CFTR (cystic fibrosis transmembrane conductance regulator) is expressed in the apical membrane of epithelial cells. Cell-surface CFTR levels are regulated by endocytosis and recycling. A number of adaptor proteins including AP-2 (μ2 subunit) and Dab2 (Disabled-2) have been proposed to modulate CFTR internalization. In the present study we have used siRNA (small interfering RNA)-mediated silencing of these adaptors to test their roles in the regulation of CFTR cell-surface trafficking and stability in human airway epithelial cells. The results indicate that μ2 and Dab2 performed partially overlapping, but divergent, functions. While μ2 depletion dramatically decreased CFTR endocytosis with little effect on the half-life of the CFTR protein, Dab2 depletion increased the CFTR half-life ~3-fold, in addition to inhibiting CFTR endocytosis. Furthermore, Dab2 depletion inhibited CFTR trafficking from the sorting endosome to the recycling compartment, as well as delivery of CFTR to the late endosome, thus providing a mechanistic explanation for increased CFTR expression and half-life. To test whether two E3 ligases were required for the endocytosis and/or down-regulation of surface CFTR, we siRNA-depleted CHIP [C-terminus of the Hsc (heat-shock cognate) 70-interacting protein] and c-Cbl (casitas B-lineage lymphoma). We demonstrate that CHIP and c-Cbl depletion have no effect on CFTR endocytosis, but c-Cbl depletion modestly enhanced the half-life of CFTR. The results of the present study define a significant role for Dab2 both in the endocytosis and post-endocytic fate of CFTR.
Clathrin-mediated endocytosis is responsible for the internalization of a wide variety cell-surface molecules, including receptors, signalling molecules and channel proteins. Internalization of these proteins is mediated by cytoplasmic adaptor molecules that recruit the cell-surface molecules into clathrin-coated pits by simultaneously binding to clathrin and to internalization signals in the cytoplasmic domains of cell-surface protein (reviewed in ). The best-characterized adaptor protein is AP-2, a tetramer composed of α2, β2, μ2 and σ2 subunits . AP-2 binds to some internalization signals via its μ2 subunit  and to clathrin via its β2 subunit , whereas the α2 subunit binds accessory proteins involved in coated pit assembly .
AP-2 is not the only adaptor that is important for internalization of cell-surface molecules. Another example is Dab2 (Disabled-2), an adaptor that has been shown to facilitate the internalization of the LDLR [LDL (low-density lipoprotein) receptor] . The selective use of different adaptors is mediated, at least in part, by the different internalization signals found in the receptors. AP-2 recognizes a YXXΦ cytoplasmic tail signal, where X is any amino acid and Φ is a large hydrophobic residue, as found in the TR [Tfn (transferrin) receptor] , whereas Dab2 recognizes FXNPXY signals present in the LDLR . It is now clear that multiple adaptor molecules are used selectively for different cell-surface molecules, thus making the clathrin-mediated endocytosis much more complicated than originally envisioned (reviewed in ).
CFTR (cystic fibrosis transmembrane conductance regulator) is a cAMP-activated chloride channel expressed in airway, intestinal and a number of other epithelial tissues [10–12]. CFTR is expressed on the apical surface of epithelia and its surface expression is regulated by endocytosis and recycling [13–15]. Efficient endocytosis of CFTR has been shown to require C-terminal cytoplasmic tail internalization signals, Y1424DSI and L1430L [16–18]. Furthermore, the tyrosine-based signal has been shown to interact with the μ2 subunit of AP-2 .
Other proteins have also been shown to be involved in CFTR endocytosis. These include myosin VI and c-Cbl (casitas B-lineage lymphoma proto-oncogene), an E3 ligase [20,21]. Myosin VI, a motor travelling towards the minus end of the actin filaments, is targeted to clathrin-coated structures by its interactions with Dab2 . Moreover, CFTR immunoprecipitation studies in human airway cells and intestinal epithelial cells have demonstrated that Dab2 co-immunoprecipitates in a complex with myosin VI and AP-2 [20,23], suggesting that it is a CFTR-binding partner in the clathrin-coated pit. In intestinal cells, CFTR was shown to interact directly with the α2 subunit of AP-2, but not Dab2, whereas Dab2 interacts directly with the α2 subunit of AP-2 . Interestingly, Dab2-knockout mice have higher levels of CFTR expression in jejunum sections , suggesting that Dab2 regulates CFTR surface expression.
In the present study, we wanted to determine the roles of AP-2 and Dab2 in human airway epithelial cells and define what their different functions were. The goal was to determine their relative importance in CFTR endocytosis, but more importantly, to determine whether either played a role in cell-surface CFTR down-regulation. Using polarized human epithelial cells for the endocytosis assays, we found that siRNA (small interfering RNA) KD (knockdown) of either the μ2 subunit of AP-2 or Dab2 dramatically inhibited CFTR endocytosis. In fact, both KDs had very similar inhibitory effects on CFTR endocytosis, indicating that both adaptors contribute to CFTR internalization. Analysis of the half-life of CFTR protein, however, revealed striking differences in CFTR stability in the different KDs. Depletion of AP-2 (μ2) had little effect on the half-life of CFTR, whereas Dab2 depletion increased the half-life ~3-fold. This suggested that although both adaptors are necessary for the endocytosis of CFTR, Dab2 is also a part of the post-endocytic trafficking machinery that directs CFTR to the degradative pathway.
CFBE41o-WT (WT is wild-type; expressing WT-CFTR) cells were cultured as described previously . For the functional analysis in Ussing chambers and endocytosis assays using polarized monolayers, CFBE41o-WT cells were seeded on to 12-mm Transwell® filters (Costar, Corning) and cultured on an air–liquid interface for 5 days before analysis, as described previously .
Antibodies and chemicals
Anti-CFTR mouse monoclonal [clone 24-1 (A.T.C.C.) and MM13-4 (Millipore)] and rabbit polyclonal (anti-NBD1) antibodies were used as described previously . Mouse anti-AP50/μ2 and anti-EEA1 (early endosome antigen 1) antibodies were from BD Transduction Laboratories. A rabbit polyclonal antibody against M6PR (mannose 6-phosphate receptor) and c-Cbl were purchased from Abcam. A rabbit polyclonal antibody against Dab2 was from Santa Cruz Biotechnology. A rabbit polyclonal anti-CHIP [C-terminus of the Hsc (heat-shock cognate) 70-interacting protein] antibody was from Thermo Scientific. A polyclonal anti-actin antibody was purchased from Sigma–Aldrich. Alexa Fluor® 488–labelled goat anti-mouse IgG antibody, Alexa Fluor® 594-labelled goat anti-rabbit IgG antibody and Alexa Fluor® 594-conjugated human Tfn were from Invitrogen. HRP (horseradish peroxidase)-labelled goat anti-mouse IgG antibody and HRP-labelled goat anti-rabbit IgG antibody were from Bio-Rad. The SuperSignal West Pico chemiluminescence substrate was from Pierce. All other chemicals were from Sigma–Aldrich or Fisher Scientific.
CFTR, μ2, Dab2, c-Cbl and CHIP protein levels in control or siRNA-depleted samples were determined as described previously . Briefly, cells were lysed in RIPA buffer [50 mM Tris/HCl (pH 8.0), 1% Nonidet P40, 0.5% deoxycholate, 0.1% SDS, 150 mM NaCl and Complete™ Protease Inhibitor (Roche)]. Proteins were separated by SDS/PAGE and transferred on to PVDF membranes (Bio-Rad). CFTR was detected with a monoclonal antibody (MM13-4, 1:500 dilution; Millipore). Dab2 and μ2 were detected with a polyclonal anti-Dab2 antibody (1:200 dilution; Santa Cruz Biotechnology) and a monoclonal anti-μ2 antibody (1:1000 dilution; BD Transduction Laboratories) respectively. c-Cbl and CHIP were examined using antibodies against c-Cbl (1:500 dilution; Abcam) and CHIP (1:200 dilution, Thermo Scientific). HRP-labelled secondary antibodies and Super; Signal West Pico chemiluminescence substrate (Pierce) were used for visualization of specific protein bands. Densitometry was performed using ImageJ software (NIH).
Co-immunoprecipitation of CFTR and Dab2
CFTR and Dab2 were co-immunoprecipitated from CFBE41o-WT cells expressing WT-CFTR. CFBE41o- cells with no detectable CFTR expression were tested as a negative control. Briefly, cells were lysed in a buffer containing 150 mM NaCl, 50 mM Tris/HCl (pH 8.0), 1% Nonidet P40 (Sigma) and Complete™ Protease Inhibitor (Roche). After centrifugation at 15000 g for 15 min at 4°C, the soluble fraction (supernatant) was incubated with Protein G–agarose beads (Roche) pre-loaded with anti-CFTR (24-1) or anti-Dab2 antibodies at 4°C for 2 h. Proteins were eluted in Laemmli sample buffer and separated by SDS/PAGE followed by Western blot transfer to PVDF membranes as described previously . CFTR was then detected with an anti-CFTR antibody (MM13-4) in the anti-Dab2 precipitated samples, and Dab2 was detected in the anti-CFTR precipitated samples using an anti-Dab2 antibody.
Indirect immunofluorescence microscopy was performed as described previously . Briefly, cells were grown on glass coverslips, fixed in 4% paraformaldehyde/PBS and permeabilized with 0.1% Triton X-100/PBS for 5 min, washed three times for 2 min each with PBS, and then blocked with 2.5% goat serum/PBS. Cells were incubated with primary antibodies diluted in blocking solution for 2 h. Following washing steps, the secondary antibodies (1:500) were applied and incubated for 45 min and mounted with Vectashield®/DAPI (4′,6-diamidino-2-phenylindole) (Vector Labs). All incubations and washing steps were at room temperature (approximately 25°C) unless otherwise stated. Microscopy was performed using a Leitz epifluorescence microscope equipped with a step motor, filter-wheel assembly (Ludl Electronics Products) and an 83000-filter set (Chroma Technology). Images were obtained with a Sensys-cooled charge-coupled high-resolution camera (Photometrics). IpLab Spectrum software (Signal Analytics) was used for image acquisition. For the ammonium chloride experiments, cells were treated with 5 mM ammonium chloride for 16 h prior to indirect immunofluorescence microscopy using anti-CFTR and anti-M6PR antibodies.
For confocal microscopy, CFBE41o-WT cells were grown on a Transwell® membrane support for 5 days and then processed for immunofluorescent staining as described above. Confocal imaging was performed using a Nikon 2000U inverted microscope equipped with a PerkinElmer UltraVIEW ERS 6FE-US spinning disk laser apparatus. Images were processed with Volocity 5 software (Improvision).
Cell-surface CFTR was biotinylated as described previously .
Metabolic pulse–chase assay
siRNA-mediated depletion of Dab2, μ2, c-Cbl and CHIP
siRNA duplexes corresponding to non-conserved regions of human Dab2, μ2 and c-Cbl were purchased from Qiagen. The specific sequence 5′-TAGAGCATGAACATCCAGTAA-3′ was chosen as a target for the siRNA depletion of Dab2. The targeting sequence for knock down of μ2 was 5′-TGCCATCGTGTGGAAGATCAA-3′. The targeting sequence for c-Cbl was 5′-CCCGCCGAACUCUCAGAUATT-3′. The double-stranded non-silencing control siRNA sequence 5′-AATTCTCCGAACGTGTCACGT-3′, which has no significant homology with any other genes, was also purchased from Qiagen. Depletion of human STUB1 (encoding CHIP) was achieved by using siRNA oligonucleotides from Dharmacon as described previously . Transfection of siRNA oligonucleotides was performed using siLentFect™ lipid reagent (Bio-Rad) according to the manufacturer's protocol. Briefly, cells at ~70–80% confluency were transfected with the optimized transfection mixture. After 24 h incubation at 37°C, the transfection mixture was replaced with fresh cell culture medium. Experiments were conducted 3–6 days after transfection. The depletion efficiencies of individual genes were assessed by Western blot analysis.
Ussing chamber analyses
Short-circuit current (ISC) was measured under voltage-clamp conditions using MC8 voltage clamps and P2300 Ussing chambers (Physiologic Instruments) as described previously . Monolayers were initially bathed on both sides with identical Ringer's solutions containing 115 mM NaCl, 25 mM NaHCO3, 2.4 mM KH2PO4, 1.24 mM K2HPO4, 1.2 mM CaCl2, 1.2 mM MgCl2 and 10 mM D-glucose (pH 7.4). Bath solutions were vigorously stirred and gassed with 95%O2/5% CO2. ISC measurements were obtained using an epithelial voltage clamp (Physiologic Instruments). A 1 s, 3 mV pulse was imposed every 10 s to monitor resistance calculated using Ohm's law. Where indicated, the mucosal bathing solution was changed to a low Cl− solution containing 1.2 mM NaCl and 115 mM sodium gluconate, and all other components as above. Amiloride (100 μM) was added to block residual Na+ current, followed by the CFTR agonists forskolin (20 μM, Calbiochem) and genistein (50 μM, Sigma–Aldrich) as indicated (minimum 5-min observation at each concentration). CFTRInh-172 (10 μM) was added to the apical bathing solution at the end of experiments to block CFTR-dependent ISC. All chambers were maintained at 37°C during experiments.
Following siRNA depletion of Dab2, CFBE41o-WT cells were seeded on to coverslips and incubated at 37°C. At 72 h after siRNA transfection, the cells were serum-starved for 2 h and then pulsed with Alexa Fluor® 594-conjugated Tfn (Alexa 594–Tfn, 20 μg/ml) for 5 min at 37°C. After multiple washing steps with ice-cold PBS to remove unbound Alexa 594–Tfn and inhibit transport, the coverslips were incubated with complete growth medium pre-warmed to 37°C for the time periods indicated. Then the coverslips were quickly washed with ice-cold PBS and fixed in 4% formaldehyde, mounted and visualized using fluorescence microscopy.
Data analysis and statistics
All experiments were repeated at least three times, and the data were expressed as the means±S.D. Statistical analysis was performed using Student's t test (two-tailed) in Microsoft Excel, and significance was determined at the P<0.05 level.
Depletion of AP-2 (μ2 subunit) and Dab2 in airway epithelial cells increases total CFTR levels
In order to determine the roles of AP-2 and Dab2 in CFTR trafficking at the cell surface, we first examined how siRNA KD of each of these adaptors affected CFTR expression in human airway epithelial cells (CFBE41o-WT). Using different concentrations of siRNA, we determined the maximum depletion conditions for each adaptor and their effects on total CFTR expression. Figure 1(A) illustrates that more than 90% of μ2 was depleted using siRNA KD, and this results in a 2–3-fold increase in the CFTR C band, the mature form of CFTR. Figure 1(B) shows that Dab2 levels were reduced more than 95% of the control and the CFTR C band was increased 5–6-fold. The core-glycosylated form (B band) of CFTR was also slightly increased upon Dab2 KD, whereas this was not seen in the μ2 KD. The results suggest that although depletion of μ2 increased total CFTR levels, the Dab2 depletion had a more pronounced effect on CFTR expression (Figures 1A and 1B). Interestingly, depletion of both adaptors did not have an additive effect, suggesting that the two adaptors were not acting independently (Figure 1C). Next, we performed co-immunoprecipitation experiments in order to confirm the interaction between CFTR and Dab2 by immunoprecipitating either CFTR and Western blotting for Dab2 or immunoprecipitating Dab2 and immunoblotting for CFTR (Figure 1D). The results confirm that CFTR and Dab2 are present in the same complex.
Increased CFTR levels in CFBE41o-WT cells following μ2 and Dab2 depletion and confirmation of CFTR and Dab2 interactions
During its biogenesis, CFTR is first synthesized as a core glycosylated B band in the endoplasmic reticulum and then is further modified to the mature glycoslylated C band as it passes through the Golgi complex. Because the total pool of CFTR was increased by the depletion of the adaptors, and the effects on the fully processed band C CFTR were the most pronounced, we next examined how depletion of μ2 and Dab2 affected the surface pool of CFTR. To test this, we performed cell-surface biotinylation and immunocytochemistry . We first labelled the cell-surface CFTR with biotin and measured the levels of biotinylated CFTR following either μ2 or Dab2 depletion. The results indicate that the cell-surface CFTR was increased ~3- and ~5-fold when μ2 and Dab2 were depleted respectively (Figures 2A and 2B). The increase of the cell-surface CFTR is comparable with that of the total pool. To validate that the increased CFTR is on the cell surface, we performed confocal microscopy on polarized CFBE41o- cells grown on permeable supports (Figure 2C). Although μ2 KD increased the total (Figure 2C, Top) and the surface (Figure 2C, Side) pool, the Dab2 depletion had a much more pronounced effect, suggesting that Dab2 might be affecting more than one step in the pathway. Importantly, these results established significant roles for AP-2 and Dab2 in the regulation of the cell-surface and consequently the total CFTR pools in airway epithelial cells. These results are consistent with the idea that these adaptors regulate the endocytosis of CFTR. Furthermore, the more significant changes in CFTR surface levels following Dab2 depletion suggested divergent functions for these adaptors in the post-endocytic compartments. To confirm this, we designed CFTR endocytosis, half-life, and functional and morphological studies in order to follow the fate of CFTR after depletion of the adaptors.
Increased cell-surface CFTR levels following μ2 and Dab2 depletion
AP-2 (μ2) and Dab2 are both required for CFTR endocytosis
Previous studies have shown that disruption of AP-2 function in non-polarized cells inhibits CFTR endocytosis [19,23]. In the present study we examined this question in polarized human airway epithelial cells and directly compared the roles of AP-2 and Dab2 in CFTR endocytosis from the apical cell surface. First, we depleted μ2 or Dab2 and cultured the cells as polarized monolayers on filter supports. Next, we tested how depletion of μ2 or Dab2 affected CFTR internalization using a surface biotinylation assay to monitor endocytosis . Immunoblotting experiments revealed that both adaptors were depleted >90% in all experiments. Importantly, the results indicated that both μ2 and Dab2 equally contribute to CFTR endocytosis (Figure 3). Specifically, in 2.5 min, 19.9±2.1% (mean±S.D.) of the cell-surface CFTR was internalized in the control cells, whereas in μ2-depleted cells, CFTR internalization was reduced to 9.0±1.8%, indicating a 55% decrease. Dab2 depletion had a similar effect (9.5±1.3% endocytosis in 2.5 min), indicating a 53% inhibition of CFTR endocytosis. To determine whether the effects of μ2 and Dab2 depletion were additive, we depleted both adaptors (Figure 3, bottom panels). In the double-KD experiments, CFTR internalization was similar to the single μ2- or Dab2-depletion experiments (8.1±2.5% in 2.5 min). These results suggested that the two adaptors do not work independently and both are necessary for efficient CFTR endocytosis.
Reduced CFTR endocytosis rates following μ2, Dab2 or μ2+Dab2 depletion
Dab2 depletion dramatically increases the half-life of the CFTR protein
Following endocytosis, CFTR either recycles to the cell surface or enters the degradative pathway . To identify the roles of μ2 and Dab2 in the intracellular trafficking of CFTR following endocytosis, we tested whether depletion of these adaptors affected the half-life of CFTR. For these experiments, we monitored the half-life of CFTR using a metabolic pulse–chase protocol. We determined that, in the control samples, the half-life of CFTR was 12±0.6 h, which is consistent with previous reports . In the μ2-depleted cells, the half-life of CFTR was 16±2.8 h (Figure 4A), suggesting only a mild, but significant, effect. In contrast, in the Dab2-depleted cells, the half-life of CFTR was 37±7.1 h, indicating that CFTR was 3-fold more stable in Dab2-depleted cells (Figure 4B). The results indicate that although μ2 and Dab2 are both required for the endocytosis of CFTR, Dab2 also plays a significant role in regulating the half-life of CFTR.
Increased half-life of CFTR in μ2- and Dab2-depleted cells
The function of cell-surface CFTR is increased in the absence of Dab2
CFTR is present in a functional complex at the cell surface . Although numerous components of the complex have been identified, the role of Dab2 in the maintenance of the complex has not been examined. Since Dab2 depletion enhanced the cell-surface levels of CFTR and might alter the composition and more importantly the functionality of the complex, we performed Ussing chamber studies to test whether higher CFTR levels resulted in enhanced CFTR function. Following Dab2 depletion or transfection with a control siRNA, monolayers were mounted into Ussing chambers and forskolin was used to indirectly elevate intracellular cAMP levels to activate CFTR (Figure 5). Genistein was added to maximally stimulate the channels and to determine the non-cAMP component activity. When the current reached a maximum, CFTR-inh172 was added to block ISC. Under these conditions, the ΔISC was 88.4±7.4 μA/cm2 in control cells and 175.4±1.9 μA/cm2 in the Dab2-depleted cells. The results demonstrated that cell-surface CFTR function was increased following Dab2 depletion and suggest that Dab2 is not required to keep CFTR in a functional complex, since the ratio of the cAMP- and genistein-activated currents did not change compared with the control. Specifically, the results demonstrate that in Dab2-depleted cells, the CFTR currents were increased ~2-fold, and this is consistent with the elevated cell-surface levels of CFTR. Furthermore, the baseline currents were increased in the Dab2-depleted cells, and this is consistent with a previous study suggesting that the constitutive activity of CFTR increased as the expression levels increased .
Increased transepithelial chloride transport following Dab2 depletion in CFBE41o-WT monolayers
c-Cbl and CHIP depletion have no effect on CFTR endocytosis, but c-Cbl does affect the half-life of CFTR protein
Since the half-life of CFTR was dramatically increased by Dab2 depletion, we tested whether Dab2 depletion was somehow interfering with the function of two E3 ligases known to interact with CFTR, c-Cbl and CHIP [21,33]. In the first experiment, we examined the role of c-Cbl that has been shown to regulate CFTR endocytosis and lysosomal targeting . After siRNA KD of c-Cbl in polarized airway epithelia, we monitored CFTR endocytosis. Surprisingly, we saw no evidence that cCbl depletion affected CFTR internalization, even though c-Cbl depletion was greater than 90% (Figure 6A). In metabolic pulse–chase experiments, the half-life of CFTR changed from 13.1±2.4 h under control conditions to 18.5±3.1 h upon c-Cbl KD, indicating a modest effect on CFTR stability (Figure 6B). These results indicate that although c-Cbl does not play a significant role in CFTR endocytosis, it may contribute to CFTR ubiquitination at a post-endocytic step. A similar analysis on CHIP KD revealed that CHIP depletion had no effect on either CFTR internalization or half-life (Figures 7A and 7B). In summary, the depletion of the two E3 ligases had no effect on CFTR internalization, and only c-Cbl depletion caused a modest increase in CFTR stability, suggesting that the Dab2-associated effects on CFTR stability are independent of these two E3 ligases.
Effects of c-Cbl depletion on CFTR endocytosis rates (A) and protein half-life (B)
CHIP depletion has no effect on CFTR endocytosis (A) or protein half-life (B)
Dab2 depletion results in enlargement of the early endosomal compartments
Because Dab2 depletion dramatically increased the half-life of CFTR, next we investigated morphological changes in post-endocytic intracellular compartments following Dab2 depletion by immunocytochemistry. CFTR staining was greatly elevated in Dab2-depleted cells (Figures 8A and 8B), but more interestingly, the vesicles of the EEA1 compartment, the early sorting endosome, were dramatically enlarged compared with control cells (compare Figures 8A and 8B, middle panels). Furthermore, the number of EEA1-positive vesicles in each cell was also increased, as was the total amount of EEA1 protein as monitored by Western blot analysis (Figure 8C). This suggests that Dab2 may be necessary for protein exit from the early endosome. On the basis of a recent study by Penheiter et al. , who demonstrated that Dab2 depletion in NIH 3T3 cells inhibited type II TGF-β (transforming growth factor-β) receptor recycling, we tested Dab2 depletion on protein recycling to the cell surface. We used TR as our model, since it efficiently recycles to the cell surface . We utilized a fluorescently labelled Tfn for these studies. The cells were pulsed for 5 min with fluorescent Tfn (Alexa 594–Tfn) and then Tfn trafficking was monitored by immunofluorescence microscopy over a 30 min period. Tfn remains associated with the receptor until it is recycled to the cell surface and is released as apotransferrin, and therefore the loss of cell-associated fluorescence over time indicates recycling efficiency. The results indicate that in the control cells, the cell-associated Tfn was completely cleared after 30 min (Figure 9), whereas in the Dab2-depleted cells, much of the loaded fluorescent Tfn remained cell-associated for 30 min. This demonstrates that in the absence of Dab2, TR recycling is dramatically compromised. The results indicate, therefore, that Dab2 is important for protein recycling.
Dab2 depletion leads to the enlargement of early endosomes
Dab2 depletion reduces Tfn recycling
Delivery of CFTR to the late endosomal compartment is inhibited in Dab2-depleted cells
Although our results indicated that in Dab2-depleted cells endocytosis and recycling were reduced, we reasoned that inhibition of delivery to the late endosomal/lysosomal compartments  must be inhibited if the CFTR half-life is extended during Dab2 depletion. To test this idea, we examined CFTR trafficking to the late endosomal compartment by monitoring CFTR co-localization with a marker of this compartment, M6PR. To inhibit protease activity in this compartment, we treated the cells with a weak base, ammonium chloride, to enhance visualization of CFTR in this compartment prior to degradation. We have successfully employed this method in the past to monitor TR chimaeras that were targeted to the lysosome . The results indicate that under control conditions, CFTR and the M6PR did not co-localize (Figure 10A), whereas in the presence of protease inhibition, there is clear evidence that CFTR accumulated in these compartments (Figure 10B), presumably on its way to degradation. However, in Dab2-depleted cells, CFTR does not co-localize with the M6PR if proteases are inhibited, suggesting that CFTR trafficking to the late endosome is compromised in Dab2-depleted cells (Figure 10C). The results therefore suggest that CFTR delivery to the later stages of the endocytic pathways is compromised in Dab2-depleted cells, and this is consistent with the dramatically enhanced CFTR half-life that we monitored in the metabolic pulse–chase experiments.
CFTR delivery to the late endosomes is inhibited in Dab2-depleted cells
A major focus of the present study was to determine the role of two adaptor molecules, AP-2 (μ2) and Dab2, in CFTR endocytosis, recycling and degradation from the cell surface. Understanding the processes that regulate WT CFTR cell-surface levels will provide a background for understanding the cell-surface instability of rescued ΔF508 CFTR . We examined two endocytic adaptor molecules, AP-2 and Dab2, and two E3 ligases, CHIP and c-Cbl. c-Cbl has also been reported to serve as an endocytic adaptor molecule . Since AP-2 and Dab2 recognize different target recognition sequences, e.g. YDSI compared with FXNPXY, it was unclear how two different adaptors function in CFTR endocytosis. If both adaptors were used interchangeably, then inhibition of one might only have limited effects on CFTR endocytosis. In contrast, we see similar and very significant effects of AP-2 or Dab2 KDs on CFTR internalization. Furthermore, since depletion of both μ2 and Dab2 resulted in a similar effect as the depletion of one, these results are consistent with the hypothesis that they work in concert and disruption of one of them eliminates the function of the other during endocytosis.
The results of the present study are consistent with a recent report indicating that in intestinal epithelial cells, Dab2 directly interacts with the α2 subunit of AP-2, but not CFTR . In the model by Collaco et al. , the α2 subunit of AP-2 interacts directly with both CFTR and Dab2, and Dab2 interacts with myosin VI, which interacts with the actin cytoskeleton. Thus Dab2 forms a bridge between AP-2 and myosin VI and does not bind directly to CFTR. Therefore Dab2 does not serve as a CFTR adaptor in the classic sense. This model is consistent with the earlier observations by Morris et al. [36,37] that Dab2 interacts with the α2 subunit of AP-2 and myosin VI. Our CFTR internalization data in μ2 and Dab2-depleted cells are consistent with the model proposed by Collaco et al.  for CFTR endocytosis. Since both the μ2 and the α2 subunits interact with CFTR [19,23], this suggests that both subunits of AP-2 have critical contact sites for CFTR. In the Collaco et al.  study, depletion of the α2 subunit interfered with CFTR endocytosis; we found similar results following KD of the μ2 subunit of AP-2. These two findings are also consistent with the idea that depletion of either AP-2 subunit interferes with the assembly of the AP-2 molecule and therefore inhibits AP-2 function. It will be interesting to see which region of CFTR is recognized by the α2 chain of AP-2.
Because the present study indicated that Dab2 depletion dramatically increased the half-life of CFTR, we tested the idea that c-Cbl might be mediating this effect. We based our hypothesis on a recent report indicating that in human airway epithelial cells depletion of c-Cbl enhanced the half-life of CFTR and also inhibited endocytosis . These results suggested that c-Cbl, a known E3 ligase, also functioned as an adaptor molecule. We tested whether c-Cbl and μ2 were interchangeable during endocytosis. Surprisingly, c-Cbl depletion (~90%) did not affect CFTR endocytosis. The only effect of c-Cbl depletion was a modest increase in the half-life of CFTR. This supports the view that c-Cbl functions at a post-endocytic step . Therefore the results of the present study, in combination with those of Collaco et al. , confirm that the only adaptor for CFTR endocytosis is AP-2, as was originally proposed by Weixel and Bradbury . However, the most important finding of the present study is that we determined the differences in the function of AP-2 and Dab2.
Because we observed a dramatic increase in the half-life of CFTR following Dab2 depletion, we wanted to understand the mechanism of this effect. Our first analysis indicated that two E3 ligases known to associate with CFTR, c-Cbl and CHIP [21,33], were not necessary for CFTR endocytosis, and more importantly depletion of these E3 ligases did not dramatically increase the half-life of CFTR. However, by examining the morphology of the endocytic compartments following Dab2 depletion, we observed that the sorting endosome or EEA1 compartment was enlarged in these cells. This result is consistent with those of Penheiter et al. , who were examining type II TGF-β receptor recycling in NIH 3T3 cells. They studied the role of Dab2 in TGF-β receptor endocytosis and found that siRNA KD of Dab2 had no effect on TGF-β receptor internalization. Dab2 depletion did, however, result in enlarged EEA1-positive endosomes, and interfered with TGF-β receptor trafficking to the Rab11-positive recycling endosome and TGF-β receptor recycling . How this affected TGF-β receptor half-life, however, was not examined in these studies. This sorting block is also consistent with the results of Chibalina et al. , who demonstrated that siRNA depletion of myosin VI or its interacting protein LMTK2 (lemur tyrosine kinase 2) blocked protein transport out of the sorting endosome, resulting in enlarged endosomes and a delay in TR recycling. Our present studies in Dab2-depleted cells indicate a similar phenotype of enlarged endosomes, with a corresponding inhibition of TR recycling. Therefore the results of the present study, in combination with those discussed above, indicate significant roles for myosin VI, Dab2 and LMTK2 in sorting of proteins from the early endosome to the recycling endosome. The results of the present study also establish that Dab2 is necessary for both endocytosis and recycling.
In previous studies we examined mutations in CFTR that both enhanced  and inhibited  endocytosis, and neither one of these mutations affected the half-life of CFTR protein. Furthermore, in the present studies, μ2 depletion only had a modest effect on the half-life of CFTR. Therefore it is unclear why the block at the sorting endosome stage would have a dramatic effect on the half-life of CFTR in Dab2-depleted cells. This was even more surprising since previous studies reported that blocking the exit from the recycling endosome has no effect on CFTR expression levels , suggesting that the half-life was not affected. Therefore, in order to gain a deeper understanding of this mechanism, we examined how CFTR transport to the late endosome was affected in Dab2-depleted cells. The results indicated that Dab2 is necessary for CFTR delivery into the degradation compartment. Interestingly, it has been demonstrated that depletion of the C-terminal EHD4 (Eps15 homology domain 4) by siRNA leads to the generation of enlarged early endosomal structures that contain Rab5 and EEA1, and accumulation of LDL particles destined for degradation . Because the LDL particles accumulated rather than being delivered to the late endosome and lysosome, the authors suggested that EHD4 KD indicated that EHD4 was not only involved in the recycling of molecules through the sorting endosome, but also in delivery of cargo destined for degradation. On the basis of the similar phenotype of the enlargement of the sorting endosome where recycling receptors and cargo are trapped, it is tempting to speculate that myosin VI and Dab2 are required for efficient exit from the sorting endosome for delivery to both the recycling compartment and the late endosome, and therefore Dab2 is required at multiple steps in the endocytic pathway.
In summary, the results of the present study indicate the both μ2 and Dab2 are necessary for CFTR endocytosis; however, only Dab2 is essential for the post-endocytic trafficking of CFTR either for recycling or for delivery to the late endosome. On the basis of this, we propose a model (Figure 11) that illustrates that three sites in the endocytic pathway are affected by Dab2 depletion: (1) internalization from the cell surface; (2) delivery from the sorting endosome to the recycling endosome; and (3) delivery to the late endosome. Given the complexity of the trafficking of the WT CFTR protein in the endocytic pathway, it will be intriguing to see how these types of manipulations with siRNA KDs will affect the trafficking of a rescued mutant protein such as ΔF508 CFTR.
A model for Dab2 function in CFTR trafficking
594–Tfn, Alexa Fluor® 594-conjugated transferrin
adaptor protein 2
casitas B-lineage lymphoma
cystic fibrosis transmembrane conductance regulator
C-terminus of the Hsc (heat-shock cognate) 70-interacting protein
Eps15 homology domain 4
early endosome antigen 1
lemur tyrosine kinase 2
mannose 6-phosphate receptor
small interfering RNA
transforming growth factor-β
Lianwu Fu and James Collawn designed the experiments and Lianwu Fu conducted the experiments. Andras Rab conducted the EEA1 Western blotting experiments and Li Ping Tang conducted the Ussing chamber experiments. Steven Rowe and Zsuzsa Bebok provided scientific advice on the functional analysis of CFTR. Lianwu Fu, Zsuzsa Bebok and James Collawn wrote the paper.
This work was supported by the National Institutes of Health [grant numbers R01 DK60065 (to J.F.C.), R01 HL076587 (to Z.B.), K23 DK075788 (to S.M.R.)] and the American Lung Association [grant number RG82840N (to L.F.)]. This work was also supported by Dr Eric Sorscher and grants to the Gregory Fleming James Cystic Fibrosis Research Center [grant numbers P30 DK 072482, R474-CR11].