Address correspondence and reprint requests to Francisco Ciruela, Departament de Bioquímica i Biologia Molecular, C/Martí i Franquès, 1, 08028 Barcelona, Spain. E-mail: firstname.lastname@example.org
Recent evidence suggest that many G protein-coupled receptors (GPCR) and signalling molecules localize in microdomains of the plasma membrane. In this study, flotation gradient analysis in the absence of detergents demonstrated the presence of the metabotropic glutamate receptor type 1α (mGlu1α) in low-density caveolin-enriched membrane fractions (CEMF) in permanently transfected BHK cells. BHK-1α cells exhibit a similar pattern of staining for caveolin-1 and caveolin-2, and these two proteins show a high degree of co-localization with mGlu1α receptor as demonstrated by immunogold and confocal laser microscopy. The presence of mGlu1α in CEMF was also demonstrated by co-immunoprecipitation of mGlu1α receptor using antibodies against caveolin proteins. Activation of the mGlu1α receptor by agonist increased extracellular signal-regulated kinases phosphorylation in CEMF and not in high-density membrane fractions (HDMF), suggesting that mGlu1α receptor-mediated signal transduction could occur in caveolae-like domains. Overall, these results clearly show a molecular and functional association of mGlu1α receptor with caveolins.
sodium dodecyl sulphate polyacrylamide gel electrophoresis
vascular endothelial growth factor receptor
Glutamate is the major excitatory neurotransmitter in the CNS (Hollmann and Heinemann 1994) and its function through ionotrophic (iGlu) and metabotropic (mGlu) glutamate receptors can be modulated by other neurotransmitters/neuromodulators (Conn and Pin 1997). Eight members of the mGlu receptor family have been identified and categorized into three subgroups on the basis of their sequence homology, agonist selectivity and signal transduction pathway. Group I contains mGlu1 and mGlu5 subtypes, which are coupled to phospholipase C in transfected cells, and have quisqualic acid (Quis) as their most potent agonist. The second group consists of mGlu2 and mGlu3 receptors, which couple negatively to adenylyl cyclase in transfected cells and for which l-2-(carboxycyclopropyl) glycine is the most potent agonist. Group III contains mGlu4, mGlu6, mGlu7 and mGlu8, which again couple negatively to adenylyl cyclase but have l-2-amino-4-phosphonobutyric acid as their most potent agonist. Several studies showed that different groups of the mGlu receptors exhibit differential neuronal targeting, with group II and III mGlu receptors being predominantly pre-synaptic, whereas the group I being post-synaptic (Baude et al. 1993; Ohishi et al. 1995; Shigemoto et al. 1996, 1997) and, furthermore, mGlu1 and mGlu5 show a precise and highly ordered synaptic location in an annulus that surrounds the post-synaptic density (PSD; Baude et al. 1993; Lujan et al. 1996, 1997). This specific location could be directed by protein–protein interactions with cytosolic proteins, or as a result of receptor association with specific plasma membrane domains, which have been shown to concentrate different signalling molecules as well as some neurotrasnmitter receptors (Bruses et al. 2001; Suzuki et al. 2001).
Caveolae are small invaginations of the plasma membrane that have been implicated in functions such as signal transduction, cholesterol transport and endocytosis (Anderson 1998; Okamoto et al. 1998; Kurzchalia and Parton 1999; Smart et al. 1999), and they are considered as a subtype of rafts, membrane microdomains that share in common a characteristic lipid composition but that can be differentiated as they present different solubility in different detergents (Röper et al. 2000) and different degrees of organization (Madore et al. 1999). Caveolae are present in most cell types but are specially abundant in adipocytes, endothelial cells, fibroblasts and smooth muscle cells (Severs 1998). Caveolins form the structural network of caveolae and are necessary for all the above functions. Recent studies have reported the presence of at least three mammalian subtypes of caveolin which show similarities in structure and function but with differential tissue distribution. Caveolin-1 and caveolin-2 show a similar tissue distribution, whereas caveolin-3 shows a muscle-specific tissue distribution and may replace caveolin-1 as the major caveolar protein in differentiated muscle cells (Scherer et al. 1996; Way and Parton 1996). It has also been shown the presence of caveolin-1 and caveolin-2 in brain (Galbiati et al. 1998), where they have been located at high levels of expression in growth cones of dorsal root ganglion neurones what suggest that they could have a role in neurite outgrowth.
Here we report an association of metabotropic glutamate receptor type 1α with low-density caveolin-rich plasma membrane fractions and its molecular interaction with caveolin-1 and caveolin-2. Furthermore, it is shown that mGlu1α receptors mediates signalling through the extracellular signal-regulated kinase (ERK)1/2 mitogen-activated protein kinase (MAPK) pathway in these low-density caveolin-rich plasma membranes.
Antibodies and cell culture
The primary antibodies used for immunolabelling were: affinity-purified antimGlu1 polyclonal antibody F1-Ab (pan-mGlu1; Ciruela and McIlhinney 1997), affinity-purified antimGlu1α polyclonal antibody F2-Ab (Ciruela et al. 1999), rabbit anticaveolin-1 (Transduction Laboratories, Lexington, KY, USA), mouse anticaveolin-1 (Zymed Laboratories, Inc., San Francisco, CA, USA), mouse anticaveolin-2 (Clone 65; Transduction Laboratories), mouse antitransferrin receptor (Clone H68.4, Zymed), mouse antiβ-tubulin (Clone TUB 2.1, Sigma Chemical Co., St Louis, MO, USA), rabbit anti-ERK1/2 (M-5670, Sigma Chemical Co.) and mouse antiphosphorylated ERK1/2 (M-8159, Sigma Chemical Co.). The secondary antibodies used were: horseradish-peroxidase (HRP)-conjugate swine anti-rabbit IgG and HRP-conjugate swine anti-mouse IgG (Dako, Glostrup, Denmark), AlexaFluor488-conjugated goat anti-mouse IgG antibody and TexasRed-conjugated goat anti-rabbit IgG antibody (Molecular Probes, Leiden, the Netherlands).
BHK-1α cells permanently transfected with the mGlu1α receptor, kindly given by ZymoGenetics (Seattle, WA, USA) were grown in Dulbecco's modified Eagle medium (DMEM; Whittaker, Walkersville, NY, USA) supplemented with 1 mm sodium pyruvate, 2 mm l-glutamine, 100 U/mL penicillin/streptomycin, 10% (v/v) fetal bovine serum (FBS) and 0.5 mg/mL of G418 sulfate (Gibco, Grand Island, NY, USA) at 37°C and in an atmosphere of 5% CO2 (Ciruela and McIlhinney 1997).
For immunocytochemistry, BHK-1α cells were fixed in 4% paraformaldehyde for 15 min and washed with phosphate-buffered saline (PBS) containing 20 mm glycine (buffer A) to quench the aldehyde groups. Then, cells were permeabilized with buffer A containing 0.2% Triton X-100 for 5 min. Cells were treated with PBS containing 1% bovine serum albumin (BSA, buffer B). After 1 h at room temperature, cells were labelled with the indicated primary antibody for 1 h, washed and stained with the indicated secondary antibody. Coverslips were rinsed for 30 min in buffer B and mounted with Immuno Floure mounting medium (ICN Biomedical Inc., Costa Mesa, CA, USA). Confocal microscope observations were made with a LeicaTCS-SP (Leica Lasertechnik GmbH, Heidelberg, Germany) confocal scanning laser microscope adapted to an inverted Leitz DMIRBE microscope.
Immunogold localization of mGlu1α
BHK-1α cells grown as described above but in T25 flask were washed with PBS and fixed at room temperature in 0.1% glutaraldehyde and 2% formaldehyde in 0.1 phosphate buffer pH 7.2 during 15 min. After washing, the cells were scraped off the flasks, pelleted and embedded in 7.5% gelatin in PBS. The pellets were incubated on ice with first 2.1 m sucrose and then 2.3 m sucrose, and frozen in liquid nitrogen. Ultrathin sections were cut on a Reichert Ultracut S microtome (Leica, Glostrup, Denmark), collected with 2.3 m sucrose, and mounted on Formvar-coated copper or nickel grids. F2-Ab and monoclonal anticaveolin-2 Ab, followed by 20-nm protein A-gold and 5-nm protein G-gold were used to detect mGlu1α receptor and endogenous caveolin-2, respectively. Sections were analysed in a Philips 100 Cm electron microscope (Philips, Eindhoven, the Netherlands).
Confluent BHK-1α cells were washed with ice-cold PBS, scraped and placed into 2 mL of 500 mm sodium carbonate, pH 11.0. Homogenization was carried out sequentially using a loose-fitting Dounce homogenizer (10 strokes), a Polytron tissue grinder (three 10-s bursts; Kinematica GmbH, Brinkmann Instruments, Westbury, NY, USA), and a sonicator (three 20-s bursts; Branson Sonifier 250, Branson Ultrasonic Corp., Danbury, CT, USA). The homogenate was then adjusted to 45% sucrose by addition of 2 mL of 90% sucrose prepared in MBS (25 mm Mes, 0.15 m NaCl, pH 6.5) and placed at the bottom of an ultracentrifuge tube. A discontinuous 5–35% sucrose gradient was formed above (4 mL of 5% sucrose/4 mL of 35% sucrose; both in MBS containing 250 mm sodium carbonate) and centrifuged at 105 000 g for 16–20 h in a SW41.Ti rotor (Beckman Instruments, Palo Alto, CA, USA). A light-scattering band confined to the 15–20% sucrose was observed. Gradient fractions of 1 mL were collected and analysed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS–PAGE) and western blot (see below) was performed using antibodies against mGlu1α receptor (F2-Ab, 5 µg/mL), caveolin-1 (polyclonal Ab, 1/5000), caveolin-2 (1/250), tubulin (1/200) and transferrin receptor (1 µg/mL).
Immunoprecipitation and western blot
BHK-1α cells and low-density caveolin-enriched membrane fractions (CEMF) obtained from BHK-1α cells were solubilized in ice-cold lysis buffer [Tris-buffered saline (TBS), pH 7.4, containing 0.5% (v/v) Nonidet p-40, 0.5% (v/v) Triton X-100 and 60 mmn-octylglucoside] for 30 min on ice. The solubilized preparation was then centrifuged at 1 000 g for 20 min. The supernatant (1 mg/mL) was processed for immunoprecipitation, each step of which was conducted with constant rotation at 0–4°C. The supernatant was incubated with a mouse anticaveolin-1 (2 µg/mL), mouse anticaveolin-2 (2 µg/mL) or and irrelevant mouse antibody (2 µg/mL) for 2 h. Then 40 µL of a suspension of protein G cross-linked to agarose beads were added and the mixture incubated overnight. The beads were washed twice with ice-cold lysis buffer, twice with ice-cold lysis buffer containing 0.05% (v/v) Nonidet p-40, 0.05% (v/v) Triton X-100 and 6 mmn-octylglucoside), once with ice-cold TBS, pH 7.4, and aspirated to dryness with a 28-gauge needle. Subsequently, 30 µL of SDS–PAGE sample buffer was added to each sample. Immune complexes were dissociated by heating to 60°C for 10 min and resolved by SDS–PAGE using 6% and 12% acrylamide gels. Proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Immobilon-p, Millipore, Watford, UK) using a semidry transfer system and immunoblotted using the F2-Ab (5 µg/mL) and HRP-conjugated goat anti-rabbit IgG. The immunoreactive bands were developed with the SuperSignal chemiluminescent detection kit (Pierce, Rockford, IL, USA).
Phosphorylation of ERK
In vitro phosphorylation assays were carried out mainly as described by Suzuki et al. (2001). In brief, the light membrane fraction and a mixture of high-density fractions of a sucrose gradient was diluted with deionized water and centrifuged at 105 000 g for 1 h at 4°C in a 75Ti rotor. The pellet was resuspended in 1 mL of phosphorylation buffer [50 mm HEPES/KOH pH 7.4, containing 15 mm MgCl2, 1 mm dithiothreitol (DTT), protease inhibitor mixture (Sigma Chemical Co.), 1 mm NaVO4, 50 mm NaF and 500 µm adenosine triphosphate (ATP)], and divided into two 500-µL aliquots, one was treated with 100 µm quisqualate and the other with vehicle. Phosphorylation was carried out at 25°C for 5 min 50 µL of each reaction were resolved on 10% SDS–PAGE, blotted onto Immobilon-p membrane and incubated with rabbit anti-ERK1/2 (1/40 000) or mouse antiphospho-ERK1/2 (1/10 000). Quantitative analysis of detected bands was carried out by densitometric scanning.
Co-localization between mGlu1α receptor and caveolin
By double immunolabelling experiments in fixed and permeabilized BHK-1α cells, the subcellular distribution of mGlu1α receptor, caveolin-1 and caveolin-2 was studied. As observed in Fig. 1, a punctate staining pattern for the three proteins throughout the cell is shown. The overlap of the images resulted in a high degree of co-localization between mGlu1α receptor and caveolin-1 (Figs 1c and d) and between mGlu1α receptor and caveolin-2 (Figs 1h and i) showing the close spatial localization of mGlu1α receptor and both caveolins. The co-localization with caveolin-1 was predominantly observed in the plasma membrane (Fig. 1d, arrow), while the co-localization with caveolin-2 was observed both in the plasma membrane (Fig. 1i, arrow) and in intracellular compartments (Fig. 1i, arrow head).
Subcellular localization of mGlu1α and caveolin-2 in BHK-1α cells was examined by electron microscopic immunocytochemistry. Ultracryosections of BHK-1α cells labelled with F2-Ab followed by protein A-gold and with mouse anticaveolin-2 followed by protein G-gold (Figs 1e and j), showed co-localization of mGlu1α receptor with caveolin-2 (Figs 1e and j, arrowheads) both in intracellular compartments (Fig. 1e) and in the plasma membrane (Fig. 1j). These observations correlated well with those obtained by confocal microscopy and confirmed the close association of mGlu1α receptor with caveolins.
mGlu1α receptor associates with CEMF
To determine whether the mGlu1α receptor is present in CEMF, BHK-1α cells were homogenized and fractionated using the non-detergent-based protocol of Smart et al. (1995) which separates caveolin-rich membrane from the bulk of cellular membranes and cytosolic proteins. Fractions were recovered from the top of the gradient and analysed by western blot using antibodies against mGlu1α, caveolin-1, caveolin-2, tubulin and transferrin receptor. This procedure permitted the isolation of one fraction clearly enriched in both caveolin-1 and caveolin-2 (Fig. 2a, lane 4) in the low-density region of the gradient. To discard possible contaminants from the high-density membrane fractions (HDMF) of the gradient, the blots were probed against tubulin and transferrin receptor and no reactivity was found. When those blots were probed against mGlu1α receptor, we showed that the receptor was also highly enriched in the same fraction (Fig. 2a, lane 4), confirming the association of an important amount of the mGlu1α receptor with CEMF in BHK-1α cells.
mGlu1α receptor co-immunoprecipitates with caveolin-1 and caveolin-2
Caveolin proteins are small structural proteins associated with the flask-shaped membrane pits named caveolae. Because mGlu1α receptor appears to be present in low-density membrane fractions (LDMF) and co-localizes with caveolins in the plasma membrana (see Fig. 1) we investigated whether the basis for this preferential distribution could be explained by a molecular association of the mGlu1α receptor with caveolins. To further characterize this possible interaction, we performed co-immunoprecipitation experiments in BHK-1α cells permanently transfected with mGlu1α receptor (Fig. 3). Cells were solubilized and the extracts expressing mGlu1α (Fig. 3a, lane 1) were immunoprecipitated with an irrelevant antibody (Fig. 3a, lane 2), mouse anticaveolin-1 (Fig. 3a, lane 3) and mouse anticaveolin-2 (Fig. 3a, lane 4), and subsequently the immunoprecipitates were analysed by SDS–PAGE and immunoblotted using F2-Ab, an antibody against mGlu1α receptor. An immunoreactive band with an identical migration as in the lysate (Fig. 3a, lane 1) corresponding to the mGlu1α receptor could be detected in the immunoprecipitate with anticaveolin-1 (Fig. 3a, lane 3) and with anticaveolin-2 (Fig. 3a, lane 4), but could not be detected when an irrelevant antibody was used (Fig. 3a, lane 2). Moreover, co-immunoprecipitation experiments were also performed in CEMF isolated from BHK-1α cells. The immunoprecipitates were analysed by SDS–PAGE and immunoblotted using F2-Ab. An immunoreactive band corresponding to the mGlu1α receptor in the lysate (Fig. 3b, lane 1) and in the immunoprecipitate with anticaveolin-1 (Fig. 3b, lane 4) could be detected, but not in the immunoprecipitate of any of the negative controls (Fig. 3b, lanes 2 and 3). The intensities of the immunoreactive bands on X-ray film corresponding to mGlu1α protein were measured by densitometric scanning. The relative densitometric scanning values revealed an enrichement of six times in the mGlu1α co-immunoprecipitates obtained in CEMF (Fig. 3b) when compared with the ones performed in total membrane extract (Fig. 3a). These results clearly confirm an association between mGlu1α receptor and caveolins, an interaction that might explain the location of mGlu1α receptor in low-density CEMF.
Signalling of mGlu1α receptor within CEMF
It has been previously described that stimulation of mGlu1 receptors involves the activation of the MAPK ERK via protein kinase C (PKC; Calabresi et al. 2001). The functional integrity of mGlu1α receptor found in the light membrane fraction was tested, using an in vitro phosphorylation assay. After isolating this fraction (Fig. 2, lane 4) and adding 100 µm quisqualate, it could be observed a marked phosphorylation of ERKs (Fig. 4, CEMF + Quis) when compared with the untreated sample (Fig. 4, CEMF – Quis). Interestingly, the fraction at the bottom of the sucrose gradient (Fig. 2, lanes 8 and 9), also containing mGlu1α receptor, did not showed ERKs phosphorylation after Quis challenge (Fig. 4, HDMF). These results show that mGlu1α receptor and the molecules involved in its signal transduction are located in those light membrane fractions and that the signalling cascade is properly coupled to be fully functional.
The results presented here show that mGlu1α receptor co-fractionates with caveolin-1 and caveolin-2 in a low-density fraction obtained by ultracentrifugation in a sucrose gradient in permanently transfected BHK-1α cells, interacts with caveolae proteins and co-localizes with caveolins by immunogold and confocal fluorescence microscopy.
Several transmembrane receptors have been associated with CEMF, membrane microdomains (Kifor et al. 1998; Bruses et al. 2001; Gines et al. 2001; Marmor and Julius 2001; Suzuki et al. 2001) that have been shown to accumulate a large number of signalling molecules, such as trimeric G proteins, Ras, PKC and kinases of the src family (Harder and Simons 1997; Simons and Ikonen 1997). It has been suggested that CEMF can recruit a specific set of proteins and exclude all others (Simons and Ikonen 1997). This specific recruitment would help in the association of particular molecules, for example transmembrane receptors and their signalling effectors. Here we show that mGlu1α receptor associates with CEMF and interacts with caveolin-1 and -2. More important is the fact that only the mGlu1α receptor in CEMF signals through the ERK1/2 kinase pathway, suggesting that the underneath signalling machinery is functionally integrated in the same light membrane domains where mGlu1α receptor is found, as it has already been described for the platelet derived growth factor receptor (PDGFR), whose signalling Ras-ERK module has been localized to caveolae (Liu et al. 1997) or, even more recently, for the vascular endothelial growth factor receptor-2 (VEGFR-2; Labrecque et al. 2003). Therefore, CEMF could constitute the proper environment for mGlu1α receptor signalling by concentrating both receptors and signalling molecules.
Caveolins posses a remarkable capacity to bind cholesterol as well as a whole series of proteins involved in signal transduction pathways, such as trimeric G proteins subunits, Src kinases and Raf. Recently, the Drosophila melanogaster metabotropic glutamate receptor (DmGluRA), expressed in photoreceptor cells, was purified to homogeneity and reconstituted into liposomes of varying composition. Interestingly, glutamate binding was strictly dependent on the presence of ergosterol, a sterol analogue whose structure closely resembles that of cholesterol (Eroglu et al. 2002). Here we demonstrated a CEMF dependence of mGlu receptor function which could be related with the presence of cholesterol and their analogues in these membrane fractions.
Recently, mGlu1α receptor has been shown to heterodimerize with the calcium-sensing receptor (CaR; Gama et al. 2001) and with adenosine A1 receptor (A1R; Ciruela et al. 2001). Interestingly, both CaR and A1R associate with caveolin-rich plasma membrane domains (Kifor et al. 1998; Gines et al. 2001), suggesting that caveolae may act as signalling platforms where these receptors could concentrate and, therefore, their interaction be facilitated. It should be noted that Becher et al. (2001) reported that mGlu1α receptor did not associate with lipid rafts resistant to Triton X-100 extraction in rat cerebellum. This apparent contradictory result could be explained taking into account that Triton X-100-based extraction procedures have been found to extract some of the proteins associated with rafts (Yamabhai and Anderson 2002), and different detergents seem to extract different rafts (Madore et al. 1999; Röper et al. 2000). This is the reason we selected the non-detergent method reported by Smart et al. (1995) in our studies. On the other hand, it has been demonstrated that post-synaptic mGlu1 has a perisynaptic localization in neurones (Baude et al. 1993; Lujan et al. 1996, 1997) and this specific localization could be due to its association with these membrane microdomains capable of organizing different groups of proteins in different membrane locations or it might be due to its molecular association with different proteins of the PSD, such as Homer (Ciruela et al. 1999, 2000), linking the receptor to the scaffolding multimeric signalling protein Shank (Sheng and Kim 2000), or most probably by a combination of both possible mechanisms.
In different cell types, especially polarized cells, it has been shown that two different post-Golgi circuits exist for membrane trafficking, one associated with the apical membrane and the other with the basolateral membrane. It has also been shown that neurones have an equivalent apical pathway for delivery of proteins and lipids to axons, and a basolateral similar pathway for delivery to dendrites (Simons and Ikonen 1997; Ikonen 2001). Moreover, in MDCK cells, it was shown that although both homooligomers of caveolin-1 and hetero-oligomers of caveolin-1 and caveolin-2 are both formed in the endoplasmic reticulum, they are distinctly targeted to the cell membrane, the first being apically sorted, whereas the second is associated with the basolateral membrane (Scheiffele et al. 1998). Our biochemical data show that mGlu1α receptor associates with both caveolin-1 and caveolin-2, so we suggest that the predominantly dendritic (post-synaptic) localization of this receptor could be explained by its association with caveolin-1 and caveolin-2 hetero-oligomers, a caveolin complex that would be targeted to the basolateral membrane (dendrites in neurones) as shown in MDCK cells.
In conclusion, our data presented here demonstrate the presence of mGlu1α receptors in CEMF and their molecular association with caveolin proteins. Moreover, they also show a functional association of the mGlu1α receptor with CEMF in BHK1α cells, linking the receptor activation to the MAPK cascade. Given the key role of the MAPK cascade and immediate–early genes in the coupling of early neuronal responses to long-term adaptive changes (Sweatt 2001), the present results suggest that the specific localization of mGlu1α receptor in CEMF could play a critical role in the mGlu1α receptor induction of the long-term potentiation (Calabresi et al. 2001).
FC is currently holding a Ramón y Cajal research contract signed with the Ministerio de Ciencia y Tecnología. This study was supported by grants from Ministerio de Ciencia y Tecnología (BIO99-0601-C02-02, SAF2002-03293 to RF and SAF2001-3474 to EC), Fundació La Marató de TV3 (Marató/2001/012710 to RF) and European Community (QLRT-2000–01056 to RF). We are grateful to the personnel from Serveis Científic i Tècnics de la Universitat de Barcelona for their excellent technical assistance in confocal microscopy.