Rab11 regulates the recycling and lysosome targeting of β₂-adrenergic receptors

The intracellular trafficking of cell surface receptors plays a crucial role in the modulation of their activity. For β₂-adrenergic receptors (β₂ARs), members of the G protein-coupled receptor (GPCR) superfamily, their distribution during prolonged exposure to agonist is regulated by a complex array of intracellular trafficking steps. Initially, phosphorylated receptors undergo rapid internalization into early endosomes via clathrin-coated vesicles. In the early endosome, β₂ARs are likely to be exposed to protein phosphatases and dephosphorylated, following which most receptors recycle to the cell surface in a fully sensitized state (Krueger et al., 1997; Moore et al., 1995; Morrison et al., 1996; von Zastrow and Kobilka, 1992, 1994; Yu et al., 1993). With prolonged agonist exposure, receptors undergo continuous rounds of endocytosis and recycling (Morrison et al., 1996), but over time, some β₂ARs are diverted to lysosomes where receptors are degraded (Kallal et al., 1998; Moore et al., 1999b).

Some cell surface receptors, such as the G protein-coupled chemokine receptor CCR5, V2 vasopressin, endothelin A, and M₄-muscarinic receptors, and the transferrin receptor, may utilize an indirect route in recycling from early endosomes to the cell surface (Bremnes et al., 2000; Innamorati et al., 2001; Ren et al., 1998; Ullrich et al., 1996; Volpicelli et al., 2002). This indirect route involves transit from early endosomes to the pericentriolar recycling endosome (RE), then to the plasma membrane. The effect of trafficking through the RE on receptor function is unknown, other than that it results in slowed receptor recycling (Innamorati et al., 2001; Ren et al., 1998; Ullrich et al., 1996). We have reported two observations raising the possibility that some β₂ARs also utilize this slow recycling pathway. First, during prolonged exposures to agonist, the rate of β₂AR recycling from intracellular compartments to the plasma membrane significantly decreases (Moore et al., 1999a). Additionally, in cells pretreated with bafilomycin A₁, a proton pump inhibitor that raises endosome pH and inhibits traffic between early and late endosomes (Clague et al., 1994), the normal trafficking of β₂ARs to lysosomes is blocked and receptors are markedly localized with rab11, a marker for the RE (Moore et al., 1999b). We hypothesized, therefore, that the RE and slow recycling pathway is important in the regulation of β₂AR trafficking in a rab11-dependent manner during receptor downregulation.

Rab proteins are members of the small GTPase superfamily and as with all GTPases, they shuttle between two activity states determined by the phosphorylation status of a bound guanine nucleotide. Every rab protein has a characteristic subcellular distribution where it functions as an essential regulator of vectorial traffic (Novick and Zerial, 1997; Zerial and McBride, 2001). Rab11 localizes to and regulates traffic through the RE (Green et al., 1997; Ren et al., 1998; Ullrich et al., 1996) as well as between endosomes and the trans-Golgi network (TGN) (Chen et al., 1998; Wilcke et al., 2000). Studies of the transferrin receptor indicate that rab11 activity, and subsequently receptor traffic through the RE, can be modulated by overexpressing wild-type or mutant rab11 proteins (Ren et al., 1998; Ullrich et al., 1996; Wilcke et al., 2000). Overexpression of rab11S25N, which is deficient in GTP binding and functions as a dominant negative mutant, inhibits the delivery of transferrin to the RE and is believed to slow the recycling of receptors from early endosomes to the cell surface (Ren et al., 1998). Alternatively, overexpression of rab11Q70L, a mutation that confers constitutive activity for many rab proteins, causes the accumulation of transferrin in the RE (Ren et al., 1998; Ullrich et al., 1996). Among GPCRs, recycling of the M₄-muscarinic receptor and CXCR2 are significantly slowed in cells expressing the rab11S25N mutant (Fan et al., 2003; Volpicelli et al., 2002).

In the present study, we modulated rab11 activity by overexpressing wild-type and mutant rab11 proteins in human embryonic kidney 293 (HEK293) cells expressing β₂ARs. The expression of rab11S25N significantly impaired β₂AR recycling from transferrin-positive but EEA1-negative vesicles to the cell surface after both brief and prolonged exposure to agonist. Conversely, the overexpression of both wild-type rab11 (rab11WT) and the Q70L mutant induced the redistribution of β₂ARs into rab11-positive compartments and inhibited the trafficking of β₂ARs to lysosome-associated membrane protein 2 (LAMP-2)-positive late endosomes and lysosomes. The morphologic changes following prolonged exposures to agonist in cells overexpressing rab11WT were associated with a reduced rate of receptor recycling and a marked inhibition of receptor degradation. These findings indicate a pivotal role for the RE and rab11 in regulating the endosome sorting of β₂ARs.

12β6 cells (a gift of B. Kobilka, Stanford, CA) were cultured in DME with 10% fetal bovine serum and 200 μg/ml Geneticin (Life Technologies, Gaithersburg, MD). This cell line was derived from HEK293 cells and expresses human β₂ARs with an N-terminal hemaglutinin (HA) epitope tag at a level of 300,000/cell (Moore et al., 1995; von Zastrow and Kobilka, 1992). Mouse monoclonal antibodies against the HA-tag (mHA.11) and LAMP-2 were obtained from Berkeley Antibody Co. (Berkeley, CA) and BD Pharmingen (San Diego, CA), respectively, and mouse monoclonal antibodies against rab11 and EEA1 were obtained from BD Transduction Laboratories (Franklin Lakes, NJ). Rabbit polyclonal antibodies against the terminal 15 amino acids of the β₂AR C-terminus and rab11 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and Zymed Laboratories (South San Francisco, CA), respectively. Species-appropriate goat secondary antibodies and transferrin conjugated to Alexa 594 were purchased from Molecular Probes (Eugene, OR). Cy5-transferrin was purchased from Jackson ImmunoResearch Labs (West Grove, PA). FuGene 6 transfection reagent was obtained from Roche Molecular Biochemicals (Basel). The pEGFP (enhanced green fluorescent protein) expression vector was obtained from CLONTECH Laboratories (Palo Alto, CA) and rab11 cDNAs cloned into pEGFP were kindly provided by Dr Angela Wandinger-Ness (Albuquerque, NM). All other reagents were obtained from Sigma Chemical Co. (St Louis, MO) unless otherwise noted.

12β6 cells growing on poly-D-lysine coated cover slips in 6-well clusters were treated as indicated in the figure legends. Following treatment, the cells were washed with PBS containing 1.2% sucrose (PBSS), fixed with 4% paraformaldehyde in PBSS at 4°C for 10 minutes, then washed again with PBSS. The following steps were done at 25°C, with PBSS used for washes. The fixed cells were incubated in 0.34% L-lysine, 0.05% Na-m-periodate for 20 minutes, then washed and permeabilized with 0.2% Triton X-100. After a further wash, the cells were blocked with 10% normal goat serum for 15 minutes. Primary antibodies were diluted in PBSS with 0.2% normal goat serum and 0.05% Triton X-100, then added to the cells and left for 1-12 hours. The cells were washed four times before labeling with secondary antibodies using the same procedure as for primary antibodies. We used the following concentrations of antibodies: mHA.11, 5 μg/ml; polyclonal anti-β₂AR C-terminus 2 μg/ml; anti-rab11, 10 μg/ml; anti-LAMP-2, 5 μg/ml; anti-EEA1, 1:100 dilution; Alexa 488- and Alexa 647-anti-mouse IgG, 1:200 dilution; and Alexa 594 anti-rabbit IgG, 1:200 dilution. The cover slips were mounted in Mowiol and viewed using a DeltaVision deconvolution microscopy system (Applied Precision Inc., Issaquah, WA) equipped with a Zeiss Axiovert microscope. Imaging was performed using a Zeiss ×100 oil immersion lens and sections were collected at an optical depth of 150 nm in the z-plane. The localization of vesicles was verified by the examination of multiple planes. Images were optimized using DeltaVision deconvolution software, pseudocolored using MetaMorph software (Universal Imaging Corporation, West Chester, PA), and transferred to Photoshop 6.0 (Adobe Systems, San Jose, CA) for the production of final figures.

To quantify β₂AR trafficking to late endosomes and lysosomes following prolonged exposures to agonist, the colocalization of receptors and LAMP-2 was determined using MetaMorph software. Quantification of overlap was performed following the construction of a binary image for each channel based on the merged image. The extent of colocalization was determined from four independent sections from each cell. The final values reflect the means±s.e.m. for 7-10 cells from 3 separate experiments.

12β6 cells were plated onto poly-D-lysine-coated 22-mm-diameter glass coverslips at 50% confluence in 6-well clusters. After 24 hours and just prior to transfection, the media was removed and replaced with antibiotic-free media containing 3% fetal bovine serum. Cells were transfected with 2 μg per well of empty pEGFP vector as a control, pEGFP-rab11WT, or pEGFP-rab11S25N using FuGene 6 transfection reagent according to the manufacturer's instructions, with a DNA to FuGene 6 ratio of 2 μg/3 μl. The transfection mixture remained on cells for 48 hours, at which time the cells were treated as indicated in the figure legends. In control studies, FuGene 6 transfection reagent alone had no effect on cell morphology or the distribution of β₂ARs and organelle markers (data not shown).

To study the effects of rab11 overexpression on the distribution of β₂ARs following prolonged exposures to ISO (isoproterenol), leupeptin (100 μM) was added prior to ISO to allow the accumulation of receptors in lysosomes (Moore et al., 1999b). In some experiments cells were fed Alexa 594-transferrin (50-100 μg/ml) in serum-free media as described in the figure legends and receptors identified using mHA-11 antibody followed by anti-mouse Alexa 647 IgG. In specified experiments, cells were washed, fixed, and labeled for β₂ARs and either LAMP-2 or EEA1, and the secondary antibodies were anti-rabbit Alexa 594 IgG (1:200 dilution) and anti-mouse Alexa 647 IgG (1:100).

For morphologic studies of β₂AR recycling, cells were treated with ISO (5 μM) for 20 minutes or 6 hours, washed thoroughly, and placed in warm media containing 5 μM propranolol for varying times. In some experiments, cells were fed Alexa 594-transferrin as indicated in the figure legends to selectively label the endocytic pathway.

EcR293 cells stably express the regulatory protein VgRXR, a chimeric steroid receptor that is activated by synthetic ecdysteroids such as ponasterone (No et al., 1996). Cells were initially transfected to stably express human β₂ARs with an N-terminus FLAG epitope using FuGene transfection reagent as previously described, with selection for G418 resistance. EGFP-rab11 constructs were cloned into the pIND-Hygro vector, which has a promoter element responsive to VgRXR, cloning sites for insertion of genes and a hygromycin resistance gene. pIND-Hygro/EGFP-rab11 and pIND-Hygro/EGFP rab11S25N were stably transfected into EcR293 cells using FuGene6, and hygromycin-resistant clones screened for ponasterone-inducible fluorescence. For induction of rab11 overexpression, cells were treated for 72 hours with 5 μM ponasterone, whereas control cells were treated with vehicle alone (0.125% ethanol).

Following treatment with ponasterone A (PON) or ethanol for 72 hours, transfected cells were washed with PBS, then dissolved in Laemmli sample buffer (Laemmli, 1970). Samples were electrophoresed through a 4-15% Tris-HCl SDS-polyacrylamide gradient gel (Ready Gel, BioRad, Hercules, CA) and transferred to Immobilon-P membranes (Millipore, Bedford, MA). The membranes were blocked in 5% dried milk in 0.2% Tween and probed with anti-rab11 mAb at a dilution of 1:2000, and bound antibodies were detected by chemiluminescence (ECL-Pierce Chemical Co., Rockford, IL). Immunoblots were quantified by densitometry using SigmaGel software (SPSS Science, Chicago, IL).

Cells growing on poly-D-lysine coated 24-well clusters were cultured in the presence of ponasterone or ethanol for 72 hours, then treated with ISO for 20 minutes or 6 hours. The cells were washed extensively and incubated in DME with 20 mM HEPES, pH 7.4 (DME-H) prewarmed to 37°C to allow recycling. At varying times up to 90 minutes, the cells were chilled and surface receptors measured by incubation in DME-H containing 6 nM [³H]CGP12177 at 0°C for 2 hours, followed by washing with cold DME-H. The monolayers were then lysed with 1% SDS and 1% NP-40, and the lysates counted by scintillation spectroscopy. Nonspecific binding was determined by parallel incubations with 5 μM propranolol and was less than 5% of total binding. For every experiment, binding at each time point was measured in triplicate. The fraction of internal receptors remaining was plotted versus time of recycling, and the curves fitted by nonlinear regression using GraphPad Prism software (version 3.0) and recycling rate constants determined as previously described (Morrison et al., 1996). Recycling rate constants are presented as the mean±s.e.m. of three separate experiments.

Transient overexpression of rab11WT, rab11S25N, and empty vector was performed in 12β6 cells as described above. After 48 hours, the media was removed and replaced with serum-free media containing DiI-LDL (20 μg/ml) for 5 minutes at 37°C. Cells were washed four times in chilled DME, then incubated for 45 minutes at 37°C in DME containing Cy5-transferrin (25 μg/ml). Following fixation in 4% paraformaldehyde and lysine/Na-m-periodate, cells were mounted using Mowiol and viewed and processed as described above. In other experiments, the cells were fed DiI-LDL for 5 minutes as above, washed, and then after 45 minutes fixed and labeled for LAMP-2, as previously described.

Transfected cells growing on poly-D-lysine coated 6-well clusters were cultured in the presence of either ethanol or ponasterone for 48 hours to induce EGFP-rab11 expression. The monolayers were then treated with EZ-link sulfo-NHS-biotin (0.5 mg/well) for 30 minutes at 25°C to biotinylate surface receptors. The biotinylated cells were washed with Dulbecco's phosphate buffered saline (DPBS) containing 0.9 mM CaCl₂ and 0.5 mM MgCl₂ to remove excess biotin and treated with ISO (5 μM) for the indicated times up to 24 hours. Following incubation with agonist, cells were washed and harvested in DPBS containing leupeptin (10 μg/ml), then pelleted and solubilized at 4°C in n-dodecyl-β-D-maltoside (DDM) buffer (20 mM HEPES, pH 7.4, 300 mM NaCl, 5 mM EDTA, 0.8% DDM, and Complete™ EDTA-free protease inhibitor at standard concentration). Lysates were centrifuged at 16,000 g to remove cellular debris. From each sample, 50 μg of protein was added to 50 μl streptavidin agarose beads (Sigma) at 4°C for 1 hour to bind biotinylated receptors, then the beads were pelleted by centrifugation and washed using DDM buffer followed by DPBS. The final pellets were resuspended in Laemmli buffer, heated to 65°C for 15 minutes, and centrifuged at 11,000 g to free biotinylated receptors from the streptavidin-agarose. A 10-μg aliquot of eluted protein from each sample was treated with 2 μl peptide N-glycosidase F (New England Biolabs, Beverly, MA) for 90 minutes at 37°C to remove N-terminal sugar moieties and improve the resolution of β₂ARs. The samples were electrophoresed and transferred to Immobilon-P membranes as described above. The membranes were probed with the anti-β₂AR C-terminus polyclonal antibody at a dilution of 1:1000 and detected using chemiluminescence. Bands were quantified by densitometry using SigmaGel software. Results are shown as the mean±s.d. of three separate experiments.

Previous data implicates roles for rab5 and rab4 in the endocytosis and rapid recycling of β₂ARs during brief exposures to agonist (Seachrist et al., 2000). We have reported that the rate of β₂AR recycling decreases with prolonged exposure to agonist (Moore et al., 1999a), suggesting that some receptors also utilize a slow recycling pathway. Furthermore, in cells pretreated with bafilomycin A₁, β₂ARs accumulate in the RE where they localize with rab11 (Moore et al., 1999b). Thus, we predicted that β₂AR recycling should be altered in cells expressing wild-type (WT) or dominant-negative-mutant rab11 (S25N). We therefore tracked the recycling of β₂ARs in HEK293 cells stably expressing β₂ARs and transiently expressing EGFP-rab11 fusion proteins or empty pEGFP vector as a control following brief (20 minute) and prolonged (6 hour) exposures to agonist.

EGFP-rab11S25N localized to the cytosol and to a perinuclear compartment (Fig. 1A,b′), which we identified as trans-Golgi network by labeling with antibody against the TGN-specific adaptor protein AP1 (data not shown), consistent with the reported localization of rab11S25N mutant in BHK cells (Chen et al., 1998). EGFP-rab11WT was largely localized to perinuclear vesicles, consistent with REs (Fig. 1A,c′).

We determined the effects of rab11 on β₂AR recycling following a 20-minute exposure to agonist, a time point at which receptors have undergone only one or two rounds of endocytosis and recycling (Morrison et al., 1996). In control and rab11-transfected cells, there was a rapid redistribution of β₂ARs from the plasma membrane into punctate vesicles that were positive for EEA1 and transferrin (data not shown). Cells were then thoroughly washed and placed in warm media containing the antagonist propranolol to prevent any additional receptor internalization. In both control cells (Fig. 1A,a) and cells expressing EGFP-rab11WT (Fig. 1A,c), β₂ARs primarily localized to the plasma membrane within 15 minutes following the removal of agonist. The small fraction of receptors that failed to recycle were present in a perinuclear distribution in the case of control cells (Fig. 1A,a) or localized with EGFP-rab11WT (Fig. 1A,c). By 30 minutes after removal of agonist, recycling was virtually complete in control (Fig. 1B,a) and rab11WT-transfected cells (Fig. 1B,c). In cells overexpressing rab11S25N, some receptors remained localized to EEA1-positive endosomes after 15 minutes of recycling (Fig. 1A,b), although virtually all β₂ARs had exited these EEA1 endosomes after 30 minutes of recycling (Fig. 1B,b). However, many receptors failed to recycle even after 30 minutes and localized to dispersed vesicles devoid of EEA1. There was little or no colocalization of receptors with rab11S25N after 15 or 30 minutes of recycling, suggesting that β₂ARs do not transit through the TGN.

To characterize these EEA1-negative β₂AR-containing vesicles further, we treated cells transiently expressing EGFP-rab11S25N as above except during the recycling period Alexa 594-transferrin (50 μg/ml) was added to label the endocytic and recycling pathways. Nonrecycled β₂ARs were largely localized to transferrin-positive vesicles in cells expressing EGFP-rab11S25N after 15 minutes (data not shown), 30 minutes (Fig. 2A,b), and 60 minutes of recycling, (data not shown). To determine if these transferrin-positive compartments represent post-early endosome vesicles, we performed pulse-chase experiments in which Alexa 594-transferrin was added for 20 minutes during the exposure to ISO, washed, and allowed to chase for 15 minutes during β₂AR recycling. In both vector- and rab11S25N-transfected cells, transferrin was predominately localized to EEA1-positive early endosomes following the pulse and largely segregated from EEA1 after the chase (data not shown), but with strikingly different distributions. In control cells, transferrin and β₂ARs colocalized in a perinuclear compartment consistent with the RE (Fig. 2B,a), whereas in cells expressing rab11S25N they partially colocalized in vesicles dispersed throughout the cytoplasm (Fig. 2B,b). This finding suggests that rab11S25N inhibits β₂AR recycling from post-early endosome vesicles.

To quantify the effects of modulating rab11 activity on β₂AR traffic, we created EcR293 (EcR293: β₂AR) cell lines stably expressing epitope-tagged β₂ARs in which wild-type or mutant rab11 could be inducibly overexpressed upon addition of the synthetic insect hormone ponasterone A. Using immunoblot analysis, we found that EGFP-rab11 proteins were significantly overexpressed following 72 hours of induction by ponasterone A, with the levels of expression of rab11WT and the rab11S25N mutant being approximately 4-fold and 2-fold, respectively, that of endogenous rab11 (Fig. 3). The overexpression of both rab11WT and rab11S25N to these levels had no effect on either the rate or extent of β₂AR internalization (Table 1 and data not shown).

In transfected EcR293: β₂AR cells, we directly measured β₂AR recycling in cells overexpressing rab11 proteins and in vehicle control cells following brief (20-minute) exposures to agonist. As shown in Table 1, the first-order rate constants for receptor recycling (k_r) in uninduced cells were similar to those for untransfected EcR293:β₂AR cells and to those previously measured in 12β6 cells (Morrison et al., 1996), indicating that neither transfection with pIND-EGFP-rab11 proteins nor the slight leak in their expression significantly affected the rate of receptor recycling. In cells overexpressing the rab11S25N mutant, receptor recycling was significantly impaired following a brief exposure to agonist (Fig. 4A and Table 1). In contrast, the overexpression of rab11WT had no effect on recycling following a 20-minute exposure to agonist (Fig. 4B and Table 1), consistent with our morphologic data in transiently transfected cells.

Although overexpressing wild-type rab11 did not overtly alter the trafficking of β₂ARs following a brief exposure to agonist, we considered the possibility that there may be an effect following prolonged exposures to agonist as we have previously shown that β₂AR recycling slows significantly during downregulation (Moore et al., 1999a). Cells transiently expressing EGFP-rab11WT or vector were exposed to agonist for 6 hours, then immediately fixed or washed to allow receptor recycling for 15 or 30 minutes, as described above. Under both conditions, some receptors were localized in EEA1-positive early endosomes following 6 hours of agonist treatment (Fig. 5A). In both control cells and cells expressing EGFP-rab11WT, receptors also were distributed to dense perinuclear compartments that contained either pulse-chased transferrin (Fig. 5B) or EGFP-rab11 (Fig. 5A,b), indicating that β₂ARs accumulate in REs. After removal of agonist, receptors largely recycled to the cell surface in control cells after 15 minutes (Fig. 6A,a) and 30 minutes (Fig. 6B,a). However, in cells expressing EGFP-rab11WT, β₂ARs had exited EEA1-positive early endosomes within 15 minutes (Fig. 6A,b) but some receptors remained in the RE even after 30 minutes (Fig. 6B, b), suggesting that rab11WT slows recycling from the RE but not from early endosomes. The effect of expressing EGFP-rab11S25N on β₂AR recycling following a prolonged exposure to agonist was similar to that observed after a 20-minute agonist treatment, with nonrecycled β₂ARs largely appearing in dispersed EEA1-negative but transferrin-positive vesicles (data not shown).

We next quantified β₂AR recycling in transfected EcR293: β₂AR cells following a 6-hour exposure to agonist as described above. As expected, the expression of EGFP-rab11WT following ponasterone A treatment significantly decreased both the rate and extent of β₂AR recycling following a prolonged exposure to agonist (Fig. 6C and Table 1). Similar to its effect after a brief exposure to agonist, EGFP-rab11SN also significantly slowed receptor recycling following a 6-hour exposure to agonist (Table 1).

As the overexpression of rab11WT induced the accumulation of β₂ARs in the RE during prolonged exposures to agonist, we considered that overexpressing wild-type or mutant rab11 also may alter β₂AR trafficking to the degradative pathway. 12β6 cells transiently overexpressing empty vector or rab11WT were treated with leupeptin (100 μM) to allow the accumulation of β₂ARs in lysosomes (Moore et al., 1999b) and exposed to ISO (5 μM) for 6 hours. In control cells, β₂ARs (Fig. 7A,a) had a punctate labeling pattern and extensively localized with the late endosome/lysosome marker LAMP-2, similar to our previous observations (Moore et al., 1999b). However, in cells overexpressing rab11WT (Fig. 7A,b) or rab11Q70L (data not shown), there was significantly less colocalization of β₂ARs and LAMP-2 (percent colocalization in control cells, 44.60±3.07 versus 11.51±1.65 for rab11WT cells, P<0.05). Instead, β₂AR extensively localized to rab11-positive structures and to peripheral endosomes.

To correlate this finding with a physiologic outcome, we directly quantified β₂AR degradation in the inducible EcR293:β₂AR:EGFP-rab11WT cell line by measuring the loss of surface-biotinylated receptors after varying exposures to agonist in cells pretreated with either ponasterone A or vehicle. β₂AR degradation in vehicle treated cells proceeded in a manner similar to untransfected cells (data not shown). However, in cells overexpressing rab11WT (Fig. 7B,C), receptor degradation was significantly inhibited following a 6-hour exposure to agonist, consistent with our morphologic data showing decreased receptor localization to lysosomes at this time point.

The finding that overexpressing rab11WT inhibited β₂AR trafficking to LAMP2-positive late endosomes and lysosomes suggests that rab11 regulates the trafficking of β₂ARs to lysosomes during prolonged exposure to agonists. We sought to determine if this effect resulted from a general disruption of trafficking directly from early endosomes to late endosomes. To address this question, we examined the effects of overexpressing rab11WT on the trafficking of an endocytic ligand, low-density lipoprotein (LDL). LDL internalizes with its receptor into early endosomes, where it localizes with transferrin, and then is sorted to late endosomes and lysosomes (Ghosh et al., 1994). If overexpressing rab11WT has a direct effect on early endosome to late endosome/lysosome traffic, we predicted that LDL also would be unable to transit to lysosomes but would remain in early endosomes and localize with transferrin. To test this prediction, we pulsed 12β6 cells transiently overexpressing rab11 or vector with the fluorescent ligand DiI-LDL (20 μg/ml) for 5 minutes, washed, and allowed DiI-LDL to chase for 45 minutes in the presence of Cy5-transferrin (25 μg/ml) to label the endocytic pathway. As seen in Fig. 8, the overexpression of rab11WT had no effect on the distribution of DiI-LDL when compared to control cells expressing EGFP alone. In all cases, only a small amount of colocalization of DiI-LDL and Cy5-transferrin was observed (Fig. 8A), consistent with the normal transiting of LDL through early endosomes to the degradative pathway. To confirm this result, we performed similar pulse-chase experiments with DiI-LDL in control cells (data not shown) and in cells overexpressing rab11WT (Fig. 8B) and then labeled cells for LAMP-2. As expected and similar to control cells (data not shown), most DiI-LDL localized with LAMP-2 after a 45-minute chase, indicating that rab11WT does not inhibit the trafficking of LDL to late endosomes and lysosomes.

We and others have shown that during agonist treatment that β₂ARs undergo continuous rounds of endocytosis into and recycling from early endosomes, processes that are regulated by rab5 and rab4, respectively (Morrison et al., 1996; Seachrist et al., 2000; von Zastrow and Kobilka, 1994). Previous studies fail to demonstrate significant localization of β₂ARs with rab11 following 1 hour and 6 hours of agonist treatment, suggesting that these receptors do not transit through the RE (Innamorati et al., 2001; Moore et al., 1999b). In this study, we now present evidence that β₂ARs also recycle via the rab11-positive RE and that rab11 regulates multiple steps of agonist-induced β₂AR trafficking. First, in control cells, some β₂ARs localized to a perinuclear compartment that also contained pulse-chased transferrin following a prolonged exposure to agonist. Second, overexpressing rab11S25N slowed β₂AR recycling from transferrin-positive vesicles following both brief and prolonged exposures to agonist. Third, the overexpression of rab11WT inhibited β₂AR recycling from the RE following prolonged exposures to agonist. Fourth, the overexpression of rab11WT inhibited β₂AR trafficking to lysosomes and receptor degradation, whereas the sorting of DiI-LDL from transferrin-containing endosomes to late endosomes and lysosomes was not affected.

The effects of overexpressing rab11WT on β₂AR recycling following prolonged exposures to agonist are similar to those previously described for transferrin recycling (Ren et al., 1998; Ullrich et al., 1996). It is possible that overexpressing rab11WT slows receptor recycling by increasing β₂AR targeting to the RE, decreasing receptor efflux from this compartment, or both. However, studies of transferrin receptor trafficking in HeLa cells suggest that the main effect of rab11WT is in regulating the exit of receptors from the RE and not their delivery to that compartment (Wilcke et al., 2000). Our data showing that β₂ARs accumulated in the RE even after 30 minutes of recycling are consistent with this interpretation. Overexpressing rab11WT had no effect on either the rate or extent of β₂AR recycling following a brief exposure to agonist (Fig. 4B). At this time point most receptors are localized to early endosomes and receptors have undergone only one or two rounds of endocytosis and recycling. The trafficking of receptors from EEA1-positive early endosomes following prolonged exposure to agonist was also unimpaired (Fig. 6). These findings indicate that rab11WT does not regulate the direct, rapid recycling pathway but slows recycling through the RE. Further, these data suggest that the fraction of β₂ARs trafficking through the RE with each round of endocytosis and recycling must be relatively small compared with the rapid-recycling pathway.

Studies of transferrin receptor trafficking suggest that rab11S25N inhibits receptor recycling from early endosomes and prevents their delivery to the RE (Ren et al., 1998). In our studies, overexpressing rab11S25N induced a significant inhibition of β₂AR recycling following both short and prolonged exposures to agonist, with nonrecycled receptors largely being retained in dispersed transferrin-positive but primarily EEA1-negative vesicles 15 and 30 minutes after the removal of agonist (Figs 1, 2). This finding suggests that rab11S25N inhibits the recycling of β₂ARs from a post-early endosome recycling compartment and is consistent with recently published studies of M₄-muscarininc receptor trafficking in PC12 cells (Volpicelli et al., 2002) and CXCR2 trafficking in HEK293 cells (Fan et al., 2003). However, the exact nature of these vesicles remains in question. Although it is possible that this compartment reflects dispersed recycling endosomes, in the case of M₄ receptors these vesicles are believed to represent an intermediate compartment between early endosomes and REs (Volpicelli et al., 2002). M₄-muscarinic receptors are known to largely accumulate in the RE, however this does not appear to be the case for β₂ARs (Innamorati et al., 2001). Instead, β₂ARs are believed to primarily recycle to the cell surface directly from early endosomes in a rab4-dependent manner (Seachrist et al., 2000). Thus, our results raise the possibility that these β₂AR-containing vesicles are located between early endosomes and the plasma membrane in the direct recycling pathway. Such a conclusion is consistent with published data suggesting that rab11S25N inhibits transferrin transit from sorting endosomes to both the RE and directly to the cell surface (Ren et al., 1998). Further, recent studies reveal the presence of recycling domains within endosomes that are occupied by both rab4 and rab11 (De Renzis et al., 2002; Sonnichsen et al., 2000), suggesting that these small GTPases may together coordinate proper sorting along the recycling pathway. A recently described 80 kDa protein, Rab Coupling Protein (RCP), serves as an effector for both rab4 and rab11 and is required for normal endosomal recycling (Lindsay et al., 2002). Thus, it is possible that rab11S25N impairs rapid recycling by binding RCP, or some other effector protein, thus preventing the functional interaction of effector with rab4.

We have shown that the overexpression of rab11WT inhibits the trafficking of β₂ARs to late endosomes and lysosomes and significantly delays receptor degradation (Fig. 7). However, our finding that LDL sorting from transferrin-containing endosomes to late endosomes/lysosomes was apparently not altered following the modulation of rab11 activity suggests that rab11 does not directly regulate early endosome to late endosome traffic (Fig. 8). This conclusion is consistent with published data showing that epidermal growth factor (EGF) receptors can travel to late endosomes and are normally degraded in HeLa cells overexpressing rab11Q70L (Wilcke et al., 2000). However, in contrast to EGF receptors, β₂ARs predominately recycle to the cell surface following endocytosis and only slowly travel to lysosomes (Moore et al., 1999a). Thus, it is likely that overexpression of rab11WT has an indirect effect on β₂AR trafficking to lysosomes by causing β₂ARs to become `trapped' in the slow recycling pathway, resulting in a diminished pool of receptors available for targeting to lysosomes. Supporting this explanation are our data showing significant colocalization of β₂ARs and rab11WT and slowed receptor recycling following prolonged exposure to agonist in cells overexpressing rab11WT. These data, along with our previous results in bafilomycin-A₁-treated cells (Moore et al., 1999b), suggest that the slow recycling and degradative pathways are competitive and that receptor degradation can be decreased by inducing receptor accumulation in the RE.

Current data indicate that rab11 regulates the sorting of materials in the RE to at least two destinations, the plasma membrane and the TGN (Chen et al., 1998; Ren et al., 1998; Wilcke et al., 2000). Based on our results, we cannot exclude the possibility that there exists an additional rab11-regulated pathway from the RE to late endosomes or lysosomes. Recent data indicate that the delivery of cation-independent mannose-6-phosphate receptors to late endosomes is distal to their trafficking through the RE (Lin et al., 2004), although no such pathway has been described for any GPCR. It is conceivable that receptors, such as β₂ARs, that predominately recycle to the cell surface following endocytosis and only slowly move on to lysosomes, transit through the RE en route to the degradative pathway. Alternatively, rapidly degraded receptors, such as EGF receptors, are sorted directly from early endosomes to late endosomes, a process that is not dependent on rab11. Furthermore, a recent report suggests a role for rab4 in the regulation of both the recycling and degradative pathways (McCaffrey et al., 2001), indicating that a rab protein may direct multiple sorting events. Whatever the explanation, our results clearly show that although overexpressing rab11WT has no effect on the rapid recycling of β₂ARs, it inhibits their recycling from the RE and their trafficking to lysosomes and subsequent degradation.

In conclusion, these data indicate that β₂ARs are not only sorted in the early endosome to either the plasma membrane or to lysosomes, but also transit through the RE, a process that is regulated by rab11. It is unknown why some β₂ARs recycle via the RE and what targets receptors to this pathway. Studies of the V2 vasopressin receptor suggest that its accumulation in the RE is dependent upon the phosphorylation status of serine/threonine residues within a cluster in the cytoplasmic tail of the receptor (Innamorati et al., 1998; Innamorati et al., 1999; Innamorati et al., 2001). A similar serine cluster has been described for the β₂AR that is important in its rapid desensitization, but the role of these serines on receptor sorting is unknown (Seibold et al., 1998). Whatever the mechanism, our data clearly show that the trafficking of β₂ARs through the rab11-positive RE is associated with slowed rates of receptor recycling, diminished receptor delivery to lysosomes, and delayed receptor degradation. Although a recent report suggests that the RE is not essential for transferrin receptor recycling (Sheff et al., 2002), our findings raise the possibility that transit through the RE may serve dual roles in the modulation of β₂AR activity: contributing to receptor desensitization by slowing the recycling of dephosphorylated receptors to the cell surface as well as limiting receptor downregulation by protecting against receptor degradation.

We wish to thank Angela Wandinger-Ness for providing the EGFP-rab11 constructs and Douglas Eikenburg, Keith Morrison, and Klaus-Peter Zimmer for their helpful comments. This work was supported by grants from the National Institutes of Health to R.H.M. (HL64934) and B.J.K. (HL57445 and HL50047).