The binding of angiotensin II (Ang II) to the Ang II type I receptor , a member of the superfamily of G-protein-coupled receptors (GPCRs), activates and initiates a series of intracellular processing,1, 2, 3 including rapid phosphorylation, desensitization, and endocytosis of the . The internalized , via clathrin-coated vesicles (CCV), binds to -arrestins and dynamin.4, 5 CCVs fuse with sorting endosomes shortly after internalization with participation of a number of Rab GTPases.
The Rab family proteins are a 23-25kDa ras superfamily of GTPases that contain two C-terminal geranylgeranyl (20 carbons) groups serving to bind proteins tightly to membranes.6 To date, over 60 mammalian Rab proteins have been identified; each Rab protein appears to associate with a particular membrane compartment(s) and regulate intracellular protein trafficking, such as endocytosis, endosome fusion, exocytosis, and recycling of GPCRs. 7, 8
It is believed that two distinct intracellular systems regulate the recycling of internalized GPCRs: one is the Rab4-mediated rapid recycling pathway, and the other is the Rab11-mediated slow endosome pathway. 5, 6, 7, 8, 9 Rab4 is mainly localized in early endosomes and is thought to play an important role in the efflux of cargo proteins out of early endosomes and the rapid recycling of cargo proteins directly to the plasma membrane from the early endosome.10, 11 Rab11 is mainly localized in perinuclear recycling endosomes and the trans-Golgi network and regulates the slow recycling from perinuclear endosomes and the trans-Golgi network to the plasma membrane.12, 13
Internalized GPCRs recycle back to the cell surface through different pathways.5, 8 Based on their affinity to and association with -arrestins, GPCRs are divided into two classes4: class A receptors, which bind weakly with -arrestins and recycle rapidly back to the plasma membrane, and class B receptors, which bind tightly with -arrestins and recycle slowly back to the plasma membrane. The rapid recycling of -adrenogenic receptor , a typical class A GPCR, appears to be regulated by Rab4, directly from early endosomes,14 and may15 or may not16 be regulated by Rab11. , a typical class B GPCR, is suggested to be regulated by Rab11 for its recycling.16, 17 Overexpression of wild-type Rab11 and constitutively active Rab11 (Q70L) significantly increases recycling to the plasma membrane.16 Several studies have indicated that Rab11 also regulates the recycling of other class B GPCRs, such as vasopressin V2 receptor,18 CXC chemokine receptor 2,19 -isoform of the thromboxane A2 receptor,20 and M4 muscarinic acetylcholine receptor.21 Recently, Rab4 has also been suggested to participate in recycling.22 Thus, the roles of Rab4, Rab11, or both in the recycling of the are not well understood.
In the current study, fluorescence resonance energy transfer (FRET) microcopy23 was employed to detect the dynamic relationship between and endogenous Rab4 or Rab11 during the of the entire recycling course after the internalization of . FRET is a powerful tool to detect protein-protein interaction of less than 10-nm distance,24 and it can overcome the limitation (resolution is hundreds of nm) of the methods used in most of previous studies on the morphological co-localization of Rab4 and Rab11 with the cargo GPCR. Moreover, some of the Rab-dominant negative mutants used in several studies are not GTPase-defective, as originally presumed.25 Here, we provide additional insights into the relationship between Rab4/Rab11 and recycling. We observed that both Rab4 and Rab11 are associated with the recycling, and the association varies depending upon the recycling stage, as observed by FRET microscopy and co-immunoprecipitation. Moreover, gene knockdown of either Rab4 or Rab11 by specific siRNAs disrupts their dynamic association and recycling. Based on these studies, we propose a model of a Rab4 and Rab11 coordinated regulation of recycling.
Materials and Methods
Antibodies and Reagents
Monoclonal mouse Rab4 and Rab11 antibodies were purchased from BD Transduction Laboratories (Lexington, Kentucky). Polyclonal rabbit anti-GFP, polyclonal rabbit anti-Rab4, and polyclonal goat anti-Rab11 and normal mouse, rabbit, and goat antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, California). Alexa-546, Alexa-555, and Alexa-633 protein labeling kits were obtained from Molecular Probes (Eugene, Oregon). Cycloheximide and other obtained reagents were obtained from Sigma (St. Louis, MO).
Cell Culture and Transfection
Human embryonic kidney (HEK) 293 cells from ATCC (Manassas, Virginia) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) containing glucose, 10% fetal bovine serum (FBS), -glutamine, and sodium pyruvate. Human -EGFP (enhanced green fluorescence protein) or its empty vector pEGFP-N126 was transfected into HEK 293 cells using Lipofectamine 2000 transfection reagents (Invitrogen), as described previously,27 according to the manufacturer’s instructions. Transfectants were selected with G418. Stable transfectants were tested and confirmed by immunoblotting with anti-GFP and anti- antibodies, flow cytometry, and radioligand binding assay (data not shown). The EGFP tag did not interrupt the function of , determined by a dose-dependent Ang II–induced phosphorylation of extracellular signal-regulated kinase (ERK1/2) (data not shown). The stably transfected cells expressing human -EGFP were designated in this study as HEK 293 cells.
siRNA and Transfection
Specific Rab4 (AATGCAGGAACTGGCAAATCT), Rab11 (AAGAGTAATCTCCTGTCTCGA), and negative control (AATTCTCCGAACGTGTCACGT) siRNA duplexes were purchased from Qiagen (Valencia, California). HEK 293 cells were transfected with these siRNA using Lipofectamine 2000 transfection reagents (Invitrogen) according to the manufacturer’s instructions. After siRNA treatment for , cells were split into three portions. A small portion of cells was used to check Rab4 and Rab11 levels by immunoblotting with anti-Rab4 and Rab11 antibodies [Fig. 1b ]. The largest portion of cells was used for immunoprecipitation experiments [Fig. 1c]. The third portion of the cells was dispersed in DMEM containing 10% FBS and placed ( cells per well) on six-well plates with collagen I-coated coverslips; the attached cells were gently washed and fixed with 4% formaldehyde in PBS for observation of -EGFP fluorescence at the indicated recycling times after the cells have been treated with Ang II for (Fig. 2 ).
Immunoprecipitation and Immunoblotting
Vehicle- or siRNA-transfected HEK 293 cells were homogenized by freeze-thaw in a lysis buffer ( , pH sucrose), with 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF), benzamidine hydrochloride, and protease inhibitors (soybean and lima bean trypsin inhibitors, leupeptin, and aprotinin, each ). The homogenates were incubated (rocking, , ) with of anti-GFP rabbit IgG, of anti-Rab4 rabbit IgG, or of anti-Rab11 goat IgG in of the lysis buffer with , 0.1% ovalbumin, AEBSF, 0.5% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate. Controls were normal rabbit, mouse, or goat IgG (data not shown). After adding of a 50% slurry of protein G-Sepharose CL-4B (Amersham Pharmacia, Uppsala, Sweden) in PBS and incubation at overnight, the beads were washed three times with of ice-cold PBS, containing AEBSF. Proteins bound to beads were eluted in of loading buffer at for , separated by SDS/PAGE (sodium dodecyl sulfate/polyacrylamide gel electrophoresis) in 4 to 12% gradient gels, and transferred onto nitrocellulose membrane for incubation with mAb against Rab4 or Rab11 and rabbit anti-GFP, followed by appropriate horseradish peroxidase-conjugated secondary antibodies and detection using Super Signal Chemiluminescent substrate (Pierce, Rockford, IL).
Confocal Immunofluorescence Microscopy
HEK 293 cells were fixed with 4% paraformaldehyde in PBS for at room temperature. After washing with PBS, the fixed cells on coverslips were incubated overnight at with Alexa Fluor 633-conjugated monoclonal mouse Rab4 antibody ( ; Alexa Fluor 633 protein labeling kit, Molecular Probe, Eugene, Oregon) or Alexa Fluor 546-conjugated polyclonal goat Rab11 antibody ( ; Alexa Fluor 546 protein labeling kit, Molecular Probe). The coverslips were mounted in SlowFade mounting medium (Molecular Probe) and sealed onto glass slides. Samples were imaged using an Olympus Fluoview FV300 laser scanning confocal microscope equipped with a NA objective.
Receptor Recycling Assays
Vehicle- or siRNA-transfected HEK 293 cells were treated with vehicle or Ang II for at and then placed on ice. After washing three times with ice-cold, serum-free DMEM medium, the cells were recultured in complete culture medium containing cycloheximide and reincubated at (with 5% ) for the indicated time from 0 to . The cells were then fixed with 4% paraformaldehyde for fluorescence microscopy. The density of at the cell surface was determined by quantifying cell surface fluorescence using MetaMorph 7.0 (Molecular Devices, Downington, Pennsylvania).27 After identifying the plasma membrane, regions of interest (ROIs) were drawn manually in zoomed-in images. The background was subtracted from each image, and then the images were thresholded to identify specific EGFP fluorescence for at the plasma membrane. Receptor recycling was defined as the recovery of cell-surface receptors following the removal of Ang II, compared with the cell-surface expression of receptors in cells that were not exposed to Ang II (vehicle-treated cells).
FRET Microscopy and Data Processing
The fluorophore pairs used for FRET imaging in this study were -EGFP (as donor dipole) and Alexa Fluor 555 (as acceptor dipole) conjugated with Rab4 or Rab11 antibodies (Alexa Fluor 555 protein labeling kit, Molecular Probe). Seven images were acquired for each FRET analysis, as described,23 with an Olympus Fluoview FV300 laser scanning confocal microscope equipped with a NA objective, Argon and laser, and emission filters and 590-nm long press (LP) filter. Either single-labeled donor or acceptor or double-labeled samples were acquired under the same conditions throughout the image collection. The uncorrected FRET images (uFRET) were acquired by donor excitation in the acceptor channel, which contained pure FRET (pFRET) and contaminations from both donor and acceptor spectral bleed-through (SBT). pFRET images were generated by employing a described algorithm23 for pixel-by-pixel removal of donor and acceptor SBT on the basis of matched fluorescence levels between the double-labeled specimen and the single-labeled reference specimens.
ROIs were selected in the uFRET images.23 In this study, we used image (e) (donor excitation in the donor channel of the double-labeled specimen) as the reference image for selection of ROIs to determine the plasma membrane, cytoplasm, and perinuclear compartments. Image g was acquired at acceptor excitation in the acceptor channel of the double-labeled specimen.
The percentage of energy transfer efficiency images was processed on a pixel-by-pixel basis by using the following equation:and are the photo multiplier tube (PMT) gains of donor and acceptor channels; and are the spectral sensitivity of donor and acceptor channels provided by the manufacturer; and are the donor and acceptor quantum yield, measured by spectrofluorometer, as described28; Ida is the image of donor excitation in the donor channel of the double-label samples after removing the background; and pFRET is the “processed FRET” or “pure FRET.” The calculation of distance of donor and acceptor was based on the equation as described in Ref. 23: ; Förster’s distance in this study was .
Subcellular Location of Human , Rab4, and Rab11
In HEK 293 cells under basal conditions, human tagged with EGFP was observed primarily at the plasma membrane [Fig. 3a ] but also in the intracellular membrane, perinuclear compartments, and cytoplasm [Fig. 3a], consistent with previous observations.22, 27
Endogenous Rab4 staining was observed in typical vesicles scattered throughout the cytoplasm, and very occasionally concentrated in the perinuclear compartments [Fig. 3b]. Endogenous Rab11 was scattered throughout the cytoplasm as well, but most often, Rab11 was observed to concentrate in the perinuclear areas [Fig. 3c]. Overall, Rab4 was more dispersed than Rab11 throughout the cytoplasm, but neither Rab4 nor Rab11 was observed at the plasma membrane. These observations are in agreement with studies using GFP-tagged Rab4 and Rab11.6, 22
Recycling Course of
In this report, the cell surface at the basal state was set at 100%. As shown in Fig. 4 , the relative surface intensity of is ( cells) when cells were treated with Ang II for , a time considered as the start of recycling (time ); subsequent recycling of to the plasma membrane was quantified at the indicated time points after removing of Ang II. The membrane surface fluorescence intensity was directly related to recovery time : , which followed a first-order exponential curve [Fig. 4a]. After , the cell surface reached slightly over 75% of its basal level, and close to 90% of its basal level at about , similar to that reported22 for the rat , but much slower than that14, 17 for class A GPCRs, e.g., . However, only was expressed at the plasma membrane at steady state (Fig. 4), probably due to the lack of newly synthesized for replacement, implying that of were processed for degradation.
For a first-order exponential fit of the data, the following values are obtained: is 0.9931, with two degrees of freedom, and an absolute of square of 0.00168; the values for a second-order exponential fit [Fig. 4b] are: is 0.5833, with one degree of freedom, and an absolute of square is 0.1014. The second-order polynomial curve is a poor fit [Fig. 4c]. These analyses show that recovery rate best fits a first-order exponential equation, consistent with a single recycling pathway for human .
In the following studies, for convenience, we divided the entire recycling course into early (from time 0 to ), middle (from to ), and late (over ) recycling stages.
Role of Rab4 and Rab11 in the Early Stage of Recycling
FRET was not observed between and Rab4 or Rab11 during the basal state (data not shown), which is consistent with the absence of co-localization of with either Rab4 or Rab11 in the basal state (Fig. 3).
At the start of recycling (time 0), obvious FRET was observed in the cytoplasm between Rab4 and [Fig. 5a ], the efficiency of energy transfer was ( cells), the estimated distance was (Table 1 ), but no FRET was observed between Rab11 and [Fig. 5b, Table 1]. The association of and Rab4, but not and Rab11 was also observed by immunoprecipitation [Fig. 1a, lane 2].
Calculation of efficiency of energy transfer (E%) and molecular proximity (R) between AT1R and Rab4 or Rab11 during the entire AT1R recycling course. AT1R HEK 293 cells were treated and fixed as described in Sec. 2. AT1R -EGFP is the donor; the acceptor is Rab4 or Rab11. FRET analysis is described in Figs. 5, 6, 7, 8, 9. Data are mean±SE ; n , number of cells analyzed, Infin, infinity.
|Plasma membrane||Cytoplasm||Perinuclear compartment|
|Recycling time (min)||Acceptor||n||E (%)||R(Å)||E (%)||R(Å)||E (%)||R(Å)|
At time , FRET was obvious between Rab4 and in cytoplasm and perinuclear compartments, especially in vesicles in the cytoplasm [Fig. 6a ], the % was ( cells) and ( cells), respectively (Table 1). FRET was also observed between Rab11 and [Fig. 6b] in perinuclear compartments ( , cells) and cytoplasm ( , cells). No FRET was observed at the plasma membrane between with either Rab4 or Rab11 (Fig. 6). The association of Rab4 and Rab11 with was also observed by immunoprecipitation [Fig. 1a, lane 3]. These observations indicated that Rab4 played a major role(s) during the early recycling period; Rab11 could also play some role in this stage.
Role of Rab4 and Rab11 in the Middle Stage of Recycling
At time , FRET between Rab4 and in the cytoplasm minimally decreased ( , cells), but increased in perinuclear compartments ( , cells) [Fig. 7a ; Table 1], compared to the early recycling stage. In contrast, FRET between Rab11 and increased in both the cytoplasm ( , cells) and perinuclear compartments ( , cells) [Fig. 7b] compared to its early stage . At , no FRET was observed between and Rab4 or Rab11 at the plasma membrane [Fig. 7; Table 1). Of note, the calculated of Rab4 and ( , cells) was similar to that of Rab11 and ( , cells) ; both Rab4 and Rab11 were co-immunoprecipitated by the GFP antibody (for ) [Fig. 1a, lane 4] and vice versa (data not shown). Furthermore, gene knockdown of Rab4 by specific Rab4 siRNA disrupted the association of with Rab11 [Fig. 1c]; Rab11 gene knockdown also disrupted the association of Rab4 with [Fig. 1d]. All of these data indicated that Rab4 and Rab11 were in the same recycling endosomes for trafficking at this stage. Therefore, both Rab4 and Rab11 play important roles in trafficking during this period.
Role of Rab4 and Rab11 in the Late Stage of Recycling
At time , no FRET was observed between Rab4 and in the cytoplasm, perinuclear compartments, and the plasma membrane [Fig. 8a ], but obvious FRET was still observed between Rab11 and in the perinuclear compartments [Fig. 8b]. The remarkable observation in this late period is that FRET occurred at the plasma membrane between Rab11 and ( , cells) [Fig. 8b; Table 1], indicating the importance of Rab11 in the late stage of recycling.
At time , no FRET was observed between Rab4 or Rab11 and at the plasma membrane, cytoplasm, or perinuclear compartments (Fig. 9 ; Table 1), similar to that observed at the basal state (data not shown).
Interruption of Recycling by Rab4, Rab11 Gene Silencing
The recovery rate was markedly decreased in either Rab4 or Rab11 siRNA- treated cells (Fig. 2), which confirmed the conclusion obtained from the FRET studies that both Rab4 and Rab11 regulate the recycling of .
Relationship of Rab4 and Rab11 During the Entire Recycling Course
In the basal state [Fig. 10a ] and the start of recycling, time 0 [Fig. 10b] no co-localization was observed between Rab4 and Rab11. At [Fig. 10c] and [Fig. 10d], obvious co-localization of Rab4 and Rab11 was observed, and the greatest co-localization occurred at time , predominantly at perinuclear compartments. At the late stage of the recycling period, co-localization was no longer observed [Figs. 10e and 10f].
The quantity of cell surface receptors is important for the cell to maintain its normal functions and responses to environmental changes.1 Receptor expression on the cell surface is maintained by a dynamic balance among internalization, endocytosis, degradation, exocytosis, and recycling. The receptor recycling was initially thought to be a bulk flow process.29 However, subsequent studies have shown that it is a regulated process. For example, the PDZ (PSD-95, Dlg, and ZO-1) motif30 of and the MRS (MOR-derived endocytic recycling sequence) sequence of opioid receptors31 are important for their rapid recycling, respectively. In recent years, it has become widely appreciated that Rab4 and Rab11 are responsible for the regulation of receptor recycling. 5, 6, 7, 8, 9, 10, 11, 32, 33, 34, 35, 36 Transferrin receptor (Tfn) recycles efficiently to the cell surface via two distinct pathways: the rapid pathway is mediated by Rab4, while the slow pathway is initiated by tubular extension from sorting endosomes through Rab11-positive perinulear endosomes.5, 35, 36 It is also accepted that the Rab4-mediated rapid pathway is utilized for class A GPCR recycling, such as , similar to the rapid pathway of Tfn. The class B GPCRs, including V2 vasopressin receptor,18 m4AchR,21 and somatostatin receptor,37 recycle through the Rab11-mediated slow pathway, similar to the slow recycling of Tfn. , which belongs to class B GPCR, recycles slowly back to the cell surface after internalization.5, 22
The molecular mechanism for recycling has been controversial. ARAP1, or type 1 angiotensin II receptor-associated protein 1, has been shown to facilitate the recycling to the cell surface,38 but this mechanism is not yet fully accepted. Ferguson and co-workers8, 16 proposed a Rab11-mediated slow recycling model after studying the role of wild-type Rab11 and constitutively active Rab11 (Q70L) in the recycling of the rat . Hunyady 22 have observed that Rab4 also co-localizes with and proposed two distinct rat recycling models, based on differential sensitivity to wortmannin, a PI3 kinase inhibitor. In the current study, both co-immunoprecipitation of with Rab4 [Fig. 1a] and FRET was observed between Rab4 and during the early (Figs. 5 and 6) and middle (Fig. 7) periods of recycling, consistent with observations22 that Rab4 also plays a role in the recycling of . However, there may not be two separate Rab pathways for human recycling: (1) during the entire recycling period, no FRET between Rab4 and was observed at the plasma membrane or subplasma membrane. Instead, FRET between and Rab11 was observed at the plasma membrane or subplasma membrane at the late recycling stage [Fig. 8b]; (2) in the current study, the recovery rate fits a first-order [Fig. 4a] but not second-order exponential equation [Figs. 4b and 4c], consistent with one sequential recycling pathway.
FRET is revolutionizing studies in life sciences and is widely applied in biology, biochemistry, immunology, cell and molecular biology, and clinical medicine to study protein-protein interaction, cellular signaling, conformational structure, ligand-receptor interaction, and diagnosis of certain diseases.24 It is worthwhile to mention the advantage of pixel-by-pixel analysis in the current protocol, which enables the quantitative detection of FRET occurrence and removal of spectral bleedthrough from both donor and acceptor.23 Pixel-by-pixel analysis also enables the detection of the spatial and temporal protein-protein interaction inside cells. Even though FRET can be measured at 10 to distance of molecular proximity using the current protocol, the accurate location of a particular compartment is still challenging.
In the current study, FRET was observed and analyzed spatially into three compartments: the plasma membrane, cytoplasm, and perinuclear compartments. At the early stage of recycling, internalized human were mainly in Rab4-positive compartments in the cytoplasm [Fig. 5a]. No FRET was observed between Rab4 and human at the plasma membrane [Fig. 6a], indicating no rapid recycling occurred with Rab4. At the middle stage, FRET of with Rab4 and Rab11 was observed (Fig. 7), and the association of with both Rab4 and Rab11 was supported by the co-IP data [Fig. 1a]. The association was disrupted by either Rab4 or Rab11 gene knockdown [Figs. 1c and 1d], indicating that was localized mainly in compartments containing both Rab4 and Rab11 at this stage. However, at the late stage, was localized mainly in Rab11-positive endosomes, both in the perinuclear compartments and at the plasma or subplasma membrane areas (Fig. 8). Rab4 or Rab11 gene knockdown decreased the recovery rate (Fig. 2). Thus, our data indicate that both Rab4 and Rab11 coordinately regulate the entire recycling of .
Based on the current and previous reports, a single recycling pathway for human recycling coordinated by both Rab4 and Rab11 is proposed (Fig. 11 ). Internalized rapidly appears in the CCVs. CCVs then fuse with Rab5/Rab4-positive early endosomes. The exact role(s) of Rab4 is not clear in the early endosomes, but it may enable the cargoed vesicles to fuse with Rab11-positive perinuclear compartments. After the completion of fusion, Rab11 and Rab4 form common recycling endosomes mainly at the perinuclear compartments. Subsequently, the destined for recycling buds off, via Rab11, and is recycled back to the plasma membrane by fusing Rab11-positive endosomes with plasma membrane.
Both Rab4 and Rab11 are important in the regulation of human recycling. Like other small GTPases, Rab4 and Rab11 have active GTP-bound and inactive GDP-bound forms, which are regulated by their GTPase-activating proteins, guanine nucleotide dissociation inhibitor, and guanine nucleotide exchange proteins.39, 40 The identification of the specific regulators could help us to better understand the regulation of recycling.
We are grateful to Dr. Tamas Balla (Endocrinology and Reproduction Research Branch, NICHD, National Institutes of Health, Bethesda, Maryland 20892) for critical reading of the manuscript, Drs. P. M. Lanctôt and G. Guillemette (Université de Sherbrooke, Canada) for the -GFP construct and its empty vector pEGFP-N1, Dr. Susette Mueller (Lombardi Comprehensive Cancer Center, Georgetown University, Washington, DC) for the assistance with confocal microscopy. This study was supported in part by grants from the National Institutes of Health (Grant Nos. HL074940-01, HL23081, DK39308, HL68686, and DK52612), the K. M. Keck Foundation at the University of Virginia (Charlottesville), and U.S. Public Health Service Grant Nos. 2P30-CA-51008 and 1S10RR15768-01.