to logistic expressions for sigmoidal concentration-effect curves with variable slope using GraphPAD Prism (GraphPAD Software). Where appropriate, statistical significance was assessed using one-way anova with Bonferroni's multiple comparison post-test or ‘one sample’t-test.
In agreement with our previous studies (Mundell et al. 2001; Mundell et al. 2002), cell surface expression of mGluR1a and mGluR1b was decreased by 31 ± 4 and 48 ± 7%, respectively (means ± SEM, n = 6), following glutamate treatment (30 µm; 30 min). The glutamate-induced internalization of mGluR1a and mGluR1b was completely inhibited by co-addition of LY367385 (100 µm), BAY36-7620 (100 µm) or CPCCOEt (200 µm) (data not shown). However, strikingly, LY367385 and BAY36-7620 also increased mGluR1a, but not mGluR1b, surface expression above the control level for untreated receptor-expressing cells by 32 ± 5 and 29 ± 3%, respectively (means ± SEM, n = 6). The effect of LY367385, BAY36-7620 or CPCCOEt alone on the cell surface expression of mGluR1a and mGluR1b was then assessed (Fig. 1). Pretreatment of cells with LY367385 (100 µm) or BAY36-7620 (100 µm) for 30 min increased mGluR1a cell surface expression by 36 ± 6 and 30 ± 4%, respectively, whilst CPCCOEt (200 µm) had no effect (Fig. 1a). On the other hand mGluR1b cell surface expression was not affected by any of the antagonists (Fig. 1b). The increase in mGluR1a surface expression induced by LY367385 or BAY36-7620 was also present when glutamate pyruvate transaminase (GPT), an enzyme able to degrade glutamate present in the culture medium, was added to the incubation medium (Fig. 1c). To verify that the effect of these two antagonists was not due to synthesis of new receptor, LY367385 and BAY36-7620 were added in the presence of cycloheximide, a potent inhibitor of protein synthesis. HEK293 cells recombinantly expressing mGluR1a were incubated in culture medium containing 50 µg/mL of cycloheximide, and after 30 min the antagonists were added to the culture medium for a further period of 30 min and the surface expression of mGluR1a then assessed. Under these conditions, LY367385 and BAY36-7620 were still able to increase mGluR1a surface expression (data not shown). The time course of the antagonist-induced increase in mGluR1a cell surface expression was also determined (Fig. 2a). Treatment of mGluR1a-expressing cells with 100 µm LY367385 or BAY36-7620 induced a relatively rapid increase in receptor surface expression, being maximal after around 30 min treatment, which was stable for up to 24 h. Following removal of LY367385 or BAY36-7620 from the medium, the cell surface expression of mGluR1a rapidly returned back to control level. In contrast, treatment with CPCCOEt for up to 24 h induced little if any change in mGluR1a cell surface expression, whilst as expected glutamate treatment produced a reversible decrease in receptor surface expression. Furthermore, LY367385 and BAY36-7620 increased mGluR1a cell surface expression in a concentration-dependent manner (Fig. 2b), with EC50 values of 12 and 4 µm, respectively.
The increase in mGluR1a surface expression induced by LY367385 and BAY36-7620 was confirmed by immunoblotting (Fig. 3). Membrane proteins were separated from intracellular proteins using a surface biotinylation protocol and the expression of mGluR1a among membrane proteins was analyzed by immunoblot (Fig. 3, upper panel). The mGluR1a-specific band (∼140 kDa) was evident in the sample from transfected cells with no treatment (lane 1), but absent in sample from untransfected cells (lane 5). Treatment with 100 µm LY367385 (lane 2) or 100 µm BAY36-7620 (lane 3), but not 200 µm CPCCOEt (lane 4), increased the intensity of the band up to a maximum of around 75%. As a control, the total cell extract from recombinant HEK293 cells was analyzed by immunoblotting for the presence of mGluR1a (Fig. 3, lower panel). The intensity of the band corresponding to mGluR1a was largely unchanged in samples from recombinant cells treated with no antagonist (lane 1), 100 µm LY367385 (lane 2), 100 µm BAY36-7620 (lane 3) or 200 µm CPCCOEt (lane 4), and absent in untransfected HEK293 cells. To assess the effectiveness of the membrane preparation and of the membrane protein separation by surface biotinylation, the presence of a cytoplasmic marker protein, β-tubulin, was assessed by immunoblotting in both preparations. Using an anti-β-tubulin antibody, a single band of ∼55 kDa corresponding to β-tubulin was observed in the lane corresponding to the total cell preparation, but was completely absent in the membrane protein samples obtained by surface-biotinylation (Fig. 3, right panels). Furthermore, in ligand binding assays, LY367385 and BAY36-7620 treatment also increased specific [3H]quisqualate binding to membrane preparations of mGluR1a-expressing cells (data not shown). Taken together the results of these different approaches confirm that LY367385 and BAY36-7620 increase cell surface expression of mGluR1a without an increase in overall cell mGluR1a content.
We next undertook functional studies to investigate whether the LY367385 and BAY36-7620 pretreatment-induced increase in mGluR1a cell surface expression could alter basal or agonist-stimulated IP accumulation (Fig. 4). In the absence of drug treatment, basal IP accumulation was doubled in mGluR1a-expressing cells as compared to non-transfected controls, presumably as a consequence of the constitutive activity of mGluR1a (compared to non-transfected cells, basal IP accumulation was not increased by transfection with mGluR1b, data not shown). Pretreatment with LY367385 or BAY36-7620 (each 100 µm) increased subsequent basal (constitutive) IP accumulation in mGluR1a-expressing cells, whereas CPCCOEt had no effect (Fig. 4a). Furthermore, glutamate-induced (30 µm; 15 min) IP accumulation was also enhanced by LY367385 or BAY36-7620 pretreatment, whereas CPCCOEt pretreatment did not affect the glutamate response. Together, these results show that pretreatment of mGluR1a-expressing cells with LY367385 or BAY36-7620 increases both basal and agonist-stimulated IP accumulation, possibly as a result of the drug-induced increase in cell surface receptor number.
The direct effect of LY367385, BAY36-7620 and CPCCOET on basal and glutamate-stimulated IP accumulation in mGluR1a-expressing cells was next assessed. Stimulation with glutamate (30 µm; 30 min) induced an almost twofold increase in IP production in mGluR1a-expressing cells, which was in turn inhibited by all three antagonists (Fig. 5a). The effect of these three antagonists on the constitutive activity of mGluR1a as assessed by basal IP accumulation was then investigated (Fig. 5b). Interestingly, the constitutive activity of mGluR1a was strongly inhibited by LY367385 or BAY36-7620, but was unaffected by CPCCOEt, indicating that the former two antagonists are also inverse agonists at mGluR1a.
The subcellular localization of HA-tagged mGluR1a and mGluR1b in HEK293 cells was next investigated by immunofluorescence confocal microscopy using a rhodamine-conjugated anti-HA antibody (Fig. 6). In the absence of drug treatment, both mGluR1a (Fig. 6a) and mGluR1b (Fig. 6b) were localized to the cell membrane. Significant amounts of mGluR1a, but not mGluR1b, could also be detected intracellularly, presumably localized to endosomes as a consequence of constitutive receptor internalization. Interestingly, treatment of mGluR1a-expressing cells with 100 µm LY367385 (Fig. 6c), but not 100 µm CPCCOEt (Fig. 6d), noticeably reduced the intracellular localization of mGluR1a. As a comparison, the internalization of mGluR1a and mGluR1b following 100 µm glutamate is also shown (Fig. 6e and f, respectively).
In an attempt to quantify constitutive internalization of mGluR1a using the ELISA technique, we initially prelabelled cell surface receptors with anti-HA antibody at 4°C, followed by warming of the cells to 37°C in order to trigger constitutive internalization, as previously described (Dale et al. 2001). Using this protocol we observed rapid and extensive loss of cell surface expression for both mGluR1a and mGluR1b. Under these conditions, in agreement with Dale et al. (2001), we found that a dominant negative mutant form of arrestin-2 (DNM-arrestin; arrestin-2(319–418)) did not inhibit the constitutive internalization of mGluR1a (after an initial 30 min of antibody labelling at 4°C, the incubation of the cells at 37°C for 60 min induced a decrease in the cell surface expression of mGluR1a of 78 ± 2% and 74 ± 3%, in the absence and presence of DNM-arrestin expression, respectively (means ± SEM, n = 3)). However, we also observed that the background reading obtained for non-transfected HEK293 cells incubated with antibody also decreased rapidly when the temperature of the incubation medium was increased from 4 to 37°C for 5 min (52 ± 7% decrease, means ± SEM, n = 3). This suggests that the loss of ELISA signal observed under these experimental conditions may be due at least in part to factors other than constitutive receptor internalization, such as a temperature-induced reduction in antibody binding. Furthermore mGluR1b, which does not show significant intracellular localization in our immunofluorescence experiments (Fig. 6b) and was therefore not expected to constitutively internalize, apparently internalized at the same rate and to the same extent as mGluR1a when the cells were prelabelled with antibody at 4°C for the ELISA assays. To examine this further, we modified the experimental conditions such that the cells were prelabelled with anti-HA antibody at 37°C instead of 4°C. Using this protocol we now found that mGluR1a internalized by 46 ± 3% after a 30 min incubation period (Fig. 7a), whereas mGluR1b showed little constitutive internalization (Fig. 7b). Furthermore, coaddition of 100 µm LY367385 or BAY36-7620 markedly reduced the constitutive internalization of mGluR1a, whereas 200 µm CPCCOEt had no effect (Fig. 7a; also LY367385 and BAY36-7620 did not inhibit internalization when assessed in cells where mGluR1a receptors were labelled with antibody at 4°C, data not shown). The role of non-visual arrestins in mGluR1a constitutive internalization was further investigated. In cells coexpressing mGluR1a and DNM-arrestin, the constitutive internalization of mGluR1a was blocked (Fig. 7b). Furthermore, in cells expressing DNM-arrestin, the cell surface expression of mGluR1a measured using a standard ELISA protocol was increased by around 50% over that in cells expressing mGluR1a alone, such that LY367385 and BAY36-7620 had no further effect on cell surface mGluR1a expression (Fig. 8a). As expected, the glutamate-induced internalization of mGluR1a was also blocked by DNM-arrestin (Fig. 8a). The clathrin-dependence of mGluR1a constitutive internalization was also investigated, using a dominant negative mutant of Eps15 (DNM-Eps15; EΔ95/295), a scaffolding protein necessary for the formation of clathrin-coated pits. As with DNM-arrestin, DNM-Eps15 inhibited the constitutive internalization of mGluR1a (Fig. 7b) and increased the receptor surface expression of mGluR1a by around 50% (Fig. 8b). Furthermore, DNM-Eps15 prevented the increase or decrease in cell surface mGluR1a normally triggered by that LY367385/BAY36-7620 and glutamate, respectively (Fig. 8b). On the other hand, results obtained with cells cotransfected with mGluR1a and the empty vectors pcDNA3 or EGFP-C2 were not significantly different from results obtained with cells transfected with mGluR1a only (data not shown). The increase in receptor surface expression induced by DNM-arrestin and DNM-Eps15 was also confirmed by immunoblot (Fig. 8c). Membrane protein samples obtained following surface biotinylation were analyzed (Fig. 8c, upper panel). The cotransfection of DNM-arrestin (lane 2) or DNM-Eps15 (lane 3) resulted in a clear increase in mGluR1a membrane expression compared to cells transfected with only mGluR1a (lane 4). For comparison (Fig. 8c, lower panel) the expression of mGluR1a in the total cell extracts was not significantly altered by the coexpression of the DNM-constructs. To check whether cotransfection with another plasmid DNA influences the surface expression of mGluR1a, we cotransfected the receptor with empty pcDNA3 and empty EGFP-C2 vectors. In this case, no change in mGluR1a surface expression was found (data not shown).
In this study the effect of mGluR1 receptor antagonists CPCCOEt, BAY36-7620 and LY367385 on the trafficking of mGluR1a and mGluR1b transiently expressed in HEK293 cells was investigated by ELISA, immunofluorescence and immunoblotting. As expected, glutamate-induced internalization of both mGluR1a and mGluR1b was inhibited by these three antagonists, but LY367385 and BAY36-7620 also increased mGluR1a cell surface expression to greater than control. This increase was detectable also when mGluR1a-expressing cells were incubated with LY367385 or BAY36-7620 in the absence of glutamate. Since CPCCOEt did not produce a similar increase in mGluR1a cell surface expression, and also since addition of GPT to the medium did not reverse the effects of LY367385 and BAY36-7620, then the latter two antagonists cannot be increasing cell surface expression by blocking receptor internalization induced by any glutamate present in the culture medium. Furthermore, the inability of cycloheximide to inhibit the increase in mGluR1a surface expression induced by LY367385 and BAY36-7620 indicates that the increase is not due to synthesis of new receptor. Also, our experiments showed that the LY367385- and BAY36-7620-induced increase in cell surface expression of mGluR1a is rapid and reversible, since even after 24 h treatment, removal of drug from the medium led to the reversion of surface receptor levels to control values within about an hour. The potency with which LY367385 increased mGluR1a cell surface expression (EC50 = 12 µm) was similar to the published value for inhibition of quisqualate-induced IP accumulation (EC50 = 8 µm; Clark et al. 1997), whereas BAY36-7620 was somewhat less potent in the present study (EC50 = 4 µm) than originally reported for inhibition of glutamate-independent IP accumulation (EC50 = 0.38 µm; Carroll et al. 2001). A recent study reported that CPCCOEt increased the cell surface expression of mGluR1a stably expressed in L929sA mouse fibrosarcoma cells (Lavreysen et al. 2002). We did observe a small increase (5–10%) in mGluR1a cell surface expression after 24 h treatment with CPCCOEt, but much smaller than the ∼300% increase observed in L929sA cells. It is possible that differences in cell type or species of origin of the receptor (human receptor was used by Lavreysen et al. whereas we used the rat version) explain this difference.
Contrary to the effects on mGluR1a, neither BAY36-7620 nor LY367385 altered cell surface expression of mGluR1b. A major difference between the splice variants is that mGluR1a is constitutively active whilst short splice variants, such as mGluR1b, are not (Prezeau et al. 1996). To investigate whether this difference might be relevant to the effects of BAY36-7620 and LY367385 on receptor surface expression, we assessed the effect of the antagonists on the agonist-independent activity of mGluR1a. Using basal IP accumulation in mGluR1a-expressing cells as a measure of constitutive activity, we demonstrated that BAY36-7620 and LY367385 are inverse agonists, whereas CPCCOEt is not (Litschig et al. 1999). Unlike BAY36-7620 (Carroll et al. 2001), LY367385 has not previously been reported to exhibit inverse agonist activity at mGluR1a. A possible explanation for the ability of BAY36-7620 and LY367385 to increase cell surface expression of mGluR1a but not mGluR1b may be that antagonists with inverse agonist activity suppress the constitutive internalization of mGluR1a. This is supported by our immunofluorescence experiments, where in the absence of agonist, mGluR1a immunofluorescence is detected both at the cell surface and intracellularly, whereas mGluR1b is only detected at the cell surface. Of further interest, in these experiments LY367385 suppressed the intracellular accumulation of mGluR1a immunoreactivity, whereas CPCCOEt did not, suggesting that the inverse agonist activity of LY367385 is able to suppress receptor constitutive internalization. To quantify and confirm these findings, the constitutive internalization of mGluR1a was assessed using a modified ELISA protocol. Under experimental conditions where the cells are prelabelled with antibody at 37°C, we observed time-dependent internalization of mGluR1a but not mGluR1b, in line with our imaging results. Having established that mGluR1a undergoes constitutive internalization, we demonstrated that the two mGluR1 antagonists with inverse agonist activity inhibited constitutive internalization. On the other hand CPCCOEt did not affect the constitutive internalization of mGluR1a, which also confirms that the internalization is not due to activation of the receptor by glutamate present in the medium. Agonist-independent internalization has been previously observed for constitutively active mutants of GPCRs (CAMs; Parnot et al. 2002; Seifert and Wenzel-Seifert 2002). Interestingly, an antagonist with inverse agonist activity has been shown to increase the cell surface expression of CAM forms of the AT1 receptor (Miserey-Lenkei et al. 2002). CAMs apparently assume an active conformation of the receptor that mimics the agonist-induced active receptor conformation (Parnot et al. 2002) and may consequently be recruited for internalization by the mechanism underlying agonist-induced internalization. More importantly, constitutive activity has also been observed with some wild type GPCRs (reviewed in Seifert and Wenzel-Seifert 2002) including the group I mGluR splice variants with a long COOH-terminus tail (mGluR1a, mGluR5a and mGluR5b; Prezeau et al. 1996). Constitutively active group I mGlu receptors also appear to constitutively internalize (Dale et al. 2001; Fourgeaud et al. 2003), indicating a link between constitutive activity and constitutive internalization of GPCRs. It therefore appears likely that the inverse agonists inhibit the constitutive internalization of mGluR1a as a result of their ability to promote an inactive conformation of the receptor. Interestingly however, the mGluR5 receptor antagonist MPEP which displays inverse agonist activity (Pagano et al. 2000), was reported not to block the constitutive internalization of mGluR5a, but rather increased internalization (Fourgeaud et al. 2003). It is therefore possible that the mechanisms regulating constitutive internalization of mGluR1a and mGluR5 are different, or that the inactive conformation of mGluR5 promoted by MPEP remains able to interact with components of the internalization pathway.
The constitutive internalization of mGluR1a in HEK293 cells was reported to be independent of both arrestin and dynamin, although the receptor was found to colocalize with clathrin and transferrin (Dale et al. 2001). Using our modified ELISA protocol where cells are prelabelled with antibody at 37°C, we assessed the arrestin- and clathrin-dependence of mGluR1a constitutive internalization by coexpressing the receptor with DNM constructs of arrestin-2 (Krupnick et al. 1997) and Eps15 (Benmerah et al. 1999). Both DNM constructs markedly suppressed the constitutive internalization of mGluR1a, indicating that both the agonist-dependent (Mundell et al. 2001; Mundell et al. 2002) and agonist-independent internalization of mGluR1a proceed by an arrestin- and clathrin-dependent mechanism. Furthermore, the DNM constructs also increased cell surface expression of mGluR1a, and in the presence of the DNM constructs the two inverse agonists were unable to further increase cell surface expression of mGluR1a. Thus the effects of the DNM constructs actually mirror those of the inverse agonists, and the nonadditivity of the DNM/inverse agonist effects suggests that the mechanism by which BAY36-7620 and LY367385 increase cell surface expression of mGluR1a is by blocking arrestin- and clathrin-dependent internalization of the receptor. It remains unclear why the constitutive internalization of mGluR1a as measured by the protocol of Dale et al. (2001) is arrestin-independent. One possible explanation is that the temperature increase from 4 to 37°C reduces the affinity of the antibody for the receptor epitope, thereby reducing antibody binding and apparently mimicking receptor internalization in the ELISA or similar antibody-based assays. The constitutive internalization of mGluR1a as measured using the 4°C protocol could therefore be arrestin-independent because it does not reflect a genuine internalization process. On the other hand, the possibility that the two different antibody-labelling protocols (labelling at 4 or 37°C) measure two different forms of constitutive internalization cannot at present be ruled out. Further study will be required to resolve this issue, however, the results we obtain by prelabelling with antibody at 37°C suggest that the constitutive internalization we measure by this protocol is a physiologically relevant process, not least because it is blocked by specific manipulations including inverse agonist treatment and coexpression of DNM-constructs that are well characterized inhibitors of internalization. Furthermore, in preliminary experiments, we find that even in the absence of agonist, coexpressed arrestin-2 is able to coimmunoprecipitate with mGluR1a in HEK293 cells (unpublished observations), suggesting that arrestins are well placed to trigger constitutive internalization. Moreover, the finding that arrestins mediate the agonist-independent internalization of constitutively active variants of the µ-opioid (Li et al. 2001) and vasopressin receptor (Barak et al. 2001) suggests a broader role for arrestins in the constitutive internalization of GPCRs.
In summary, our results demonstrate that agonist-independent internalization of mGluR1a proceeds via an arrestin- and clathrin-dependent pathway that regulates overall cell surface expression of the receptor. This constitutive internalization is blocked by receptor antagonists with inverse agonist activity, suggesting that constitutive receptor activation is obligatory for constitutive internalization. It will be important in future studies to examine the effects of inverse agonists on mGluR1 receptor trafficking in neurons.
The authors would like to thank Dr T. Mueller and Dr J. Vry (Bayer AG, Wuppertal, Germany) for providing the compound BAY36-7620, Professor J.L. Benovic (Thomas Jefferson University, Philadelphia, USA) for supplying the DNM-arrestin 2 construct and Dr A. Benmerah (Institut Cochin, INSERM U567, University of Paris, France) for supplying the DNM-Eps15 construct. This work was supported by the Medical Research Council.