Rab5 Isoforms Differentially Regulate the Trafficking and Degradation of Epidermal Growth Factor Receptors*

Ligand-mediated endocytosis is an intricate regulatory mechanism for epidermal growth factor receptor (EGFR) signal transduction. Coordinated trafficking of EGFR ensures its temporal and spatial communication with downstream signaling effectors. We focused our work on Rab5, a monomeric GTPase shown to participate in early stages of the endocytic pathway. Rab5 has three isoforms (A, B, and C) sharing more than 90% of sequence identity. We individually ablated endogenous isoforms in HeLa cells with short interfering RNAs and examined the loss-of-function phenotypes. We found that suppression of Rab5A or 5B hampered the degradation of EGFR, whereas Rab5C depletion had very little effect. The differential delay of EGFR degradation also corresponds with retarded progression of EGFR from early to late endosomes. We investigated the activators/effectors of Rab5A that can potentially separate its potency in EGFR degradation from other isoforms and found that Rin1, a Rab5 exchange factor, preferably associated with Rab5A. Moreover, Rab5A activation is sensitive to EGF stimulation, and suppression of Rin1 diminished this sensitivity. Based on our results together with previous work showing that Rin1 interacts with signal transducing adapter molecule to facilitate the degradation of EGFR (Kong, C., Su, X., Chen, P. I., and Stahl, P. D. (2007) J. Biol. Chem. 282, 15294–15301), we hypothesize that the selective association of Rab5A and Rin1 contributes to the dominance of Rab5A in EGFR trafficking, whereas the other isoforms may have major functions unrelated to the EGFR degradation pathway.

Rabs are small molecular weight guanine nucleotide-binding proteins (G proteins) that specialize in regulating different stages of the intracellular membrane trafficking based on their subcellular localization and interacting protein scaffolds (2). The increasing number of Rab family members during evolution reflects ongoing specialization of membrane trafficking pathways. At least 63 Rabs have been identified in humans. Among these, Rab5 serves as the master regulator of the endocytic trafficking (3). It functions through recruitment of specific effector proteins involved in membrane tethering and docking (4 -8). The demonstration that at least 20 cytosolic proteins specifically interact with active Rab5 highlights the complexity of the downstream regulation by this GTPase and raises the possibility that Rab5 might also manage other aspects of the endosome function (4,9). A common feature of the Rab family small GTPases is the existence of subgroups of structurally related isoforms sharing a high sequence identity (10). The Rab5 subgroup has three isoforms (A, B, and C) (3). A large scale mRNA expression profiling study from 79 human and 61 mouse nonredundant tissues shows distinct tissue distributions of the Rab5 isoforms, suggesting that the trafficking properties of the early endosomal network in developmentally distinct cell and tissue types is likely fine-tuned to fulfill subspecializations of a given pathway (10). In fact, several lines of evidence indicate that Rab5 isoforms can be functionally different. First, Rab5A is found to be the only isoform transcriptionally up-regulated in response to treatment with cytokines like interleukin-4 or interferon-␥ in macrophages (11,12). Moreover, Rab5A but not Rab5C was found to be involved in the maturation of phagosomes containing Listeria monocytogenes (13). In primary hippocampal cultures, treatment with 3,5 dihydroxyphenylglycine, a group I metabotropic glutamate receptor agonist, leads to up-regulation of Rab5B but not Rab5A, thereby reducing N-methyl-Daspartate-type glutamate receptor-mediated membrane current and cell death (14,15). Also, Rab5 isoforms can be differentially phosphorylated at a consensus motif in position 123 by distinct proline-directed Ser/Thr kinases in vitro (16).
Epidermal growth factor receptor (EGFR) 2 is the prototype of Class I transmembrane receptor tyrosine kinase operating through activation of its intrinsic tyrosine kinase upon ligand binding. Activated EGFR stimulates numerous signal transduction pathways that mediate a wide spectrum of cell responses, including cell proliferation, differentiation, and apoptosis (17, 18). To regulate the strength and duration of the signaling, activated EGFR also initiates a negative feedback mechanism that eventually leads to the removal of EGF-EGFR complexes from the plasma membrane by endocytosis (19). Previous studies have shown that overexpression of Rab5A enhances EGF-stimulated fluid phase endocytosis and EGF-EGFR internalization, whereas dominant negative Rab5 represses these processes (20,21). In addition, dominant negative Rab5 substantially inhibits the degradation of EGFR (21). On the other hand, continual expression of constitutively active Rab5 causes a ligand-inde-* This work was supported, in whole or in part, by National Institutes of Health pendent redistribution of EGFR into the intracellular vesicles (22). Interestingly, although expression of a constitutively active Rab5 shows no significant effect on EGFR levels (22), it enhances cell growth and receptor signaling (23).
In this study, we took advantage of the RNA interference silencing technique to individually knock down endogenous Rab5 isoforms in HeLa cells, so that we could examine the lossof-function phenotypes more likely to capture their physiological activities. Our results indicate that suppression of Rab5A significantly impairs the degradation of EGFR, whereas Rab5C suppression has little effect. The delay of EGFR degradation elicited by the absence of Rab5A occurs after EGFR is internalized, because the rate of EGFR internalization is unaffected by suppression of a single isoform. In addition, we identified Rin1 (Ras interaction/interference 1), a Vps9 domain-containing Rab5 exchange factor, to preferentially associate Rab5A and selectively potentiate its activity in response to EGF stimulation.

EXPERIMENTAL PROCEDURES
Antibodies and Reagents-A mouse monoclonal anti-Rab5A antibody was received as a gift from Dr. A. Wandinger-Ness (University of New Mexico, Albuquerque, NM). Anti-Rab5B antibody was kindly provided by Dr. David B. Wilson (Washington University, St. Louis, MO). Anti-Rab5C antibody is obtained from Sigma-Aldrich (Prestige Antibodies, HPA003426). Human epidermal growth factor and monoclonal anti-EGFR (Ab5) antibody were obtained from Calbiochem. Rabbit polyclonal anti-EGFR (1005) and monoclonal GFP antibody (sc-9996) was from Santa Cruz Biotechnology, Inc. Polyclonal GFP antibody was kindly provided by Dr. Phyllis Hanson (Washington University). Monoclonal anti-V5 antibody was from Invitrogen. Rabbit polyclonal Hrs antibody was a gift of Dr. Tim McGraw (Cornell University). Monoclonal anti-EEA1 and anti-Rin1 antibodies were both from BD Transduction Laboratory. Anti-Cbl antibody (7G10) was from Upstate. Anti-Grb2 antibody was from Cell Signaling. Anti-ubiquitin antibody was obtained from Zymed Laboratories Inc. (Invitrogen).
Plasmid Constructs-cDNA of Rab5A, B, and C were subcloned into SalI/BamHI sites of pEGFP-C1 (Clontech) and SacI/BamHI sites of RFP-C3 expression vector (kindly provided by Arnold Hayer, Swiss Federal Institute of Technology, Zürich, Switzerland). A Rin1 expression construct was prepared by subcloning full-length human Rin1 PCR product into the mammalian expression vector pcDNA3.1/V5-His TOPO TA (Invitrogen).
Cell Culture and Transfection-HeLa, HEK293T, and DU145 cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% bovine growth serum (Hyclone Laboratories) containing penicillin and streptomycin. The transfection was performed using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's instructions.
Immunoblotting and Immunoprecipitation-Immunoblotting analysis was conducted as described before (1). Briefly, the cell lysates were prepared with the lysis buffer containing protease inhibitor mixture (Sigma). The cell lysates were clarified by centrifugation prior to separation by SDS-PAGE. The resolved proteins were transferred to nitrocellulose membranes (Whatman Schleicher & Schuell, Florham Park, NJ) and then blocked in TBST containing 5% nonfat milk. The membranes were probed with primary antibodies and then horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA), and the proteins were visualized by enhanced chemiluminescence detection reagents (Pierce). The immunoblot data were quantified by Alpha-EaseFC 4.0 software (Alpha Innotech Corp., San Leandro, CA). For immunoprecipitation, the clarified cell lysates were incubated with primary antibodies and protein G-Sepharose (Sigma) overnight at 4°C. The beads were washed extensively with lysis buffer and solubilized in SDS sample loading buffer.
EGFR Degradation Assay-Control or experimental cells were serum-starved and then treated with 100 ng/ml EGF and 25 g/ml cyclohexamide for different times. At the end of each time point, the cells were washed with PBS and then lysed in lysis buffer as indicated above. The lysates were subjected to SDS-PAGE and immunoblotting with appropriate antibodies.
Biotinylation and Internalization of Cell Surface EGFR-Confluent HeLa cells pretreated with siRNA were starved in serum-free medium. One set of experiment was kept unstimulated, whereas the other two sets were stimulated with 100 ng/ml EGF for 2 and 5 min. The internalization was stopped by washing with ice-cold PBS-Ca-Mg (PBS, pH 7.4, 0.1 mM CaCl 2 , 1 mM MgCl 2 ). The chilled cells were then incubated with biotin labeling solution containing 0.5 mg/ml EZ-Link Sulfo-NHS-LC-Biotin (Pierce) in PBS-Ca-Mg at 4°C for 30 min. The reaction was quenched by washing the cells with ice-cold PBS-Ca-Mg containing 15 mM glycine. The cells were then washed again with PBS and lysed as described above. The biotinylated EGFRs were purified using MagnaBind TM streptavidin beads (Pierce) according to the product instructions. Precipitated biotin-EGFR then were subjected to SDS-PAGE and immunoblot analysis. The rate of internalization is calculated by densitometry of EGFR bands.
Quantification of Endogenous Rab5 Isoforms-His-Rab5A, B, and C recombinant proteins were purified from bacterial lysates. HeLa and DU145 cell lysates were prepared as described above. 5 or 10 g of total proteins from DU145 or HeLa cells, respectively, were subjected to SDS-PAGE alongside purified recombinant Rab5 isoform standards. The inten-sity of the protein standard bands was quantified with Alpha-EaseFc 4.0 software and plotted against their concentration predetermined with BCA assay to define the linear range of the signal. Band signals from cell lysates were then calculated to determine the ratio of different isoforms.
Immunofluorescent Microscopy-Cells grown on coverslips were fixed with 3% paraformaldehyde (Electron Microscope Sciences), permeabilized, and then blocked with goat serum. Following incubation with primary antibodies and then Alexa Fluor 594-or 488-conjugated secondary antibodies, the coverslips were mounted with fluorescent mounting medium (Dako-Cytomation, Carpinteria, CA). Confocal microscopy was performed using a 63ϫ objective and fluorescein sets on a MRC1024 (Bio-Rad). Uptake of Alexa 488-labeled EGF or transferrin was performed by first starving cells in serum-free medium for 3 or 1 h at 37°C. The cells were then incubated with cold serum-free medium containing 400 ng/ml fluorescein EGF or 40 g/ml transferrin for 1 h at 4°C. Endocytosis were initiated by replacing the ligand binding medium with prewarmed medium. At the end of each time point, the cells were rapidly chilled, washed, fixed with paraformaldehyde, and then subjected to immunostaining as described above.
Rab5 Activation Assay-cDNA of the Rab5-binding domain (R5BD, residues 739 -862) of Rabaptin5-pGEX was a kind gift from Dr. Guangpu Li (University of Oklahoma Health Sciences Center). The resulting construct termed pGEX/Rabaptin-5(R5BD) was expressed in DH5␣. GFP-Rab5 isoforms were expressed in HeLa cells for 24 h. The cells were first starved and then treated with 100 ng/ml EGF for 5 min. The cells were subjected to lysis for 5 min in the lysis buffer (25 mM HEPES, pH 7.4, 100 mM NaCl, 5 mM MgCl 2 , 0.1% Nonidet P-40, 10% glycerol, and protease inhibitor). Aliquots of clarified lysates were incubated with GST-R5BD prebound to the glutathione-Sepharose-4B resin for 30 min at 4°C on a rotating mixer. The resin was subsequently rinsed with the lysis buffer, resuspended in SDS sample buffer, and then subjected to SDS-PAGE and immunoblot analysis.
Statistical Analysis-All of the experiments presented were repeated a minimum of three times. The data represent the means Ϯ S.E. Student's t test was used to calculate statistical significance.

Depletion of Rab5A in HeLa and DU145 Cells Leads to
Delayed EGFR Degradation-To study the functional differences of endogenous Rab5 isoforms, we designed siRNAs to specifically silence individual isoforms. At least two independent siRNAs were used to suppress each isoform. Immunoblotting with isoform-specific antibodies confirmed that the suppression was more than 90% from each siRNA, and the isoform-specific siRNAs did not cross-react with other isoforms, because the levels of the other two isoforms were unaffected when one isoform was silenced (Fig. 1A). Following siRNA treatment, HeLa cells were assayed for EGFR levels at steady state. We found that cells depleted of Rab5A or 5B consistently exhibited increased level of total EGFR, whereas the loss of Rab5C showed very few effects (Fig. 1A). The accumulation of EGFR in cells depleted of Rab5A or 5B suggests that the sorting of EGFR into degradation pathway is likely interrupted. Therefore, we examined the rate of EGFR degradation by stimulating HeLa cells with EGF for different times. The levels of EGFR at early time points (3 and 8 min) were not used for quantitative analysis, because we found that the immunoblot signals of EGFR were highly variable among experiments at these time points. The results (collected from 15, 30, and 60 min) showed that in HeLa cells, the initial rate of degradation was substantially reduced in Rab5A-or Rab5B-depleted cells; however, between 45 and 60 min, the receptors eventually underwent degradation to an extent similar to the control cells (Fig. 1, B and C, and supplemental Fig. S1). The degree of delay is more profound with Rab5A depletion than with Rab5B knockdowns. Of all three Rab5 isoforms, Rab5C suppression seems to have the least impact. This result suggests that the degradation of EGFR requires both Rab5A and 5B. It is possible that the two isoforms work in concert or sequence along the degradation pathway, so that only one of the two isoforms present is not sufficient to transport EGFR at normal rates. We also tested the rate of EGFR degradation in DU145 cells. DU145 is a prostate cancer cell line expressing EGFR at much higher level than HeLa cells. In agreement with what we found in HeLa cells, the depletion of Rab5A also leads to more delayed EGFR degradation in DU145 cells (Fig. 1D). This finding not only verified that the function of endogenous Rab5 is critical for sorting EGFR into degradation pathway but also provided new evidence indicating a preferential connection between endogenous Rab5A and the trafficking of EGFR.
Quantification of Rab5 Isoforms in HeLa and DU 145 Cells-Previously, Bucci and co-workers (16) reported that HeLa cells express all three Rab5 isoforms with Rab5B to be more enriched than the other two. To test the possibility that the differential delay of EGFR trafficking caused by individual Rab5 isoform silencing corresponds to the relative isoform abundance, we examined the concentration of Rab5 isoforms in HeLa and DU145 cells using purified recombinant Rab5 isoform proteins as standards. Whole cell lysates from HeLa or DU145 cells were analyzed alongside Rab5 protein standards during SDS-PAGE. Following protein gel transfer, the membranes were probed with isoform-specific antibodies, and the intensity of the bands from protein standards was plotted against its concentration to demonstrate the linearity of the band signals (supplemental Fig. S2) and the avidity of the isoform-specific antibodies. Using this method, we were able to estimate that the ratio of endogenous Rab5A:Rab5B:Rab5C was ϳ1:1:2 in HeLa cells and 2:3:1 in DU145 cells (Fig. 2). These data show that Rab5A is not the most abundant isoform in either cell line, yet it dominates the endocytic trafficking of EGFR.
Overexpression of Rab5A Shows More Potency in Down-regulating EGFR-We next tested whether overexpression of individual Rab5 isoforms has opposite effects on EGFR degradation. HeLa cells were transiently transfected with individual CFP-Rab5 isoforms overnight and EGFR degradation assays were carried out as described above. Our findings indicate that exogenously expressed Rab5A, but neither 5B nor 5C accelerated the degradation of EGFR in HeLa cells (Fig. 3, A and B). In accordance with the biochemical analysis, immunostaining of total EGFR in cells transfected with RFP-Rab5 isoforms revealed that Rab5A-expressing cells have weaker EGFR signals, suggesting enhanced down-regulation ( Fig. 3C and supplemental Fig. S3). We did, however, notice that the degree of acceleration was relatively mild (Fig. 3, A and B). Because the transfection efficiency of Rab5 isoform constructs generally does not exceed 80% in HeLa cells, it is possible that the acceleration of EGFR degradation is masked by untransfected cells. Alternatively, we reasoned that under our experimental conditions, endogenous Rab5 isoforms may have already maximized the rate of trafficking and sequestered the limiting factors participating in this process; therefore, excess Rab5 could not improve the transport much further. On the other hand, the finding with Rab5A expression did coincide with the outcome of its depletion, implicating Rab5A as the primary Rab5 isoform in the EGFR degradation pathway.
Depletion of Individual Rab5 Isoforms Shows No Significant Effect on EGFR Internalization-Next, to identify the sites where Rab5A and/or 5B are most active, we examined whether loss of individual isoforms could interfere the internalization of EGFR. Control or Rab5 isoform siRNAs-treated HeLa cells were starved and then stimulated with EGF for the indicated times (Fig. 4, A and B). EGFR remaining on the cell surface was biotin-labeled, precipitated by streptavidin beads, and analyzed with immunoblotting. The data show that depletion of individual Rab5 isoforms did not inhibit the internalization of EGFR, whereas simultaneous depletion of all three isoforms delayed the internalization by ϳ50% (Fig. 4, A and B). Similar results have been reported previously with 125 I-EGF based assay by Sorkin and co-workers (24). Similar to EGFR, the endocytosis of transferrin (Tfn) receptors was not altered by silencing of individual Rab5 isoforms, whereas silencing all three isoforms together showed marked reduction of Tfn uptake ( Fig. 4C and  supplemental Fig. S4). These data suggest that Rab5 isoforms, although actively participating in the internalization step of receptor trafficking, are functionally redundant at this stage. Rab5A or 5B Delayed the Exit of EGFR from Early Endosomal Compartments-After internalization, EGFR sorting into the late endosome and degradation in the lysosome are necessary to terminate receptor signaling. c-Cbl-mediated ubiquitylation has been shown to be essential for regulating these events and ensuring proper degradation of EGFR (25,26). Upon EGF stimulation, c-Cbl binds directly to the EGFR via Tyr-1045 and indirectly through the Src homology 3 domain of Grb2, whose recruitment to the EGFR is mediated h, and degradation of EGFR was stimulated with 100 ng/ml EGF for the indicated times. The cell lysates were subjected to SDS-PAGE analysis. The proteins were probed with specific antibodies as indicated. C, band intensity from 15, 30, and 60 min was quantified with AlphaEaseFc 4.0 software. The graph was acquired from three independent experiments. The data represent the means Ϯ S.E. *, p Ͻ 0.007 compared with control. D, DU145 cells were transfected with siRNA targeting GFP (as negative control), Rab5A, Rab5B, or Rab5C. 24 h post-transfection, the cells were starved overnight, and degradation of EGFR was stimulated with 100 ng/ml EGF in the presence of cycloheximide for 5 h. The cell lysates were processed as described above. The proteins were probed with specific antibodies as indicated. Band intensity was quantified with AlphaEaseFc 4.0 software. The graph was acquired from three independent experiments. The data represent the means Ϯ S.E. *, p Ͻ 0.006 compared with control. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
by phosphorylation at Tyr-1068 and Tyr-1086 of EGFR (27,28). We therefore asked whether the delayed degradaton of EGFR as a result of Rab5A or 5B depletion correlates with altered EGFR phosphorylation and ubiquitylation. By immunoprecipitating EGFR after stimulation for different periods of time and probing ubiquitylated EGFR with anti-ubiquitin antibody, we found that the relative levels of EGFR ubiquitylation over time is not altered by depletion of Rab5A or 5B (Fig. 5A). Consistently, depletion of Rab5A or 5B did not suppress the phosphorylation of EGFR at Tyr-1045 or Tyr-1068 ( Fig. 5B and supplemental Fig. S5), nor did the silencing of Rab5 isoforms interfere with the recruitment of c-Cbl or FIGURE 2. Determination of the abundance of endogenous Rab5 isoforms in HeLa and DU145 cells. 5 or10 g of total cell lysates from siRNA-treated DU145 or HeLa cells were subjected to SDS-PAGE analysis alongside purified His-Rab5 isoforms ranging from 1.25 to 10 ng as standards. The proteins were probed with isoform-specific antibodies and developed with ECL. Band intensity was quantified with AlphaEaseFc 4.0 software (supplemental Fig. S2). A, the experiment was repeated three times. The graph shows the relative abundance of the three isoforms in HeLa cells. The data represent the means Ϯ S.E. of three independent experiments. B, the abundance of Rab5 isoforms in DU145 was determined as described above. The graph shows the relative abundance of the three isoforms in DU145 cells. The data represent the means Ϯ S.E. of three independent experiments. Grb2. However, it is unclear whether depletion of Rab5A or 5B alters the phosphorylation of other tyrosine or serine sites of the EGFR and/or obstructs its access to other key adaptors. Further studies will be needed to reveal a more detailed phosphorylation map of EGFR in response to Rab5 isoform silencing.
To further investigate how the loss of individual isoforms changes endosomal dynamics, we examined the subcellular distribution of EEA1, a FYVE domain-containing protein known to associate with the early endosomal compartment (4, 6, 29), and Hrs, which also localizes on early endosomal membranes and transfer cargos with ubiquitin moieties to downstream multivesicular body sorting machineries (30). In the mean time, the transport of EGFR was visualized by either indirect EGFR immunostaining or by using Alexa 488-labeled EGF. Consistent with the immunoblotting data, we observed stronger EGF (or EGFR) signals in Rab5A or 5B-depleted cells (Fig. 6). The association of EEA1 with membrane structures was not impaired by loss of individual Rab5 isoforms, nor was the distribution of Hrs and CD63, a late endosomal marker ( Fig. 6 and supplemental Fig. S6). Similar to control cells, the EGF-EGFR complex in isoform-silenced cells entered

. Internalization of EGFR and transferrin receptors are not inhibited by individual Rab5 isoform depletion.
A, the rate of EGFR internalization was examined by labeling cell surface EGFR with biotin after stimulation briefly with EGF. Magnetic streptavidin beads then were used to pull-down biotin labeled surface EGFR. After washes, bound EGFR were released from the beads with SDS sample buffer and then analyzed with SDS-PAGE. B, the intensity of the bands were quantified with AlphaEaseFc 4.0 software and plotted as a percentage of EGFR remaining on cell surface over time. The data represent the means Ϯ S.E. of three independent experiments. C, control or specific-siRNA transfected cells were starved for 1 h to deplete endogenous Tfn before the uptake assay. Next, the cells were prebound with Alexa 488-Tfn for 1 h at 4°C. Internalization was then initiated by replacing the binding medium with prewarmed internalization medium and incubating at 37°C for 5 min. At the end of uptake, the cells were washed and fixed for confocal microscopy analysis. The scale bar represents 10 m.

FIGURE 5. Phosphorylation and ubiquitylation of EGFR is not altered by silencing of Rab5 isoforms.
A, control or Rab5 isoform-silenced HeLa cells were starved and stimulated with EGF at 37°C for the indicated times. The cell lysates were then subjected to EGFR immunoprecipitation (IP) and immunoblotting with panubiquitin antibody to detect the ubiquitylated EGFR. The blots were reprobed with EGFR antibody to show the total EGFR levels in the immunoprecipitates. B, control or Rab5 isoform-silenced cells were starved and then stimulated with EGF at 4°C for 1 h. The cell lysates were then subjected to EGFR immunoprecipitation and immunoblotting with antibodies as indicated. Co-immunoprecipitated adaptor proteins were detected with specific antibodies as indicated. OCTOBER 30, 2009 • VOLUME 284 • NUMBER 44

JOURNAL OF BIOLOGICAL CHEMISTRY 30333
EEA1-positive compartments within 10 min (Fig. 6A). At later time points, most EGFR-containing vesicles was still EEA1positive in Rab5A-or 5B-silenced cells, whereas a substantial subset of EGFR in control and Rab5C-depleted cells had exited from early endosomal compartments (Fig. 6, B and C). Similar results were obtained with Hrs co-immunostaining (supplemental Fig. S6A). 45 min after endocytosis, EGFR signals had almost completely disappeared from Hrs-positive compartments in control or Rab5C knockdown cells but still partially co-localized with Hrs in Rab5A or 5B knockdown cells (supplemental Fig. S6A). These results suggest that depletion of Rab5A or 5B slows down the progression of EGFR from early to late endosomal compartments, thereby causing the delay of EGFR degradation. Even so, Rab5A-or Rab5B-silenced cells did eventually complete the degradation of EGFR, suggesting that they can still overcome this blockage, presumably via residual activity from incompletely silenced Rab5 isoform, compensatory activities from Rab5 isoforms not targeted by siRNA or alternative mechanisms independent of Rab5.
Rin1 Preferentially Associates with Rab5A, Which Facilitates the Activation of Rab5A in Response to EGF Stimulation-Previous studies have shown that Rab5 can be activated by hepatocyte growth factor stimulation (31) and that active Ras, a consequential effector of EGFR signaling, potentiates the activity of Rab5A in vitro (32). Both studies suggest that the function of Rab5 can be regulated by receptor signaling. In an effort to better understand why Rab5A displays more potency in the EGFR degradation pathway, we tested whether Rab5 isoforms are differentially activated in response to EGF stimulation. We took advantage of a Rabaptin-5 R5BD-based GST pulldown assay (33) to study the activation Rab5 isoforms in vivo. We found that after 5 min of EGF treatment, more Rab5A was bound to GST-R5BD, indicating more activation of Rab5A, whereas Rab5B and Rab5C activation did not increase in response to EGF at this time point (Fig. 7). This result further supports the hypothesis that Rab5A is more potent in regulating the trafficking of EGFR because its activation is more responsive to the EGF stimulation. Our previous work shows that Rin1, a Vps9 domain-containing Rab5 exchange factor, relays EGFR signals to Rab5 activation (32). Also, Rin1 has an Src homology 2 domain that couples Rab5 activation to the proximity of EGFR (34). Recently, we found that Rin1 works in concert with the ESCRT-0 (endosomal sorting complex required for transport 0) complex (signal transducing adapter molecule 1/Hrs) in down-regulation of EGFR (1). These data make Rin1 a prime target in Rab5-mediated EGFR trafficking. Therefore, we tested whether Rin1 shows any specificity toward Rab5 isoforms. Dominant negative (S34N) and GTP hydrolysisdeficient constitutively active (Q79L) forms of Rab5 isoforms were used to evaluate Rin1 specificity. We carried out a coimmunoprecipitation study with cells transiently co-expressing one of the Rab5 mutant isoforms along with Rin1. Characteristic of exchange factors, Rin1 binds preferentially to dominant negative Rab5, which is primarily GDP-bound. Interestingly, we found that Rin1 bound to Rab5A with higher affinity than to Rab5B or Rab5C (Fig. 8A). Another Rab5 exchange factor, GAPex-5 (human RME-6) (35), was also tested for Rab5 isoform specificities, and no apparent selectivity was found (data FIGURE 6. Trafficking of EGFR into EEA1-positive compartments is not altered by depletion of Rab5 isoforms, but exit from early endosomal compartments is delayed. HeLa cells were transfected with control or Rab5 isoformspecific siRNA. A and B, 48 h post-transfection, the cells were starved, incubated with Alexa 488-EGF for 1 h at 4°C, and then allowed to endocytose for 10 min (A) and 25 min (B) at 37°C. The cells were then fixed and immunostained for EEA1 (2°antibody: Alexa 594-IgG). The scale bar represents 10 m. C, co-localization between EGF/ EGFR and EEA1 was quantified with ImageJ, and the means Ϯ S.E. of the Pearson's coefficient from seven images/condition are illustrated in the graph. not shown). To determine whether the activation of Rab5A by EGF stimulation is Rin1-dependent, HeLa cells transfected with Rab5A and scramble or Rin1-specific siRNA were subjected to R5BD pulldown assay as described above. We found that in the absence of Rin1, the activation of Rab5A in response to EGF stimulation is inhibited (Fig. 8B). These results suggest that Rab5 isoforms may be differentially regulated by various exchange factors, whose abundance, subcellular localization, and responses to extracellular stimulation can potentially direct their favored isoforms to regulate certain pathways but not others.

DISCUSSION
The focus of this study is to investigate the functional significance of each Rab5 isoform. Work from our lab and others has demonstrated the many facets of Rab5 in regulating endocytosis, signal transduction, and cytoskeletal dynamics (36 -39). Here, we provide new evidence suggesting that Rab5 isoforms have the distinct capacity to coordinate the trafficking of EGFR rather than acting only as redundant mechanisms.
Rab5A Is the Predominant Isoform in EGFR Degradation Pathway-After effective down-regulation of Rab5 isoforms with siRNAs, we observed that the trafficking of EGFR was delayed by Rab5A and to a lesser extent by 5B suppression but barely affected by Rab5C silencing in both HeLa and DU145 cells. This result suggests that both endogenous Rab5A and 5B take part in EGFR trafficking, although their efficacy in mediating this process and sites of action may vary. We then looked into the expression level of endogenous Rab5 isoforms in HeLa and DU145 cells and found that Rab5A is no more enriched than Rab5B or Rab5C, indicating that the delayed degradation could not simply be attributed to overall reduction of Rab5 levels in cells. The retardation of EGFR trafficking was further analyzed by light microscopy to clarify the subcellular compartments retaining EGFR in the absence of each Rab5 isoform. We showed that EGFR entered early endosomal compartments positive for EEA1 within 10 min of internalization regardless of Rab5 isoform depletion. At later time points, most of the receptors had progressed out of EEA1-or Hrs-positive vesicles in control and Rab5C-depleted cells yet remained partially localized to early endosomal compartments in Rab5A-and Rab5B-depleted cells. Although the loss of Rab5A or 5B impaired the trafficking of EGFR, it did not appear to alter the normal covalent modifications of EGFR essential for its proper sorting. Taking these FIGURE 7. Rab5A is activated in response to EGF stimulation. HeLa cells were transfected with GFP-Rab5A, 5B, or 5C. 24 h post-transfection, the cells were starved for 3 h and then stimulated with EGF for 5 min. The cell lysates were collected and subjected to R5BD-GST pulldown assay as described under "Experimental Procedures." Bound GFP-Rab5 isoforms were then analyzed with SDS-PAGE and immunoblotting. Band intensity from blots was quantified and normalized to total GFP-Rab5 isoform signals from whole cell lysates (WCL). The adjacent graph represents the means Ϯ S.E. of the relative activation of Rab5 isoforms from three independent experiments. *, p Ͻ 0.04. FIGURE 8. Rin1 preferentially binds Rab5A to facilitate the degradation of EGFR. A, HEK293 cells were co-transfected with V5-tagged Rin1 along with constitutively active (Q79L) or negative (S34N) mutants of GFP-Rab5A, Rab5B, or Rab5C. The cell lysates were precleared with centrifugation and then immunoprecipitated with anti-GFP antibody. The immunoprecipitates were then separated by SDS-PAGE and blotted with antibodies as indicated above. The total lysates were also analyzed to show the expression level of recombinant proteins in each sample. B, HeLa cells were co-transfected with GFP-Rab5A and scrambled or Rin1-targeting siRNA. 24 h post-transfection, the cells were starved for 3 h and then stimulated with EGF for 5 min. The cell lysates were collected and subjected to R5BD-GST pulldown assay as described under "Experimental Procedures." Bound GFP-Rab5 isoforms were then analyzed with SDS-PAGE and immunoblotting. Band intensity from blots was quantified and normalized to total GFP-Rab5A signals from whole cell lysate (WCL). The adjacent graph represents the means Ϯ S.E. of the relative activation of Rab5A from three independent experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. OCTOBER 30, 2009 • VOLUME 284 • NUMBER 44 results together, following the down-regulation of the endogenous Rab5 isoforms, we observed a differential contribution of Rab5 isoforms in EGFR trafficking. In agreement with silencing results, overexpression of Rab5A appeared to accelerate EGFR trafficking more effectively than other isoforms. In fact, previous work from our lab has also demonstrated a close connection between Rab5A and EGFR-related pathways. Barbieri et al. (20) have shown that 125 I-EGF internalization and EGF-stimulated horseradish peroxidase uptake in fibroblasts are more upregulated by overexpression of Rab5A compared with Rab5B or 5C. It is to be noted that the phenotypes of Rab5B suppression and overexpression did not coincide. In the absence of Rab5B, EGFR degradation is moderately impaired, but overexpression of Rab5B hardly strikes any difference. One possible explanation is that the type of cells we used for overexpression does not have sufficient and appropriate combinations of Rab5B regulators and/or effectors to transduce its activity into phenotypes, whereas RNA interference ablation of Rab5B directly perturbs the homeostasis maintained by endogenous Rab5B, which is more easily detected. We propose that endogenous Rab5B may work in conjunction or in sequence with Rab5A to facilitate the trafficking of EGFR. Interestingly, when we employed an in vitro system to evaluate the strength of Rab5 isoforms in promoting vesicle fusion, our preliminary data indicated that purified recombinant Rab5A was more potent in activating endosome-endosome fusion, and loss of endogenous Rab5A in cell membrane fractions lead to decreased fusion activities (data not shown). This result not only further supports the concept of differential potency from Rab5 isoforms but also suggests that Rab5B participates in EGFR trafficking in a different manner. Recent studies point to a role of Rab5-GTP in coordinating membrane tethering and fusion with cytoskeletalbased organelle motility (9,40). Spatial trafficking of the early endosomes from the cell periphery to the cell center is critical for their ability to generate endosomal carrier vesicles that bud from early endosomes and fuse with late endosomes or mature into late endosomes (40,41). It is possible that Rab5B, instead of promoting endosome-endosome fusion, governs this aspect of the EGFR trafficking. In fact, we did, on multiple occasions, notice that internalized EGF-EGFR vesicles were scattered to sites more adjacent to the plasma membrane in Rab5B knockdown cells than in control or Rab5A knockdown cells. These data suggest that after EGFR is internalized, the inward movement of vesicles is likely impeded when Rab5B is repressed. More studies will be needed to clarify the involvement and activity of different Rab5 isoforms in modulating the motility of endosomes.

Rab5A Preferentially Interacts with Rin1, a Multi-domain Guanine Nucleotide Exchange Factor That Binds EGFR and
Regulates the EGFR Degradation-The question remains unclear as to what causes the difference among Rab5 isoforms in the EGFR degradation pathway. Overexpressed Rab5 isoforms in cells have similar subcellular localization, suggesting that all three isoforms are properly targeted to membrane structures, and Rab5A is more active in EGFR trafficking pathway not because of better membrane association. However, it is yet to be determined whether endoge-nous Rab5 isoforms are localized differently. Another possibility would be differential regulation of the isoform activities by GEFs and GAPs. At least nine different Rab5 GEFs have been identified, and seven of them are proven to have GEF activity in vitro (7,8,32,35,(42)(43)(44)(45). In addition to vesicle fusion, Rab5 can also control the motility of organelles along microtubules (9,40), actin remodeling (37)(38)(39), and participation in proliferative cell signaling pathways (23,36). It is reasonable to speculate that the wide range of Rab5 exchange factors contribute to the complexity of these proposed functions. Indeed, we tested the binding specificity of some of these GEFs toward Rab5 isoforms and found that Rin1 preferentially binds to Rab5A, whereas GAPex-5 showed no significant specificities toward isoforms. Because we tested the activation of Rab5 isoforms in response to EGFR stimulation, Rab5A displayed strong and acute activation, which is in good correlation with its potency in EGFR trafficking. More importantly, the activation of Rab5A upon EGF stimulation is inhibited by the silencing of Rin1. That is, the preferential interaction between Rab5A and Rin1 appears to control the ability and sensitivity of Rab5A in regulating EGFR degradation pathway in HeLa cells. However, it remains unclear whether the preferential interaction between Rin1 and Rab5A selectively regulates the trafficking of only a certain receptors. Barbieri et al. (34) showed that Rin1 associates with several signaling receptors, but not cargo receptors, such as Tfn receptor and mannose receptor. Moreover, overexpression of Rin1 does not alter the uptake of Tfn, nor was the recycling of Tfn in HeLa cells affected by Rin1 depletion (supplemental Fig. S7). These results suggest that Rin1 may provide selectivity for Rab5A in regulation of the trafficking of specific receptors. Because more functions have been shown to involve Rab5, it is possible that different Rab5 isoforms play more significant roles in distinct functions. In fact, preliminary studies examining Rab5-mediated cell migration implicates Rab5C to be more engaged to cell migratory activities than other Rab5 isoforms. 3 In agreement with this finding, a recent report indicated that Rab5C is the only isoform expressed during gastrulation of zebrafish and mediates Wnt11-controlled mesendodermal cell cohesion and migration (46). Ulrich et al. (46) proposed that Wnt11 promotes Rab5Cmediated endocytosis and recycling of E-cadherin to regulate the dynamics of E-cadherin turnover at the plasma membrane. Alternatively, Wnt11 might regulate Rab5-dependent actin remodeling, which in turn could affect the adhesive and cohesive properties of mesendodermal cells.
It is important to point out that lower eukaryotic model organisms such as Drosophila and Caenorhabditis elegans have only one Rab5 paralogue, and loss of its function has been shown to result in embryonic lethality (47,48), whereas three isoforms are identified for higher organisms (i.e. human and zebrafish). The multiplication of Rab5 paralogues from its ancestral origin implicates continued evolution and specialization of the early endosomal systems. The most defining biological attributes of higher organisms are the development of nervous and immune systems, both of which require highly organized endocytic networks to manage the intricacy of their respective functions. In the nervous system, not only is Rab5 shown to associate with vesicles involved in cycling of neurotransmitters during synaptic transmission (47), but its activity appears to initiate neuronal cell differentiation and dendrite branching (33,49,50) and participates in the delivering of neurotrophic "signaling endosomes" via dynein and kinesin (50,51). Several neurodegenerative diseases are linked to faulty axonal transport (52). The juvenile amyotrophic lateral sclerosis-associated gene ALS2, encoding the protein Alsin, whose Vps9 domain regulates the activation of Rab5 (43). Depletion mutation of ALS2 within its Vps9 domain has been identified to cause hereditary spastic paraplegia (53), further stressing the role of Rab5 in neuronal pathogenesis. Recent evidence also revealed multiple endocytic mechanisms governing antigen presentation (54). Depending upon the nature of the antigens as well as the endocytosis receptors, antigens can be routed to phagosomes or endosomes for peptide processing and loading. The molecular basis that assigns each Rab5 isoform to a specific function is still unclear. In this study, we provided new evidence suggesting that Rab5 isoforms have preferences toward their exchange factors and certain pathways. Given the need for flexible yet precise regulation of these neuronal and immune responses, it is reasonable to anticipate that Rab5 isoforms differentially integrate their binding specificities and/or tissue distribution into the regulation of multiple pathways. It will be of great importance to dissect the roles of Rab5 isoforms in these pathways by means of individual Rab5 isoform ablation in the future.