Agonist-induced internalization followed by subsequent return to the cell surface regulates G-protein-coupled receptor (GPCR) activity. Because the cellular responsiveness to ligand depends on the balance between receptor degradation and recycling, it is crucial to identify the molecules involved in GPCR recovery to the cell surface. In this study, we identify mechanisms involved in the recycling of the M4 subtype of muscarinic acetylcholine receptor. M4 is highly expressed in the CNS, plays a role in locomotor activity, and is a novel therapeutic target for neurologic and psychiatric disorders. Previous studies show that, after cholinergic stimulation, M4 internalizes from the cell surface to endosomes in cell culture and the rat brain. Here, we show that, after activation, M4 traffics to transferrin receptor- and Rab11a-positive perinuclear endosomes. Expression of the constitutively GDP-bound, inactive mutant Rab11aS25N inhibits M4 trafficking to recycling endosomes. Expression of the C-terminal tail of myosin Vb, a Rab11a effector, enhances M4 accumulation in perinuclear endosomes. Both Rab11aS25N and the myosin Vb tail impair M4 recycling. The results demonstrate that GPCR recycling is mediated through a discrete pathway using both Rab11a and myosin Vb.
- PC12 cells
- muscarinic acetylcholine receptors
- G-protein-coupled receptor
- myosin Vb
- unconventional myosins
- transferrin receptor
Regulated intracellular trafficking after stimulation controls activity and cell surface expression of neurotransmitter receptors, including G-protein-coupled receptors (GPCRs) and ion channels (Koenig and Edwardson, 1997; Lefkowitz, 1998;Carroll et al., 2001; Roche et al., 2001; St. John and Gordon, 2001). Agonist-induced GPCR internalization has been demonstrated in cell culture and in the rat and human brain (Bloch et al., 1999; Muriel et al., 1999). The balance between GPCR internalization and recycling to the cell surface or degradation dictates the neuronal responsiveness to ligand. Furthermore, GPCR trafficking plays a role in the physiological tolerance to drugs (Whistler et al., 1999; Finn and Whistler, 2001; He et al., 2002). Because most treatments of CNS disorders include long-term administration of agonists that target GPCRs, drug development and therapeutics depend on gaining a better understanding of the intracellular fate of GPCRs.
Early endosomes comprise distinct compartments identified primarily through studies of trafficking of the transferrin receptor (TfnR), a single transmembrane receptor that constitutively internalizes. At steady state, the TfnR resides in perinuclear recycling endosomes that are distinguished from other endosomes by selective retention of recycling proteins and the presence of the small GTPase Rab11a (Hopkins et al., 1994; Ullrich et al., 1996; Green et al., 1997). Rab11a (Ullrich et al., 1996; Ren et al., 1998) and myosin Vb, a member of the class V unconventional myosins (Lapierre et al., 2001) that interacts with Rab11a, are involved in TfnR recycling to the plasma membrane.
Although mechanisms involved in agonist-induced GPCR endocytosis have been extensively characterized, less is known about the intracellular pathways and molecules involved in GPCR recycling. The mechanisms mediating GPCR recycling may be distinct from those for the constitutively recycling TfnR (Cao et al., 1999). Also, different GPCR subtypes show distinct endocytic trafficking patterns (Trejo and Coughlin, 1999; Anburgh et al., 2000; Krudewig et al., 2000; Tsao and von Zastrow, 2000; Innamorati et al., 2001). Given the central role that GPCR recycling plays in regulating cellular responses to ligand, it is critical to identify the mechanisms that control GPCR recycling. Furthermore, because the mechanisms that control TfnR trafficking do not necessarily generalize to GPCRs and because the different GPCR subtypes show distinct endocytic pathways, it is crucial to characterize the intracellular trafficking of individual GPCR subtypes.
We examined the pathways responsible for the recycling of the M4 muscarinic ACh receptor (mAChR). M4 is highly expressed in the CNS (Levey et al., 1991; Hersch et al., 1994), is the primary autoreceptor in the striatum, plays a role in locomotor activity (Gomeza et al., 1999), and is a novel therapeutic target for schizophrenia (Felder et al., 2001) and Alzheimer's disease (Bodick et al., 1997). Activated M4 internalizes to endosomes in cell culture and the rat brain, and M4 recycling to the plasma membrane is required for resensitization (Bogatkewitsch et al., 1996;Bernard et al., 1999; Volpicelli et al., 2001). To define the mechanisms that control receptor recycling, we studied the endogenously expressed M4 mAChR in neuronotypic PC12 cells (Berkeley and Levey, 2000). This study identifies Rab11a and myosin Vb as critical regulators of GPCR return to the plasma membrane.
MATERIALS AND METHODS
cDNA constructs. The Rab11aS25N construct was a generous gift from Dr. Marino Zerial (Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany) (Ullrich et al., 1996). To construct the GFP-Rab11aS25N plasmid, the Rab11aS25N cDNA was amplified by PCR with a 5′ primer (5′-TTATATCTCGAGAATGGGCACCCGCGA-3′) that introduced a XhoI restriction enzyme site and a 3′ primer (5′-CGCGCGGAATTCGCCTTATATGTTCTG-3′) that introduced anEcoRI restriction enzyme site. The products were digested and ligated into the pEGFP-C2 (Clontech, Palo Alto, CA) vector. The sequence of the Rab11aS25N construct was confirmed by MWG Biotech (High Point, NC). The myosin Vb tail–GFP construct has been described previously (Lapierre et al., 2001).
Cell culture and transfections. PC12 cells were maintained at 37°C and 5% CO2 in DMEM (Mediatech, Herndon, VA) containing 10% heat-inactivated horse serum (Invitrogen, Grand Island, NY), 5% fetal clone serum (HyClone, Logan, UT), and 1% penicillin–streptomycin. Cells grown in six-well trays were transfected with 1 μg of plasmid DNA using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA). Approximately 5 hr after the transfection, the cells were rinsed three times with PBS and passaged onto Matrigel-coated coverslips in six-well trays. Immunocytochemistry was performed 2 d after transfections.
Analysis of M4 recycling. To selectively examine the proportion of internalized receptors that return to the cell surface, newly synthesized receptors were eliminated by 30 min pretreatment with cycloheximide (20 μg/ml) (Szekeres et al., 1998). Cycloheximide was also included in all subsequent drug treatments. Cells were treated with media containing 100 μm carbachol (CCh) for 60 min to induce M4 internalization. The dishes were then placed on ice, rinsed two times with chilled DMEM, returned to 37°C, and incubated with media containing cycloheximide for various time points. To calculate M4 recycling, the percentage of M4 colocalization with cell surface Na+/K+ ATPase after CCh washout was quantified (see below) and normalized to cell surface M4 in untreated cells. The residual cell surface M4 remaining after 60 min CCh treatment was then subtracted from this value to obtain the amount of M4 that recovers to the cell surface after CCh treatment and washout.
Immunocytochemistry. Cells were fixed for 30 min with 2% paraformaldehyde in 0.1 m phosphate buffer, pH 7.3 and rinsed several times with PBS containing 0.5 normal horse serum (PBS+). The cells were blocked with PBS containing 5% normal horse serum and 1% bovine serum albumin (blocking buffer) and permeabilized with 0.05% Triton X-100. The cells were rinsed several times with PBS+ and incubated in primary antibody diluted in blocking buffer overnight at 4°C. For double labeling, both primaries were included in the incubation. The following primary antibodies were used: early endosome autoantigen 1 (EEA1) mouse monoclonal (1:250; Transduction Laboratories, Lexington KY); M4 mouse monoclonal (1:500; Chemicon, Temecula, CA) (Bernard et al., 1999) or rabbit polyclonal (0.5 μg/ml) (Levey et al., 1991); Na+/K+ ATPase α-1 subunit mouse monoclonal (1:500; Upstate Biotechnology, Lake Placid, NY); Rab4 rabbit polyclonal (1:500; Biosource, Camarillo, CA), Rab7 rabbit polyclonal (1:100) (Chavrier et al., 1990), or Rab11 rabbit polyclonal (1:500; Zymed, San Francisco, CA); and TfnR mouse monoclonal (1:500; Zymed). The cells were rinsed and incubated for 60 min with rhodamine red-X or Cy5-conjugated donkey anti-rabbit or mouse secondary antibodies diluted in blocking buffer (1:100; Jackson ImmunoResearch, West Grove, PA). Both secondary antibodies were included in the incubation. The cells were then rinsed and mounted onto slides with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). The cells were scanned with a Zeiss (Heidelberg, Germany) LSM 510 laser scanning confocal microscope as described previously (Volpicelli et al., 2001), and the images were prepared using Adobe Photoshop (Adobe Systems, San Jose, CA).
Quantitation of colocalization. MetaMorph imaging system software (Universal Imaging Corporation, West Chester, PA) was used to quantitate colocalization on unprocessed images as described previously (Volpicelli et al., 2001). Briefly, background was subtracted from the images, and the percentage of overlapping pixels between M4 and Na+/K+ ATPase (or other markers) pixels was determined. Unless indicated otherwise, for each time point, eight cells from three experiments were analyzed for n = 24. All data were analyzed by independent samples t test or ANOVA and Fisher's least significant difference post hoc test.
Binding assays. After drug treatments, cells plated on six-well trays were rinsed three times with 2 ml of cold HBSS (Mediatech) and incubated overnight at 4°C with 1 nm 3H-quinuclidinyl benzilate (QNB) or 1 nm 3H-N-methyl-scopolamine (NMS). Nonspecific binding was determined using 20 μmatropine. The cells were transferred to polystyrene culture tubes and filtrated through Whatman GF/B fired glass-fiber filters. Filters were rinsed three times with cold PBS, and radioactivity was determined by liquid scintillation spectroscopy.
After prolonged agonist stimulation, M4 traffics to a TfnR-positive perinuclear compartment
In nonpolarized cells, the constitutively endocytosed and recycled TfnR localizes to a perinuclear compartment defined as recycling endosomes (Hopkins et al., 1994). To begin to study the role of this endosomal compartment in M4 recycling, we determined whether M4 colocalizes with the TfnR after sustained agonist stimulation (Fig.1 A). In untreated PC12 cells, M4 localized primarily to the cell surface, whereas the TfnR showed a concentrated localization near the nucleus. Brief stimulation with CCh (10 min) resulted in M4 internalization to peripherally distributed early endosomes (Volpicelli et al., 2001). After prolonged (60 min) CCh stimulation, M4 internalized from the cell surface and showed a perinuclear localization similar to the TfnR. M4 and the TfnR colocalized extensively in this perinuclear endosomal compartment, indicating that M4 traveled to a similar recycling endosomal compartment as the TfnR after agonist stimulation. Quantitation of the confocal images revealed minimal colocalization of M4 with the TfnR in untreated cells and a significant increase of M4 overlap with TfnR after 60 min prolonged CCh treatment.
After agonist washout, M4 returns to the cell surface
After prolonged CCh treatment, M4 may either recycle to the cell surface or be targeted to lysosomes for degradation. We therefore determined whether internalized M4 returns to the cell surface after agonist washout. To distinguish receptors that recycle to the cell surface after agonist treatment and internalization from newly synthesized receptors, cells were pretreated with cycloheximide to prevent new protein synthesis. In nonstimulated cells, M4showed colocalization with the cell surface marker Na+/K+ ATPase (Fig. 2 A). After CCh treatment, internalized M4 resided near the nucleus and showed minimal overlap with the Na+/K+ATPase. However, by 60 min of CCh washout, M4returned to the cell surface, in which it colocalized with the Na+/K+ATPase. A portion of M4 remained in an intracellular localization 60 min after CCh washout.
To quantitate recycling after agonist washout, M4overlap with the cell surface marker Na+/K+ ATPase was measured on confocal images. Using this method to quantify M4 internalization in single cells has been established previously and shown to provide results similar to measurements of internalization use binding assays with the membrane impermeant, nonselective mAChR ligand3H-NMS (Volpicelli et al., 2001). We first compared the extent of M4 recycling measured by quantitative immunocytochemistry with binding assays using3H-NMS (Fig. 2 B). Cells were treated with CCh for 60 min, rinsed, and incubated in media alone for 60 or 90 min. When the recycling data obtained from confocal images and binding assays were expressed as a percentage of untreated cells, the extent of M4 recycling measured by quantitative immunocytochemistry was slightly higher than binding assay measurements. It is possible that measuring the percentage of M4 pixels that overlap with the Na+/K+ ATPase after CCh treatment and washout included receptors that are at or near the cell surface but cannot bind ligand. When the cell surface M4 remaining after 60 min CCh treatment was subtracted from the value of cell surface M4after CCh washout, the values for M4 recycling were equivalent to binding assays. Therefore, quantitative immunocytochemistry provides a valid measure of M4 recycling after agonist stimulation and washout, and this method of calculating M4recycling was used for all of the analysis throughout the paper. Furthermore, unlike binding assays, confocal microscopy allows visualization of M4 localization to specific intracellular compartments.
Quantitative immunocytochemistry was then used to analyze a more extensive time course of M4 recycling (Fig.2 C). M4 began to recycle to the plasma membrane as early as 15 min after CCh washout, and, by 3 hr, the vast majority of detectable M4 had returned to the cell surface (t 1/2 = 60 min).
It is possible that two sources of intracellular of M4 could contribute to the reappearance of M4 to the cell surface: M4internalized after agonist stimulation and a pool of intracellular M4 that existed at baseline. We therefore measured the numbers of cell surface mAChRs (using the membrane-impermeant ligand 3H-NMS) and total mAChRs (using the permeable ligand3H-QNB). The vast majority (96%) of mAChRs reside at the cell surface in untreated PC12 cells. Unlike binding assays, confocal images show a proportion of intracellular M4 in unstimulated cells (Fig. 2), and measurements of pixel intensity reveal that ∼20% of M4 pixels represent an intracellular pool of receptors. Intracellular M4 could reside in acidic endosomes or lysosomes in which the receptor conformation is altered, or the receptor is degraded such that it cannot be recognized by ligand but can be recognized by the antibody. Prolonged treatment of control PC12 cells with atropine to prevent M4activation and internalization by endogenously released ACh does not cause a statistically significantly increase of M4 colocalization with the Na+/K+ ATPase (Volpicelli et al., 2001). Thus, the intracellular M4 detected by immunofluorescence in unstimulated cells does not traffic to the cell surface and therefore does not substantially contribute to the reappearance of M4 at the cell surface.
Although quantitative immunocytochemistry can analyze receptor distribution within a cell (i.e., the percentage of detectable receptor at the cell surface vs intracellular), it does not provide information about possible changes in total receptor levels. Therefore, binding assays using 3H-QNB, which binds to cell surface and intracellular mAChRs, were performed to measure the extent of mAChR degradation after CCh stimulation and washout. After 60 min CCh treatment, total mAChR binding was unchanged (110 ± 15% of untreated cells). Because M4 accounts for >95% of total mAChR binding in PC12 cells (Berkeley and Levey, 2000), the majority of M4 does not degrade after 60 min continual CCh treatment. However, after 60 min CCh treatment followed by 180 min washout, mAChR binding was 57 ± 4% of untreated cells. Therefore, after prolonged CCh treatment, ∼40% of M4 was targeted for degradation and the remaining M4 returned to the cell surface.
Internalized M4 traffics to a Rab11a-positive perinuclear compartment
In nonpolarized cells, Rab11a shows a perinuclear distribution similar to internalized M4 and TfnR and plays a role in TfnR recycling back to the cell surface in HeLa, CHO, and BHK cells (Ullrich et al., 1996; Ren et al., 1998). We determined whether M4 also colocalized with Rab11a in the TfnR-positive compartment of CCh-treated PC12 cells. Initially after CCh stimulation (2.5 and 10 min), M4 began to internalize from the cell surface, and the majority of M4 localized to discrete puncta distributed peripherally throughout the cell (Fig.3 A), shown previously to be EEA1-positive endosomes (Volpicelli et al., 2001). Although the majority of M4 did not colocalize with Rab11a after 2.5 min of CCh stimulation, by 10 min CCh, a significant portion of M4 accumulated near the nucleus, in which it colocalized with Rab11a. After 60 min continual CCh treatment, the majority of M4 colocalized with Rab11a in the perinuclear compartment. Quantitation of M4colocalization with Rab11a (Fig. 3 B) showed that the extent of M4 colocalization with Rab11a progressively increased over the time course of CCh treatment.
Perinuclear concentration of internalized M4 is Rab11a dependent
To determine whether M4 trafficking to perinuclear endosomes required Rab11a GTPase activity, we expressed, in PC12 cells, the constitutively GDP-bound GFP-Rab11aS25N fusion protein, which cannot exchange GDP for GTP. GFP-Rab11aS25N showed a diffuse, cytosolic distribution consistent with the inability of GDP-bound Rab proteins to bind to the membrane (Ullrich et al., 1994) (Fig.4 A). In untreated PC12 cells, expression of GFP-Rab11aS25N had no effect on the localization of M4 to the plasma membrane (data not shown). In control cells transfected with the pEGFP vector, after 60 min CCh, internalized M4 showed a perinuclear localization. In cells expressing GFP-Rab11aS25N, M4 localized to small puncta dispersed throughout the cell and did not concentrate near the nucleus. Similarly, expression of GFP-Rab11aS25N also dispersed the TfnR throughout the cell compared with pEGFP vector-transfected control cells (Fig.4 B). M4 showed colocalization with TfnR in vector-transfected control cells and in Rab11aS25N-expressing cells. Quantitative immunocytochemistry revealed that M4 colocalization with the TfnR in vector-transfected cells (51 ± 3%; n = 16) was equivalent to GFP-Rab11aS25N-transfected cells (50 ± 4%;n = 16; p = NS).
It has been proposed that constitutively GDP-bound Rab11aS25N prevents trafficking of the TfnR from early endosomes to perinuclear recycling endosomes (Ullrich et al., 1996; Ren et al., 1998). Because GFP-Rab11aS5N impairs M4 trafficking to perinuclear endosomes, we sought to identify the peripheral vesicular compartment containing M4 in cells expressing this mutant. First, we analyzed the extent of M4colocalization with the early endosome marker EEA1 (Rubino et al., 2000). In both pEGFP-transfected and GFP-Rab11aS25N-expressing cells, EEA1 localized to large puncta throughout the cell (Fig.5 A). However, in both cases, after 60 min CCh treatment, we failed to observe significant overlap between EEA1 and M4. Quantitative immunocytochemistry showed that M4 colocalization with EEA1 in Rab11aS25N-expressing cells (19 ± 3%;n = 16) was not increased compared with vector-transfected control cells (22 ± 6%; n = 16; p = NS). If GFP-Rab11a S25N prevented trafficking from EEA1-positive early endosomes, M4colocalization with EEA1 would be enhanced relative to control cells. However, although M4 shows a more dispersed localization in Rab11aS25N-expressing cells compared with vector-transfected control cells, M4colocalization with EEA1 is not enhanced. Therefore, M4 can exit from the EEA1-positive early endosomal domain in cell expressing dominant negative Rab11aS25N.
Early endosomes are comprised of two domains: a Rab5a/EEA1 domain that initially receives endocytosed material and a Rab4 domain. Rab4 is involved in recycling of the TfnR and the β-adrenergic receptor directly from early endosomes (van der Sluijs et al., 1992; Seachrist et al., 2000; Sonnichsen et al., 2000). It was thus possible that, in the presence of GFP-Rab11aS25N, M4 could transit from the EEA1-positive early endosomal domain to the Rab4-positive domain on early endosomes but fail to continue into perinuclear recycling endosomes. We thus determined whether GFP-Rab11aS25N expression enhanced M4 colocalization with Rab4. In cells expressing inactive Rab11aS25N, the morphology of Rab4-positive early endosomes was similar to pEGFP vector-transfected control cells. In pEGFP vector-transfected control cells, M4 showed ∼46 ± 2% (n = 16) colocalization with Rab4 (Fig. 5 B). Expression of GFP-Rab11aS25N reduced M4 colocalization with Rab4 (35 ± 2%; n = 16; p < 0.001). Therefore, expression of GFP-Rab11aS25N did not cause retention of M4 in either EEA1- or Rab4-positive early endosomes.
GFP-Rab11aS25N expression may enhance trafficking of M4 to a late endosomal–degradative pathway. Rab7 plays a role in trafficking from early sorting endosomes to late endosomes (Chavrier et al., 1990; Feng et al., 1995). Thus, if targeting of M4 to late endosomes is enhanced by GFP-Rab11aS25N expression, M4 colocalization with Rab7 should be enhanced. In vector-transfected cells (Fig.5 C), M4 showed a perinuclear localization and Rab7 localized to small puncta throughout the cells. The morphology of Rab7-positive endosomes did not appear to be altered by expression of GFP-Rab11aS25N. Furthermore, colocalization of M4 with Rab7 did not appear to be enhanced in GFP-Rab11sS25N-expressing cells. M4colocalization with Rab7 is 48 ± 2% (n = 16) in control cells and 49 ± 1% (n = 16;p = NS) in Rab11aS25N-expressing cells. Thus, expression of inactive GFP-Rab11aS25N does not enhance targeting of M4 to a Rab7-positive late endosomal pathway.
M4 recycling is Rab11a dependent
Because expression of dominant negative Rab11aS25N prevented trafficking of M4 to perinuclear recycling endosomes, we determined whether inhibiting Rab11a activity also affected M4 recovery to the cell surface. In contrast to control cells (Fig. 2), in GFP-Rab11aS25N-expressing cells, M4 distribution remained dispersed throughout the cytosol and showed minimal return to the cell surface, followed by 60 min CCh washout (Fig.6 A).
To quantitate M4 recycling, the percentage of M4 that colocalized with the Na+/K+ ATPase was measured on confocal images. Because of low transfection efficiency in PC12 cells, measurements of the effects of mutant Rab11a expression would not normally be possible using traditional methods, such as binding assays or cell surface biotinylation. However, quantitative immunocytochemistry in single cells revealed that GFP-Rab11aS25N expression had a dramatic effect on M4 recycling. Expression of GFP-Rab11aS25N inhibited M4recycling after CCh stimulation and 60 min washout relative to vector-transfected control cells (p < 0.001) (Fig. 6 B). However, the inhibition was not complete, and, by 180 min, M4 eventually returned to the cell surface. Quantitative immunocytochemistry revealed that the extent of cell surface M4 in GFP-Rab11aS25N-expressing cells was similar to vector-transfected control cells 180 min after CCh washout (p = NS). The inhibition of M4 recycling was specific for inhibition of Rab11a activity and not a result of general overexpression of Rab proteins because expression of dominant negative Rab5a does not affect the extent M4 recycling (data not shown). These findings indicate that Rab11a GTPase activity plays an important role in the normal recovery of M4 to the cell surface.
The myosin Vb tail enhances concentration of M4 in perinuclear endosomes
Recent studies have demonstrated that the active GTP-bound form of Rab11a can interact with the C-terminal tail domain of myosin Vb, and overexpression of a myosin Vb tail construct lacking its motor domain inhibits recycling of TfnR out of the perinuclear compartment in HeLa cells (Lapierre et al., 2001). Therefore, because M4 recycling requires Rab11a activity, we determined whether myosin Vb also plays a role in M4 return to the cell surface. PC12 cells were transfected with a GFP-myosin Vb tail construct that contains the Rab11a binding domain but lacks the myosin Vb motor domain. Consistent with previous findings (Lapierre et al., 2001), GFP-myosin Vb tail showed a concentrated localization near the nucleus (Fig.7 A). At baseline, the expression of the myosin Vb tail enhanced M4intracellular localization. To quantitate the effect of myosin Vb tail expression on the proportion of intracellular M4, the percentage of cell surface M4 that colocalized with the Na+/K+ ATPase was measured. In untreated cells, M4 showed a small but statistically significant 11% decrease in colocalization with the Na+/K+ ATPase in cells expressing the myosin Vb tail compared with vector-transfected control cells (p < 0.01). These results are consistent with previous findings that M4 shows a small amount of atropine-sensitive endocytic activity at baseline resulting from basal release of ACh (Volpicelli et al., 2001).
After 60 min CCh treatment, M4 showed enhanced concentration in perinuclear endosomes in cells expressing the myosin Vb tail relative to control cells (Fig. 1), and M4 showed extensive colocalization with the GFP-myosin Vb tail (Fig. 7 B). Expression of the myosin Vb tail also significantly enhanced M4colocalization with the TfnR relative to vector-transfected control cells. M4 colocalization with the TfnR in vector-transfected control cells was 51 ± 3% compared with 71 ± 3% (p < 0.001) in myosin Vb tail-expressing cells. Thus, the dominant negative myosin Vb tail enhances M4 accumulation in perinuclear recycling endosomes.
The myosin Vb tail inhibits M4 recycling
Previous studies have suggested that expression of the myosin Vb tail inhibits transit of recycling cargo out from the perinuclear compartment back to the cell surface (Lapierre et al., 2001). We therefore determined the effects of GFP-myosin Vb tail expression on recycling. As expected, M4 returned to the cell surface after 60 min CCh treatment and 60 min washout in vector-transfected control cells. However, in GFP-myosin Vb tail-expressing cells, the majority of M4 was retained in the perinuclear compartment with little recovery by 60 min washout (Fig. 8 A). The dramatic ability of dominant negative myosin Vb tail to inhibit M4 recycling was apparent as early as 15 min after CCh washout and remained for up to 3 hr (Fig.8 B). Quantitation of the percentage of M4 recycling after agonist treatment and washout showed that GFP-myosin Vb tail expression significantly inhibited M4 recycling relative to vector-transfected control cells at 15, 60, and 180 min after CCh washout (p < 0.001). Therefore, the myosin Vb tail concentrated M4 in perinuclear endosomes and functionally impeded M4 trafficking back to the cell surface, indicating that myosin Vb plays a key role in M4 recycling.
After agonist stimulation and internalization, the dynamic regulation of GPCR responsiveness after agonist stimulation is determined by a balance between degradation and recycling. Although the molecular mechanisms regulating GPCR internalization have received extensive study, the molecules involved in targeting GPCRs to the recycling pathway have only received attention recently. We show that, during prolonged agonist stimulation, M4accumulates in perinuclear endosomes, in which it colocalizes with TfnR and Rab11a. Expression of mutant forms of Rab11a and myosin Vb dramatically impair M4 recycling. Thus, the results presented here indicate that M4 receptors traffic through a specific recycling pathway through mechanisms dependent on both Rab11a and myosin Vb.
Previously, we demonstrated that, initially after agonist-induced internalization, M4 traffics to peripherally distributed Rab5a-positive early endosomes (Volpicelli et al., 2001). We now extend these findings by showing that, with prolonged stimulation, M4 traffics from early endosomes to perinuclear recycling endosomes, in which it colocalizes with the TfnR and Rab11a, markers for the perinuclear recycling compartment (Hopkins and Trowbridge, 1983; Yamashiro et al., 1984; Hopkins et al., 1990,1994; Ghosh et al., 1994; Ullrich et al., 1996; Green et al., 1997; Ren et al., 1998; Casanova et al., 1999; Sonnichsen et al., 2000; Wang et al., 2001). Rab11a activity is required for M4recycling because expression of the constitutively GDP-bound form of Rab11a, Rab11aS25N, alters M4 trafficking through the recycling pathway and severely impairs M4recycling after CCh washout. Other GPCRs localize to perinuclear endosomes after internalization (von Zastrow and Kobilka, 1992; Tolbert and Lameh, 1996), and some subtypes have been shown to colocalize with Rab11a (Moore et al., 1999; Innamorati et al., 2001; Kreuzer et al., 2001; Hunyady et al., 2002). Our study is the first to demonstrate that recycling of the M4 subtype of GPCR requires Rab11a activity and indicates that the perinuclear endosomal compartment may be a common recycling pathway for multiple GPCR subtypes.
After internalization to early endosomes, receptors targeted for recycling traffic to perinuclear recycling endosomes (Mellman, 1996). In cells expressing Rab11aS25N, M4 does not accumulate in perinuclear endosomes but localizes to small puncta with a vesicular appearance dispersed throughout the cell. Studies in nonpolarized cells suggest that Rab11aS25N impairs recycling by preventing receptor traffic from early endosomes (Ullrich et al., 1996;Ren et al., 1998). Thus, M4 should accumulate in early endosomes in cells expressing Rab11aS25N, within either an EEA1-positive domain or a Rab4 domain. However, our results demonstrate that internalized M4 does not show enhanced colocalization with EEA1 or Rab4 in Rab11aS25N-expressing cells. Therefore, unlike the TfnR, in this cell system, inactivation of Rab11a does not appear to prevent M4 transit from early endosomes. Expression of Rab11aS5N also does not enhance M4 colocalization with Rab7 and therefore Rab11aS5N does not enhance M4 trafficking to a late endosomal–degradative pathway. Furthermore, because the cells were treated with the protein synthesis inhibitor cycloheximide, M4 localization to biosynthetic compartments is minimal. However, M4 does show colocalization with the TfnR in Rab11aS25N-expressing cells, identifying this compartment as an early endosomal compartment. These data thus suggest that M4 traffics from early endosomes to recycling endosomes via an intermediate vesicular–early endosomal compartment and that active Rab11a is required for trafficking of these intermediate vesicles to perinuclear recycling endosomes. An alternative explanation is that, in cells expressing Rab11aS25N, M4 traffics to recycling endosomes, but accumulation of tubules–vesicles near the nucleus is prevented. Inhibition of Rab11a activity could also alter recycling system morphology by blocking association of vesicles with Rab11a effectors, (Prekeris et al., 2000, 2001; Hales et al., 2001), including interaction with the myosin Vb motor protein (Lapierre et al., 2001).
Myosin Vb is an effector molecule that selectively interacts with the active GTP-bound Rab11a but not Rab11aS25N (Lapierre et al., 2001). Myosin Vb contains a tail domain that binds Rab11a, a neck domain containing calmodulin binding motifs and a motor head domain. Expression of the myosin Vb tail causes inhibition of TfnR recycling in HeLa cells, as well as basolateral to apical transcytosis and apical recycling in polarized Madin-Darby canine kidney cells (Lapierre et al., 2001). In this study, expression of the myosin Vb tail enhances accumulation of M4 in the perinuclear compartment and inhibits recycling, even at prolonged time points after agonist washout. Thus, in myosin Vb tail-expressing cells, Rab11a-positive vesicles containing M4 can accumulate in the perinuclear compartment, but they cannot exit because myosin Vb cannot mobilize vesicles out of perinuclear compartment back to cell surface. Integrity of the actin cytoskeleton is necessary for targeting GPCRs to the recycling pathway. For example, disrupting the actin cytoskeleton targets the β-adrenergic receptor to the degradative pathway (Cao et al., 1999). Although we did not address the fate of M4 in myosin Vb tail-expressing cells because of low transfection efficiency in PC12 cells, it is possible that preventing myosin Vb interaction with the actin cytoskeleton would prevent sorting of M4 to recycling pathway and enhance M4 degradation. Regardless of the mechanism by which myosin Vb directs M4trafficking back to the cell surface, we show for the first time that a specific motor protein is involved in GPCR recycling.
In summary, after agonist stimulation, M4internalizes from the cell surface sequentially to Rab5a-positive early endosomes (Volpicelli et al., 2001) and then through a recycling endosomal system containing Rab11a and myosin Vb. Active Rab11a and myosin Vb are required for recycling of vesicles containing M4 back to the cell surface, likely mediated in part via an interaction between myosin Vb and the actin cytoskeleton. M4 plays an important role in locomotor activity (Gomeza et al., 1999), as the primary cholinergic autoreceptor in the striatum (Zhang et al., 2002) and as a target for the treatment of Alzheimer's disease (Bodick et al., 1997). Rab11a and myosin Vb are expressed in the brain (Sheehan et al., 1996; Zhao et al., 1996), and, thus, identifying that Rab11a and myosin Vb play key roles in M4 recycling helps to understand the mechanisms by which neurons may maintain responsiveness to ACh stimulation. Overall, these studies indicate that specific trafficking pathways serve to direct the accurate recycling of particular GPCRs to the plasma membrane.
This work was supported by the Alzheimer's Association (A.I.L.), a Pharmaceutical Research and Manufacturers Association Foundation Advanced Predoctoral Fellowship (L.A.V.), National Institutes of Health Grant RO1 NS30454 (A.I.L.), National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases Grants DK48370 and DK43405, and a Veterans Administration Merit Award (J.R.G.).
Correspondence should be addressed to Dr. A. I. Levey, Center for Neurodegenerative Disease, Emory University School of Medicine, Whitehead Biomedical Research Building, 615 Michael Street, Fifth Floor, Atlanta, GA 30322. E-mail:.