The C-terminal Tail of the Metabotropic Glutamate Receptor Subtype 7 Is Necessary but Not Sufficient for Cell Surface Delivery and Polarized Targeting in Neurons and Epithelia*

Complex neuronal functions rely upon the precise sorting, targeting, and restriction of receptors to specific synaptic microdomains. Little is known, however, of the molecular signals responsible for mediating these selective distributions. Here we report that metabotropic glutamate receptor subtype 7a (mGluR7a) is polarized at the basolateral surface when expressed in Madin-Darby canine kidney (MDCK) epithelial cells but is not polarized when expressed in cultured hippocampal neurons. Truncation of the mGluR7 cytoplasmic tail produces a protein that is restricted to a perinuclear intracellular compartment in both neurons and MDCK cells, where this protein colocalizes with a trans-Golgi network antigen. The mGluR7 cytoplasmic domain appended to the transmembrane portion of the vesicular stomatitis virus G protein and the ectodomain of human placental alkaline phosphatase is distributed over the entire cell surface in cultured neurons. When expressed in MDCK cells, this construct remains in an intracellular compartment distinct from endosomes or lysosomes. Thus, the cytoplasmic tail domain of mGluR7 is necessary but not sufficient for polarized targeting in MDCK monolayers, whereas in neurons the cytoplasmic tail is sufficient for cell surface expression but not polarization. Additional mechanisms are likely required to mediate mGluR7 neuronal polarization and synaptic clustering.

Specialized domains at the neuronal cell surface orchestrate complex synaptic functions. Accurate sorting, trafficking, and restriction of proteins to these microdomains ensures the precise distribution of synaptic proteins for proper neuronal function. Proteins destined for secretion or incorporation into the plasma membrane transit to the cell surface through a series of subcellular compartments. Synthesis occurs in the soma at the rough endoplasmic reticulum (RER), 1 and postsynthetic proc-essing continues in the Golgi complex, with sorting of proteins into specific transport vesicles occurring in the TGN (1,2). Information contained within targeted proteins function as sorting signals to specify their ultimate destination within the cell (3,4). Although progress is being made in the categorization of sorting signals, their diversity and complexity, as well as the underlying mechanisms through which cells carry out polarized targeting, remain unresolved (1).
Neurotransmitter receptors are among the complex array of proteins at synaptic microdomains that participate in the generation and propagation of neuronal activity. Glutamate serves as the principle excitatory amino acid neurotransmitter at central nervous system synapses. Activity at glutaminergic synapses is mediated by both ionotropic and metabotropic glutamate receptors (5). Metabotropic glutamate receptors couple with heterotrimeric G-proteins to regulate cell excitability and synaptic transmission (6,7) and have been implicated in synaptic plasticity (8). Currently eight distinct mGluR subtypes have been recognized (6). These have been observed to exhibit subtype-and region-specific differences in distribution along the neuronal cell surface in vivo. The mGluR7 subtype, in particular, highlights the complexity of protein targeting. mGluR7 receptors have been reported to be targeted to postsynaptic sites in the olfactory bulb (9) and locus coeruleus (10,11), to axons in the hippocampus (12,13), to both pre-and postsynaptic sites in the retina (14), and to both axons and dendrites of cultured hippocampal neurons (15). In the hippocampal formation additional complexity is evident, because the distribution of mGluR7 is apparently dependent upon the nature of the postsynaptic target neuron. Within individual axons, mGluR7 is found to be present only in distinct subpopulations of presynaptic terminals (16). The subtype-and regionspecific placement of mGluR7 at distinct membrane domains requires the existence of mechanisms whereby sorting information is interpreted, so that receptors can be distributed precisely. mGluR7 is a member of the larger seven transmembrane or G-protein-coupled receptor family of receptors. Little is known about the cellular mechanisms that govern the intracellular trafficking of the individual members of this critically important protein family.
Because of the technical difficulties inherent in studying sorting and trafficking in neurons, cultured cell lines have been utilized as a means both to study sorting in a polarized cell system and to allow biochemical investigations of the underlying cellular mechanisms. Among these culture models, epithelial cells have been especially useful in studies of neuronal membrane protein targeting (17). This utility may be in part based upon the developmental origin of neurons from neuroepithelial stem cells (18). In some but not all instances, Madin-Darby canine kidney epithelial cells (MDCK) and neurons seem to employ similar signals and mechanisms to generate polarized protein distributions (as discussed in Refs. 4, 15, and 19). Many neuronal proteins found in axons can sort to the apical surface of epithelia, whereas most dendritic proteins can sort to the basolateral surface. We report here an analysis of the subcellular targeting of the mGluR7 receptor. By coordinated investigation in both epithelial cell lines and in cultured hippocampal neurons, we are able to examine whether the requirements for the polarized sorting of the protein are the same in both cell types. We report that mGluR7 achieves a polarized steady-state distribution at the basolateral membrane when expressed in epithelial monolayers. The cytoplasmic tail domain of mGluR7 is necessary but not sufficient for this polarized targeting in MDCK cells. In neurons however, exogenously expressed mGluR7 does not achieve a polarized distribution. The cytoplasmic tail is not sufficient for polarization, but without it the protein does not achieve a cell surface distribution.

EXPERIMENTAL PROCEDURES
Construction of Fusion Proteins-Cloned mGluR7a cDNA in the pBluescript plasmid was a kind gift from S. Nakanishi (20). An XbaI site was introduced at the Bluescript polylinker NotI site by insertion of an oligonucleotide linker. This XbaI site was used along with the pBluescript polylinker ClaI restriction site and the AatII restriction site within mGluR7a for subcloning and fusion protein construction. Fusion PCR was used to construct the GFP⅐mGluR7 fusion protein. PCR was used to generate three products from which the final fusion products were made by use of the same primers. The first PCR product was generated from the mGluR7 wt cDNA using a 5Ј-primer complimentary to and including the pBluescript polylinker ClaI restriction site and a 3Ј-primer that included both the last five codons before the intended site of GFP incorporation (10 amino acids following the signal sequence) and the first seven codons of the GFP sequence. The GFP sequence was generated in a similar manner from pEGFP-C1 (CLONTECH, Palo Alto Ca.) using primers that included both the mGluR7 sequence before the site of intended insertion and the sequence on either end of the GFP encoding sequence. The downstream mGluR7 product was generated with a 5Ј-primer including both the final five codons of GFP and the initial six codons following the insertion site and a 3Ј-primer past the unique AatII restriction site in mGluR7. The first two PCR products were combined and allowed to prime off each other, outside primers were added, and the fusion product was generated. This initial fusion product was then used in a similar manner with the third PCR product from the initial set to produce the complete fusion PCR product. The resulting fragment was digested at the unique ClaI and AatII restriction sites for replacement into the original cDNA plasmid of mGluR7. The GFP⅐mGluR7-tail deletion construct was generated by PCR with a 5Ј-primer at the interface of the GFP and mGluR7 fusion and a 3Јprimer selective to a sequence following the last transmembrane domain. The primer was designed to incorporate a stop codon 11 amino acids following the final transmembrane domain and an XbaI restriction site. The resulting fragment was digested at the unique XbaI and AatII restriction sites for replacement into the cDNA plasmid of GFP⅐mGluR7. The PLAP⅐mGluR7-tail fusion construct was generated using a PLAP⅐VSVG plasmid (21). A PCR fragment containing the cytoplasmic tail of mGluR7 was generated using a 5Ј-primer containing following the VSVG transmembrane domain up to and including a FIG. 1. Metabotropic glutamate receptor constructs. The predicted membrane topologies and relevant sequence domains are illustrated for each mGluR7 cDNA construct. mGluR7 wt contains the entire wild-type receptor. The GFP⅐mGluR7 fusion protein (GFP-mGluR7) has an extracellular GFP incorporated inframe at a position 10-amino acids distal to the signal sequence. PLAP-mGluR7 tail has the mGluR7 cytoplasmic tail incorporated inframe with a single spanning membrane protein containing the extracellular placental alkaline phosphatase domain and the VSVG transmembrane domain. GFP-mGluR7 tail minus represents the GFP fusion receptor protein with its cytoplasmic tail deleted. HindIII restriction site, and a 3Ј-primer to incorporate an XbaI restriction site following the stop codon. The resulting fragment was digested at the XbaI and HindIII restriction sites for replacement into the cDNA plasmid of PLAP⅐VSVG. All constructs were sequenced through the respective ligation points. The constructs were then subcloned into the mammalian expression vector pCB6 (kindly provided by M. Roth, University of Texas, Southwestern) between the ClaI and XbaI sites prior to transfection.
COS-1 cells were transiently transfected using DEAE-dextran (22). Cells (2 ϫ 10 5 per 35-mm dish) were plated on glass coverslips the night before transfection. Cells were washed in Tris-buffered saline and exposed to 5 g of DNA, and 0.5 mg/ml DEAE-dextran (Amersham Pharmacia Biotech) was added for 40 min at 37°C. The solution was discarded. and the cells were incubated in 1 ml of DMEM containing 10% fetal bovine serum and 0.1 mM chloroquine for 3 h at 37°C. The cells were washed in Tris-buffered saline and incubated in complete DMEM for 48 -72 h prior to immunofluorescent immunohistochemistry. Subconfluent MDCK cells were prepared as previously described (22) and transfected by the PerFect lipid method (Invitrogen, San Diego, Ca.), with fusion constructs that had been subcloned into the pCB6 vector. This vector carries resistance to the antibiotic G418 (Life Technologies, Inc.). After selection in 1.8 mg/ml G418, clones were screened for expression by immunofluorescence, Western blotting, or phosphatase assay (BluePhos Microwell, Kirkegaard & Perry, Gaithersburg, MD).
Primary hippocampal cultures were prepared as described (19). Cultured neurons were transfected at stage 5, a stage when synapses have been formed and endogenous synaptic receptors have polarized and localized to the synapses. Cultured hippocampal neurons were transfected at 7 days in vitro (DIV) by the LipofectAMINE 2000 method (Life Technologies). In brief, 1 g of DNA was diluted into 50 l of serum-free medium and incubated at room temperature for 5 min. Then, 3 l of LipofectAMINE 2000 reagent was diluted into 50 l of serum-free medium in a second tube. The diluted DNA and the diluted Lipo-fectAMINE 2000 reagent were combined, mixed, and incubated for 15 min at room temperature. Neurons were transferred to a new 6-well plate that contained 1 ml of the neuronal maintenance medium from the hippocampal culture. The DNA-LipofectAMINE 2000 reagent complexes were mixed into the medium bathing the cells and incubated at 37°C with 5% CO 2 for 2 h. The coverslips were then placed back into the original 6-well plate and incubated for 2 days.
Immunofluorescent Microscopy-Immunofluorescence was performed as described (4) on stably expressing cell lines. Briefly, cells were grown to confluence on Transwell porous cell culture inserts (Corning Costar Corp., Cambridge, MA) or, for screening clones, on 8-well Lab-Tek slides (Nalge Nunc International, Naperville, IL) and fixed in Ϫ20°C methanol for 7 min at room temperature. Cells were then permeabilized in a phosphate-buffered saline-based wash buffer containing 0.3% Triton X-100 and 0.1% bovine serum albumin for 15 min. Nonspecific binding of antibody was blocked by incubating the cells in goat serum dilution buffer (16% filtered goat serum, 0.3% Triton X-100, 20 mM NaP i , pH 7.4, 0.9% NaCl) at room temperature for 1 h. c, d, f, g, i, j, l) and confocal xz cross-section (b, e, h, k) immunofluorescence images of MDCK cells were obtained. a-c illustrate an intracellular distribution of mGluR7 wt as determined by staining with anti-mGluR7-wt. c, the intracellular distribution of mGluR7 is highlighted (arrowheads). The GFP⅐mGluR7 receptor construct displays a mainly basolateral cell surface distribution (d-f) in MDCK cells as determined by native GFP fluorescence. f, minimal intracellular mGluR7 is observed (arrowheads). In contrast, however, immunofluorescence labeling of GFP⅐mGluR7 with the cytoplasmic tail-directed anti-mGluR7-wt antibody (g-i) shows a primarily intracellular distribution (i, arrowheads) as observed in c. Immunofluorescence labeling of GFP⅐mGluR7 with anti-GFP (j-l) illuminates both extracellular and additional intracellular labeling as compared with (d-f). The intracellular mGluR7 is highlighted (l, arrowheads).

FIG. 4.
Cellular distribution of PLAP⅐mGluR7 and GFP⅐mGluR7-tail-minus receptor constructs in polarized MDCK cells as determined by immunofluorescence. MDCK cells were stably transfected with PLAP⅐mGluR7-tail and GFP⅐mGluR7-tail-minus in the pCB6 mammalian expression vector. Confocal en face immunofluorescence images of MDCK cells are presented. In a and b, immunofluorescence of PLAP⅐mGluR7-tail shows the construct to be contained in a punctate intracellular compartment as determined by staining with anti-PLAP. In c and d, immunofluorescence of GFP⅐m-GluR7-tail-minus illustrates the construct to be contained in a different intracellular compartment, which shows a perinuclear distribution similar to that seen in Fig. 2. temperature. Between primary and secondary antibody incubations, the cells were subjected to three 5-min washes in the phosphate-buffered saline-based wash buffer. After incubation with the secondary antibody, cells were washed in phosphate-buffered saline three times for 5 min each and finally in 10 mM NaP i for 10 min before being mounted on coverslips with Vectashield (Vector Laboratories, Burlingame, CA). Confocal sections were taken using a Zeiss LSM 410 laser scanning confocal microscope. Images are the product of 8-fold line averaging. Xz cross sections were generated with a 0.2 micron motor step. Contrast and brightness were set so that all pixels were in the linear range. Immunofluorescence staining of neurons was performed as described (19). Immunofluorescent images were captured with IP Labs on a Zeiss Axiophot epiflurescence photomicroscope. Fluorescent dextran uptake was performed as described (23), cells were incubated for various times with DMEMϩ containing 1 mg/ml rhodamine-dextran (lysine fixable) (Molecular Probes) M r ϭ 10,000 in both upper and lower Transwell filter chambers.
Cell Surface Biotinylation-Cells were grown on 24-mm Transwell inserts (0.4-m pore, Costar) for 1 week, and the medium was replaced with complete MEM (without Geneticin) before the experiment. Steadystate biotinylation of either apical or basolateral proteins was carried out separately at pH 9.0 as described (22). Cells were incubated with N-hydroxysulfosuccinimide (NHS)-biotin for 2ϫ 20 min. Following biotinylation, filters were excised from the cups with a razor blade, and the attached monolayers were lysed in 1% Triton X-100 in 150 mM NaCl, 5 mM EDTA, 50 mM Tris, pH 7.5. The biotinylated proteins (either apical or basolateral cell surface proteins) were recovered from cell lysates by incubation with 100 l of packed immobilized streptavidinagarose beads (Pierce). Bound proteins were eluted from the beads in Laemmli sample buffer, analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting as described previously (22). The blot was incubated with anti-GFP or anti-PLAP (1:200) primary antibodies followed by horse radish peroxidase-conjugated secondary antibodies (Sigma) and developed with enhanced chemiluminescence reagent (Amersham Pharmacia Biotech).

Metabotropic Glutamate Receptor Constructs: Substitution, Deletion, and GFP Fusion Protein
Constructs-mGluR7 is a 7-transmembrane domain receptor with a large extracellular domain and a 60-amino acid cytoplasmic tail. The construct designated R7 wt (Fig. 1, mGluR7 wt) was investigated by use of a specific antiserum (11) directed against the C-terminal cytoplasmic tail of mGluR7. For subsequent investigations we chose to epitope-tag the receptor by creating GFP⅐mGluR7 fusion proteins. The GFP⅐R7 fusion protein construct has GFP incorporated in frame at a position 10 amino acids distal to the signal sequence (Fig. 1, GFP-mGluR7). For investigations of the role of the cytoplasmic tail in receptor localization, the cytoplasmic tail was deleted from the GFP⅐R7 fusion protein at a position 11 amino acids distal to the predicted end of the final transmembrane domain creating the GFP⅐R7-tail-minus construct ( Fig. 1, GFP-mGluR7 tail minus). To investigate whether the cytoplasmic tail of mGluR7 may function autonomously to direct localization, a fusion protein PLAP⅐R7-tail ( Fig. 1, PLAP-mGluR7 tail) was generated with the mGluR7 cytoplasmic tail incorporated inframe with a construct containing the VSVG transmembrane domain and placental alkaline phosphatase as the extracellular domain.
R7 wt and GFP⅐R7 Fusion Protein Distribution in Polarized and Nonpolarized Cells-Constructs were transiently expressed in COS-1 cells. Both R7 wt (Fig. 2a) and GFP⅐R7 (Fig.  2b) attain a cell surface distribution in COS-1 cells. Stably transfected MDCK cells also expressed mGluR7 at the plasma membrane. MDCK cell lines expressing mGluR7 were grown to confluence and examined by indirect immunofluorescence confocal microscopy (Figs. 3 and 4) using the anti-mGluR7 antiserum for R7 wt and either native GFP fluorescence or anti-GFP for GFP⅐R7. GFP⅐R7 displayed a polarized steady-state distribution at the basolateral surface of MDCK epithelial cells (Fig. 3, d, e, and f). To assess biochemically the polarized distributions of the mGluR7 proteins, we employed the cell surface biotinylation technique to differentially isolate MDCK apical and basolateral proteins. Steady-state cell surface biotinylation supported the immunofluorescent observation, by confirming GFP⅐R7 distribution at the basolateral surface (Fig.  5A).
To examine how neurons might sort and target GFP⅐R7, we determined whether the construct is appropriately targeted and polarized at steady state in cultured neurons. Use of the GFP fusion construct allowed detection of only exogenously expressed mGluR7. Whereas the majority of the immunofluorescence signal for GFP⅐R7 was intracellular or perinuclear, GFP⅐R7 was found at the cell surface throughout both axons and dendrites and with no obvious clustering. Exogenously expressed mGluR7 did not achieve a polarized distribution in cultured hippocampal neurons and was found at apparently equal intensity at the cell surfaces of both axons and dendrites (see Fig. 8, a and b).
Masking of the Cytoplasmic Tail of mGluR7 in MDCK Cells-Comparison of R7 wt and GFP⅐R7 fusion protein distributions suggests that the C-terminal tails of these proteins participate in a macromolecular complex. In MDCK cells stably expressing R7 wt, the protein distribution appears to be intracellular (Fig. 3, a-c) when immunolabeling is performed using an antisera directed against the final 18 amino acids of the mGluR7 carboxyl terminus. Importantly, when examined using the same C-terminal-directed antisera, the distribution of GFP⅐R7 also appears to be intracellular (Fig. 3, g-i). In con-  (lanes 3, 4). Only a weak signal is detected at the apical cell surface (lanes 1, 2). Lane 5 illustrates the absence of the construct at the surface of untransfected MDCK cells. The GFP⅐mGluR7-tail-minus construct was not detected at either the apical (lanes 6, 7) or the basolateral (lanes 8,9) cell surface. B, lane 1, consistent with its intracellular distribution (Fig. 4), cell surface biotinylation analysis did not detect the PLAP⅐mGluR7-tail construct at either the apical (lanes 2, 3) or basolateral (lanes 4, 5) cell surfaces. To ensure that the transfected MDCK cells express the PLAP⅐mGluR7-tail protein, Western blot analysis was performed on total cell lysates, revealing the presence of a protein of ϳ65 kDa corresponding to the construct.
trast, observation of the GFP⅐R7 intrinsic GFP signal in the same field (Fig. 3, d-f), together with cell surface biotinylation (see Fig. 5A), clearly illustrates a cell surface distribution (at the basolateral plasmalema, Fig. 3, d-f). These data suggest that in MDCK cells, the C-terminal portion of cell surface mGluR7 is interacting with a protein or proteins that interfere sterically with access of the C-terminal antibody to its epitope. According to this hypothesis, the C termini of mGluR7 intracellular populations do not participate in these interactions, thus permitting their detection in immunohistochemical experiments employing the C-terminal antibody. Observation of the GFP⅐R7 cell line with anti-GFP is consistent with this conclusion, revealing intracellular labeling (Fig. 3, j-l) in addition to the population detected at the cell surface. It would appear, therefore, that incorporation of mGluR7 into the basolateral plasma membrane of MDCK cells is most likely associated with the formation of protein-protein interactions involving its Cterminal tail.
Importance of the mGluR7 Cytoplasmic Tail for Sorting-Because the mGluR7 C-terminal tail appears to participate in protein-protein interactions at the MDCK cell basolateral membrane, we wondered whether the tail embodies information needed to establish protein localization. To test this pos-sibility, we examined the distributions of the GFP⅐R7-tail-minus and the PLAP⅐R7-tail constructs.
Following transient expression in COS-1 cells, GFP⅐R7-tailminus localized to perinuclear intracellular compartments whose morphology and distribution are reminiscent of the Golgi complex (Fig. 2d). When these constructs are expressed in MDCK cells, the GFP⅐R7-tail-minus construct recapitulated its COS cell distribution and was found in a perinuclear intracellular compartment (Fig. 4, c and d). To further identify the intracellular compartment in which the GFP⅐R7-tail-minus construct accumulates; we performed double labeling with an antibody specific to a Lupus auto-antigen expressed in the TGN (provided by A. Gonzalez; Ref. 24). In both COS-1 cells (Fig. 6, a-c) and in MDCK cells (Fig. 6, d-f and g-i) we found that GFP⅐R7-tail-minus colocalizes with the TGN marker. In addition, the Golgi population of this tail-minus construct does not seem to be able to cycle to the cell surface and return to the TGN via endosomes, as has been noted for TGN38 and furin. In the case of furin and TGN38, return to the TGN requires endosomal acidification and is inhibited by agents that raise endosomal pH (25). Incubation with chloroquine (100 micromolar, 4 h) did not alter GFP⅐R7-tail-minus distribution in MDCK cells (data not shown). The perinuclear Golgi-like distribution mGluR7 Targeting; C-terminus Is Necessary but Not Sufficient did not change and did not disperse to a characteristic punctate endosomal distribution during treatment. Thus, the tail-minus construct appears to be a fairly stable resident of the TGN that does not traffic from the cell surface via endosomes.
When transiently expressed in COS-1 cells, the PLAP⅐R7-tail construct attained a cell surface distribution (Fig. 2c). However, the PLAP⅐R7-tail construct did not attain a cell surface distribution in MDCK cells. Instead it was concentrated in a dispersed, punctate intracellular distribution (Fig. 4, a and b). This compartment is not labeled by antibodies to the lysosomal membrane protein LAMP-1 (Fig. 7, a-c). Subsequent experiments also indicated that it does not become labeled by internalized fluorescein isothiocyanate-dextran (Fig. 7, d-f). It would appear, therefore, that this compartment is neither lysosomal nor endosomal in origin. Steady-state cell surface biotinylation of MDCK monolayers supported immunofluorescent observations, confirming that neither PLAP⅐R7-tail nor GFP⅐R7-tail-minus were present at the cell surface (Fig. 5, A  and B).
Expression in neurons was performed to test how neurons would distribute the tail-containing and tail-minus constructs at steady state. Immunostaining with anti-PLAP and anti-GFP antibodies enabled detection of only exogenously expressed constructs. In contrast to results in MDCK cells, the immunofluorescence signal for PLAP indicated that PLAP⅐R7-tail was expressed at the neuronal cell surface throughout both axons and dendrites. It was not distributed in a polarized manner and exhibited no obvious clustering (Fig. 8, e and f). Once again, when exogenously expressed in neurons, GFP⅐R7-tail-minus was observed in a perinuclear intracellular distribution consistent with a Golgi-like pattern (Fig. 8c), with no observable label at the cell surface (Fig. 8d). DISCUSSION Neuronal function relies upon the polarized distribution of synaptic proteins. The mGluR7 subtype is a protein distributed with polarity in neurons, being found only in highly selective locations at cell surface synaptic sites. In different brain regions, mGluR7 can be found selectively distributed in either pre-or postsynaptic locations. In particular, mGluR7 distributes to selective subpopulations of axon terminals within individual neurons of the hippocampus. Recently, expression studies have examined the distribution of mGluR7 heterologously expressed in cultured hippocampal neurons (15). Results indicated that full-length mGluR7 was present at the cell surface of both axons and dendrites, suggesting that the exogenously expressed protein did not achieve a polarized distribution. The authors of this study proposed that the C terminus of mGluR7 contains an axonal targeting signal, because it permitted somatodendritic mGluR2 to enter axons.
The molecular nature of the mGluR7 sorting signal remains undefined. We wished therefore, to further examine the signals and mechanisms involved in mGluR7 sorting. Toward this end, we expressed mGluR7 and a variety of receptor constructs in neurons and epithelial cells. These constructs included fulllength mGluR7, a GFP fusion construct of full-length mGluR7, mGluR7 without its cytoplasmic tail, and a construct where the cytoplasmic tail of mGluR7 was appended to a single pass transmembrane protein. Consistent with the findings of Stowell and Craig (15), we find that in cultured hippocampal neurons, mGluR7 is expressed in a nonpolarized distribution at the surfaces of both somatodendritic and axonal domains. However in MDCK epithelia, exogenously expressed mGluR7 is polarized at the basolateral cell surface. Further experiments illustrated that the cytoplasmic tail of mGluR7 was necessary for surface delivery in both epithelia and in neurons but was not sufficient to mediate a polarized distribution in either setting.
It has been suggested that epithelial cells and neurons can use similar signals and mechanisms to sort membrane proteins to polarized distributions at the cell surface, because some neuronal proteins found in axons can sort to the apical surface of epithelia, whereas dendritic proteins can sort to the basolateral surface (22,26,27). In the case of mGluR7 this is clearly not the case. There must exist, therefore, mechanisms in neurons that interpret the mGluR7 sorting information differently from the mechanisms that interpret this information in MDCK cells. Similar discrepancies have been seen with a number of proteins, including norepinepherin and serotonin transporters (28,29).
We have demonstrated that the cytoplasmic tail of mGluR7 is clearly necessary for surface delivery, because the mGluR7tail-minus construct remained in a perinuclear intracellular pool and did not reach the cell surface in either MDCK cells or neurons. This compartment colocalizes with a trans-Golgi network antigen in both COS and MDCK cells. It would appear therefore, that the cytoplasmic tail-minus construct seems unable to exit the Golgi. The mechanism of this Golgi retention is not clear. We can state, however, that the protein does not appear to cycle between Glogi and cell surface via endosomes. This cycling behavior is observed for other proteins concentrated in the TGN, including furin and TGN 38. Membrane protein traffic from the TGN to the cell surface has been theorized to involve carbohydrate moieties as either targeting signals (30) or for modulation of the state of protein aggregation (31). In the case of mGluR7-tail-minus, glycosylation sites remain intact. Thus, the cause of retention may be more related to an inability to participate in protein-protein interactions important for stable surface expression. In addition, the cytoplasmic tail of mGluR7 is clearly not sufficient to mediate polarized sorting in epithelia, where the PLAP⅐mGluR7-tail construct remained within the cell in a cytoplasmic pool distinct from endosomal or lysosomal compartments. In neurons however, this construct was present at the surface of both axons and dendrites, as was found for exogenously expressed mGluR7, but did not achieve a polarized distribution. Thus, the capacity of mGluR7 constructs to attain cell surface expression is highly dependent upon the cell type in which they are expressed. The mGluR7 cytoplasmic tail is sufficient for cell surface delivery in neurons, but not in MDCK cells. Importantly, this portion of the protein is not by itself sufficient to specify polarized targeting in either cell type.
It is interesting to note that the cytoplasmic tails of a number of GPCRs interact with PDZ-containing cytoskeletal proteins (32). Cell surface machinery may be vital to polarized sorting (33). Interaction with cytoskeletal elements, particularly PDZcontaining proteins, have been implicated not only in cell surface interactions that retain proteins at specific locations (34), but also at earlier stages of protein targeting at the TGN (35). This is consistent with our observation that the cytoplasmic tail epitope is obscured in MDCK cells, where mGluR7 is distributed in a polarized fashion, but not in COS cells, which lack polarized membrane domains. Physical interaction of putative PDZ-containing cytoskeletal proteins with the C-terminal tail of mGluR7 could explain our inability to observe the construct at the basolateral surface with an antiserum directed to the cytoplasmic tail. In fact, the three final C-terminal residues of mGluR7 may represent a possible hydrophobic class II PDZ interaction motif (36). However, the absence of observable clustering or detergent insolubility for mGluR7 (data not shown) do not support a stable or continuous interaction with the MDCK cytoskeleton, as has been correlated with proper localization of some membrane proteins in MDCK (37,38).
In summary, the cytoplasmic tail of mGluR7 is clearly necessary for the surface delivery of the receptor but not sufficient to mediate its polarized distribution in epithelia or in neurons. Other domains of the mGluR7 molecule must be involved in contributing to sorting information. The nature of these domains, the signals they encode, and the interactions that interpret them remain to be determined. FIG. 8. Cellular distribution of mGluR7 fusion proteins in cultured hippocampal neurons determined by immunofluorescence. Cultured hippocampal neurons were transfected with mGluR7 constructs in the pCB6 mammalian expression vector, and the distributions of the exogenous proteins were determined by immunofluorescence. a and b illustrate the presence of GFP⅐mGluR7 at the surface of the axonal growth cone (a), and at the surface of both axons and dendrites (b) as detected using an anti-GFP antibody. GFP⅐mGluR7tail-minus was detected in an intracellular perinuclear distribution by native GFP fluorescence (c) and was not found at the cell surface by staining with anti-GFP (d). Immunofluorescence labeling of PLAP⅐mGluR7-tail with anti-PLAP antibody shows the construct to be present at the surface of both axons and dendrites (e, ϫ 40; f, ϫ 100).
mGluR7 Targeting; C-terminus Is Necessary but Not Sufficient