The Short Apical Membrane Half-life of Rescued ΔF508-Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) Results from Accelerated Endocytosis of ΔF508-CFTR in Polarized Human Airway Epithelial Cells*

The most common mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) gene in individuals with cystic fibrosis, ΔF508, causes retention of ΔF508-CFTR in the endoplasmic reticulum and leads to the absence of CFTR Cl- channels in the apical plasma membrane. Rescue of ΔF508-CFTR by reduced temperature or chemical means reveals that the ΔF508 mutation reduces the half-life of ΔF508-CFTR in the apical plasma membrane. Because ΔF508-CFTR retains some Cl- channel activity, increased expression of ΔF508-CFTR in the apical membrane could serve as a potential therapeutic approach for cystic fibrosis. However, little is known about the mechanisms responsible for the short apical membrane half-life of ΔF508-CFTR in polarized human airway epithelial cells. Accordingly, the goal of this study was to determine the cellular defects in the trafficking of rescued ΔF508-CFTR that lead to the decreased apical membrane half-life of ΔF508-CFTR in polarized human airway epithelial cells. We report that in polarized human airway epithelial cells (CFBE41o-) the ΔF508 mutation increased endocytosis of CFTR from the apical membrane without causing a global endocytic defect or affecting the endocytic recycling of CFTR in the Rab11a-specific apical recycling compartment.

A recent study demonstrates that in fibroblasts (BHK-21 cells) heterologously expressing CFTR, the ⌬F508 mutation reduces the plasma membrane half-life of CFTR by attenuating the endocytic recycling of rescued ⌬F508-CFTR without significantly affecting ⌬F508-CFTR endocytosis (25). Even though non-epithelial cells possess the cellular machinery to sort proteins to specific membrane domains (41)(42)(43), it is widely accepted that trafficking of heterologously expressed plasma membrane proteins in non-epithelial cells can differ from polarized epithelial cells (43,44). Thus, studies designed to elucidate the cellular defects that lead to the decrease in the apical membrane half-life of rescued ⌬F508-CFTR need to be carried out in human airway epithelial cells.
To this end, we studied CFTR trafficking in polarized human airway epithelial cell lines (CFBE41oϪ) expressing endogenous ⌬F508-CFTR or stably expressing either ⌬F508-CFTR or WT-CFTR. We report that the apical plasma membrane half-life of rescued ⌬F508-CFTR was shorter than that of WT-CFTR. The ⌬F508 mutation specifically increased endocytosis of ⌬F508-CFTR from the apical membrane without causing a global endocytic defect or altering the endocytic recycling of CFTR. Furthermore, the decrease in the apical membrane half-life of ⌬F508-CFTR was not mediated by a putative cryptic endocytic motif, YRSV (517-520). The endocytic recycling of WT-CFTR and ⌬F508-CFTR occurred in the Rab11a-specific apical recycling compartment in polarized human airway epithelial cells. Taken together, these data indicate that in polarized human airway epithelial cells (CFBE41oϪ) the ⌬F508 mutation reduces the apical membrane half-life of ⌬F508-CFTR by specifically accelerating the endocytic retrieval of ⌬F508-CFTR from the plasma membrane without affecting the endocytic recycling of ⌬F508-CFTR.
siRNA-mediated Silencing of Rab11a Expression-The doublestranded, small interfering RNA (siRNA) against a non-conserved region of the human Rab11a sequence (siRab11a; catalog #1022547) and the double-stranded, non-silencing siRNA control (Non-sil. siRNA; catalog #1022076) were purchased from Qiagen (Valencia, CA). Transfection of siRab11a and Non-sil. siRNA into CFBE41oϪ cells was conducted using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's instructions. CFBE41oϪ cells grown on 40-mm tissue culture plates were incubated for 24 h with the optimized transfection mixture (7.5 g of either siRab11a or Non-sil. siRNA and 15 l of Lipofectamine TM 2000 per plate) at 37°C. Subsequently, cells were cultured in fresh medium at 37°C for 36 h and then at 27°C for another 36 h to increase trafficking and expression of ⌬F508-CFTR in the plasma membrane. The efficiency of siRNA-mediated silencing of Rab11a expression was assessed by measuring the expression of endogenous Rab11a by Western blotting. The specificity of siRNA-mediated silencing of Rab11a expression was assessed by measuring the expression of other Rab GTPases facilitating endocytosis and recycling (Rab5a and Rab4) by Western blotting.
Plasmids and Transient Transfection-A plasmid containing GFPtagged ⌬F508-CFTR (GFP-⌬F508-CFTR) was generated as previously described using the eukaryotic expression vector pcDNA3.1 (49). To construct the GFP-⌬F508-CFTR Y517A mutant, the GFP-⌬F508-CFTR cDNA sequence in pcDNA3.1 was mutated using the QuikChange TM XL site-directed mutagenesis kit (Stratagene; La Jolla, CA). Constructs were sequence-verified by ABI PRISM dye terminator cycle sequencing (Applied Biosystems; Foster City, CA). Transient transfection of the GFP-tagged CFTR cDNAs into parental CFBE41oϪ cells was performed using Lipofectamine TM 2000 according to the manufacturer's instructions. CFBE41oϪ cells grown on 40-mm tissue culture plates were incubated for 24 h with the transfection mixture (2 g of cDNA and 4 l of Lipofectamine TM 2000 per plate) at 37°C. Subsequently, cells were cultured in fresh medium at 27°C for 36 h to increase the expression of GFP-⌬F508-CFTR and GFP-⌬F508-CFTR Y517A in the plasma membrane.
A plasmid containing FLAG-tagged wild type mouse Rab11a (FLAG-Rab11a WT) was generated as described previously using the eukaryotic expression vector pEF (50)(51)(52). To construct the FLAG-Rab11a S25N and FLAG-Rab11aS20Vmutants,theRab11acDNAsequenceinpEFwasmutated using the QuikChange TM XL site-directed mutagenesis kit. Constructs were sequence-verified by ABI PRISM dye terminator cycle sequencing. Transfection of the FLAG-Rab11a cDNAs into CFBE41oϪ cells stably expressing ⌬F508-CFTR was carried out using Lipofectamine TM 2000 according to the manufacturer's instructions. CFBE41oϪ cells grown on 40-mm tissue culture plates were incubated with the transfection mixture (4 g of cDNA and 8 l of Lipofectamine TM 2000 per plate) for 24 h at 37°C and in fresh medium for another 36 h.
Biochemical Determination of the Apical Membrane CFTR and BCRP-The biochemical determination of apical membrane CFTR and BCRP at steady state was performed by domain-selective cell surface biotinylation using EZ-Link TM Biotin-LC-Hydrazide or EZ-Link TM Sulfo-NHS-LC-Biotin (Pierce), as described previously in detail (35,54). Studies to determine the half-life of apical membrane proteins were conducted essentially as described by Heda and Marino (26). CFBE41oϪ cells, stably expressing either WT-CFTR or ⌬F508-CFTR, were grown on Transwell permeable growth supports. Parental CFBE41oϪ cells transiently transfected with either GFP-⌬F508-CFTR or GFP-⌬F508-CFTR Y517A were grown on plastic tissue culture plates. Confluent cells were cultured for 36 h in a CO 2 incubator at 27°C to increase the trafficking and expression of ⌬F508-CFTR in the apical plasma membrane. Incubation of cells with cycloheximide (Sigma-Aldrich; 20 g/ml), a protein synthesis inhibitor, was performed at 37°C, and the disappearance of CFTR or BCRP from the apical membrane was monitored over time. The half-lives were calculated by SPSS software using the one-phase exponential decay model, with plateau and span parameters constrained to 0 and 100, respectively.
Endocytic Assay and Endocytic Recycling Assay-Endocytic and endocytic recycling assays were performed in CFBE41oϪ cells expressing either WT-CFTR or ⌬F508-CFTR grown on Transwell permeable supports, using EZ-Link TM Sulfo-SS-Biotin (Pierce), as described previously in detail (35). The temperature in the incubator was reduced (27°C for 36 h) to increase trafficking and expression of ⌬F508-CFTR in the apical membrane of CFBE41oϪ cells.
Fluid-phase Endocytosis-Studies were conducted to determine the cellular uptake of Alexa 647-dextran (10,000 molecular weight; Molecular Probes), a fluorescent marker of fluid phase endocytosis (33,55) in CFBE41oϪ cells stably expressing either WT-CFTR or ⌬F508-CFTR. CFBE41oϪ cells were grown on glass coverslips at 37°C for 48 h and then at 27°C for 36 h to increase trafficking and expression of ⌬F508-CFTR in the plasma membrane. Alexa 647-dextran (1 mg/ml) was added to the cell culture medium (37°C) for 15 min. Subsequently, surface Alexa-647 dextran was removed by thorough washing at 4°C. Thereafter, cells were fixed for 20 min in 4% paraformaldehyde before being mounted in antifade medium (ProLong Gold, Molecular Probes). Thirty random images of cells expressing either WT-CFTR or ⌬F508-CFTR were examined from six different slides using an Olympus IX70 wide field microscope fitted with an Orca AG deep cooling CCD camera (Hamamatsu Photonics; Japan) and Openlab 4.0.3 software (Improvision, Inc.; Lexington, MA) using an 40ϫ oil immersion objective. All images were collected using the same settings for exposure, gain, offset, and binning. Alexa 647 fluorescence was quantified using Volocity 3.5.1 software (Improvision, Inc.).
ZO-1 Immunostaining-Studies were conducted to determine the ability of CFBE41oϪ cells, grown on glass or plastic to polarize. Cells expressing either WT-CFTR or ⌬F508-CFTR, plated on glass coverslips, were grown at 37°C for 48 h and then at 27°C for 36 h. Subsequently, cells were fixed in Ϫ20°C methanol. Nonspecific binding sites were blocked with 10% normal goat serum. Cells were stained with polyclonal anti-ZO-1 antibody in 10% normal goat serum, washed with phosphate-buffered saline, postfixed in 4% paraformaldehyde, and incubated with glycine (100 mM) to block the aldehyde groups. After washing in phosphate-buffered saline and blocking for the second time with 10% normal goat serum, cells were incubated with Alexa Fluor 647 goat anti-rabbit secondary antibody in 10% normal goat serum. Cells were mounted with antifade medium (ProLong Gold; Molecular Probes). Z-stacks of the immunolabeled cells were acquired with an Olympus IX70 wide field microscope (40ϫ oil immersion objective) fitted with an Orca AG deep cooling CCD camera (Hamamatsu Photonics, K.K.; Japan) and Openlab 4.0.3 software (Improvision, Inc.), and the volumes FIGURE 1. Summary of experiments performed to determine the effects of reduced temperature (27°C for 36 h) on the expression of endogenous ⌬F508-CFTR (⌬F508e) in parental CFBE41o؊ cells and the expression of WT-CFTR (WT) and ⌬F508-CFTR (⌬F508) in stably transduced polarized CFBE41o؊ cells. A, summary of experiments demonstrating by domain-selective cell surface biotinylation that reduced temperature had no significant effect on the apical membrane expression of transduced WT-CFTR or endogenous ⌬F508-CFTR but increased the expression of ⌬F508-CFTR in stably transduced CFBE41oϪ cells to levels equivalent to the expression observed in the WT-CFTR-transduced cells. Data are expressed as percent of the endogenous ⌬F508-CFTR expressed in the apical membrane of parental CFBE41oϪ cells at 37°C. The asterisk indicates p Ͻ 0.05. 3-4 experiments were performed per group. B, representative Western blots demonstrating that in CFBE41oϪ cells stably transduced with ⌬F508-CFTR, reduced temperature 1) increased expression of the core glycosylated (band B) (*) and the fully glycosylated (band C) (**) form of ⌬F508-CFTR in cell lysate (the left panel) and 2) increased trafficking of the ⌬F508-CFTR band C to the apical plasma membrane (right panel). Reduced temperature had no effect on ezrin expression in cell lysate. It is important to note that at 37°C a small but detectable amount of ⌬F508-CFTR, predominantly band B, was observed in the lysate from stably transduced CFBE41oϪ cells (left panel, lower image, 3-min exposure). In addition, the presence of a small amount of ⌬F508-CFTR band C was evident in the apical membrane at 37°C in the ⌬F508-CFTRtransduced CFBE41oϪ cells (right panel). Proteins were separated by SDS-PAGE using a 7.5% gel.
Immunoprecipitation and Immunoblotting-CFTR and Rab11a were immunoprecipitated from CFBE41oϪ and Calu-3 cell lysates by methods described previously (33). For experiments with endogenous and stably expressed proteins, cells were grown on Transwell permeable supports, and for experiments with transiently transfected FLAG-Rab11a WT, cells were grown on plastic tissue culture plates. Briefly, cells were solubilized in lysis buffer containing 150 mM NaCl, 50 mM Tris, pH 7.2, 0.1% IGEPAL (Sigma-Aldrich), 5 mM MgCl 2 , 5 mM EDTA, 1 mM EGTA, 30 mM NaF, 1 mM Na 3 VO 4 , and Complete Protease Inhibitor mixture (Roche Applied Science). After centrifugation at 14,000 ϫ g for 15 min to pellet insoluble material, the soluble lysates were incubated for 10 min at 30°C with 20 M GTP␥S (Sigma-Aldrich), a nonhydrolysable analog of GTP (56 -58). After cooling to 4°C, the soluble lysates were pre-cleared by incubation with protein G conjugated to Sepharose beads (Pierce) at 4°C. CFTR was immunoprecipitated by incubation with the M3A7 antibody-protein G-Sepharose complexes. Immunoprecipitated proteins were eluted from the protein G-Sepharose complexes by incubation at 85°C for 5 min in sample buffer (Bio-Rad) containing 80 mM dithiothreitol. Immunoprecipitated proteins were separated by SDS-PAGE using 15% gels (Bio-Rad) and analyzed by Western blotting with an appropriate primary antibody and an antimouse or anti-rabbit horseradish peroxidase secondary antibody.
Data Analysis and Statistics-Statistical analysis of the data were performed using GraphPad Prism version 4.0 for Macintosh (Graph-Pad Software Inc.; San Diego, CA) and SPSS software (SPSS Inc.; Chicago, IL). The half-lives were calculated by SPSS software using the one-phase exponential decay model, with plateau and span parameters constrained to 0 and 100, respectively. The half-life means were compared by a two-tailed t test with assumed unequal variances. The means for the remaining data were compared by a two-tailed t test. A p value Ͻ0.05 was considered significant. Data are expressed as the mean Ϯ S.E.

The Apical Membrane Half-life of Rescued ⌬F508-CFTR Compared with WT-CFTR Is Decreased in Polarized Human Airway Epithelial
Cells (CFBE41oϪ)-First, studies were conducted to determine the relative amount of WT-CFTR and ⌬F508-CFTR in CFBE41oϪ cells grown on permeable growth supports at 37 or at 27°C, a temperature that, at least for some cells, increases the expression of ⌬F508-CFTR in the plasma membrane (15,16). As illustrated in Fig. 1, ⌬F508-CFTR was detected in the apical plasma membrane of parental and ⌬F508-CFTRtransduced cells at 37°C. Reduced temperature (27°C for 36 h) increased the apical membrane expression of ⌬F508-CFTR in the transduced cells and had no significant effect on the expression of endogenous ⌬F508-CFTR in the parental cells or WT-CFTR in the transduced cells (Fig. 1A). It is worth noting that in stably transduced CFBE41oϪ cells, the plasma membrane expression of WT-CFTR and ⌬F508-CFTR was comparable at reduced temperature (Fig. 1A). Reduced temperature did not affect the expression of the cytoskeletal protein, ezrin, in cell lysates (Fig. 1B). Thus, reduced temperature increased the plasma membrane expression of ⌬F508-CFTR to a level sufficient to perform endocytic and recycling assays and similar to the level of WT-CFTR expression.
Next, studies were conducted to determine whether the ⌬F508 mutation affected the biochemical half-life of CFTR in the apical membrane of polarized human airway epithelial cells (CFBE41oϪ). As described above, the temperature was decreased before studies (27°C for 36 h) to increase expression of ⌬F508-CFTR in the apical membrane. The disappearance of WT-CFTR and ⌬F508-CFTR from the apical membrane was monitored over time at 37°C in the presence of 20 g/ml cyclohexamide, a protein synthesis inhibitor, as described under "Materials and Methods." As illustrated in Fig. 2, the apical membrane half-life of ⌬F508-CFTR (1.0 h) was significantly shorter than that of WT-CFTR (3.0 h; p Ͻ 0.05). Reduced temperature had no effect on the apical membrane half-life of WT-CFTR (3.3 versus 3.0 h for cells grown at 37 or at 27°C for 36 h, respectively). These data demonstrate for the first time that the ⌬F508 mutation decreases the biochemical half-life of ⌬F508-CFTR in the apical membrane of polarized human airway epithelial cells.
Endocytosis of Rescued ⌬F508-CFTR Compared with WT-CFTR Is Increased in Polarized CFBE41oϪ Cells-The decreased apical membrane half-life of ⌬F508-CFTR in CFBE41oϪ cells could result from either accelerated endocytosis or attenuated endocytic recycling of ⌬F508-CFTR or both. Accordingly, studies were conducted to test the  Trafficking of ⌬F508-CFTR in Human Airway Epithelial Cells NOVEMBER 4, 2005 • VOLUME 280 • NUMBER 44 hypothesis that the ⌬F508 mutation increased ⌬F508-CFTR endocytosis. Endocytosis of CFTR was measured at 2.5, 5, and 7.5 min, as described under "Materials and Methods." Endocytosis of WT-CFTR and ⌬F508-CFTR increased linearly between 0 and 5 min (not shown). Thus, data are reported at the 5 min time point. The percent of ⌬F508-CFTR endocytosed at 5 min was significantly greater than that of WT-CFTR (Fig. 3A). The accelerated endocytosis of ⌬F508-CFTR is consistent with the decreased apical membrane half-life of rescued ⌬F508-CFTR observed in polarized CFBE41oϪ cells.
The Endocytic Recycling of Rescued ⌬F508-CFTR Is Similar to WT-CFTR in Polarized Human Airway Epithelial Cells (CFBE41oϪ)-As noted above, the decreased apical membrane half-life of ⌬F508-CFTR could also result from inhibition of the ⌬F508-CFTR recycling from endosomes to the apical membrane. Accordingly, endocytic recycling of CFTR was measured at 2.5, 5, and 7.5 min as described under "Materials and Methods." Endocytic recycling of WT-CFTR and ⌬F508-CFTR increased linearly between 0 and 5 min (not shown). Thus, data are reported at the 5 min time point. The percent of endocytosed ⌬F508-CFTR that recycled back to the plasma membrane was similar to that of WT-CFTR (Fig. 3B). Our data suggest that the ⌬F508 mutation reduces the half-life of CFTR in the apical membrane of polarized human airway epithelial cells by accelerating the endocytosis of CFTR from the apical membrane without affecting the endocytic recycling of CFTR.

Accelerated Endocytosis and Decreased Apical Membrane Half-life of Rescued ⌬F508-CFTR Do Not Result From a Global Endocytic Trafficking Defect in Polarized CFBE41oϪ
Cells-Studies in polarized epithelial cells demonstrate that the ⌬F508 mutation causes a global endocytic trafficking defect; thus, CFTR might be a pleiotropic regulator of endocytic trafficking (59 -61). However, the effects of the ⌬F508 mutation on endocytic trafficking appear to be cell type-specific (62)(63)(64). Accordingly, studies were conducted to determine whether the accelerated endocytosis and decreased apical membrane half-life of rescued ⌬F508-CFTR in CFBE41oϪ cells resulted from a generalized defect in endocytic trafficking.
CFTR (27,28) and other ABC transporters (65) are endocytosed from the apical membrane by a clathrin-dependent pathway. Hence, a generalized defect in clathrin-mediated endocytosis should alter the apical membrane expression and half-life of another ABC transporter. The BCRP is an ABC transporter (66) highly expressed in the apical membrane of respiratory epithelial cells (67) including polarized CFBE41oϪ cells (Fig. 4C). As illustrated in Fig. 4A, the apical membrane expression of BCRP at steady state did not differ in polarized CFBE41oϪ cells expressing either WT-CFTR or ⌬F508-CFTR. Furthermore, the apical membrane half-life of BCRP did not differ in CFBE41oϪ cells expressing WT-CFTR (4.4 h) or ⌬F508-CFTR (4.6 h) as monitored by the disappearance of BCRP from the apical membrane over time at 37°C in the presence of 20 g/ml cyclohexamide (Figs. 4, B and C). Further- Trafficking of ⌬F508-CFTR in Human Airway Epithelial Cells more, the ⌬F508 mutation had no effect on fluid phase endocytosis in CFBE41oϪ cells as determined by measuring the uptake of the Alexa 647-conjugated dextran (Fig. 5A). These data are consistent with the conclusion that increased endocytosis and decreased apical membrane half-life of rescued ⌬F508-CFTR observed in polarized CFBE41oϪ cells did not result from a generalized defect in clathrin-mediated endocytosis or in fluid-phase endocytosis.
Rab5a and CFTR Endocytic Trafficking-The small GTPase, Rab5a, is an important regulator of clathrin-mediated endocytosis (68, 69) and fluid phase endocytosis (70 -72). A recent study demonstrates that endocytosis of WT-CFTR and ⌬F508-CFTR is Rab5-dependent in heterologous cells (24). Interestingly, the Rab5a gene expression is up-regulated in bronchial epithelial cells expressing ⌬F508-CFTR (IB3-1) compared with the IB3-1 cells corrected with WT-CFTR (S9 cells) (73). Thus, accelerated endocytosis of ⌬F508-CFTR observed in this study could result from increased expression of the Rab5a protein. However, as illustrated in Fig. 5B, the expression of endogenous Rab5a was similar in polarized CFBE41oϪ cells transduced with either WT-CFTR or ⌬F508-CFTR. Thus, we conclude that a difference in Rab5a expression does not account for accelerated endocytosis of rescued ⌬F508-CFTR in polarized CFBE41oϪ cells.

A Putative Cryptic Endocytic Motif YRSV (517-520) Does Not Contribute to the Decrease in the Apical Membrane Half-life of Rescued
⌬F508-CFTR in Polarized CFBE41oϪ Cells-A recent study based on the crystal structure of the first nucleotide binding domain (NBD1) of human CFTR (74) suggests that the sequence YRSV (517-520) is a cryptic tyrosine-based endocytic motif, which becomes functional in ⌬F508-CFTR due to conformational changes in the ⌬F508-NBD1 (75). The sequence YRSV (517-520) conforms to the canonical YXXØ endocytic motif (Y is tyrosine, X is any amino acid, and Ø is an amino acid with a bulky hydrophobic side chain) (76) and is conserved across species. If exposed by the ⌬F508 mutation, this motif may account for accelerated endocytosis of rescued ⌬F508-CFTR observed in the present study. Thus, studies were conducted to test the hypothesis that the putative YRSV (517-520) endocytic motif might be responsible for increased endocytosis of ⌬F508-CFTR. Mutation of Tyr-517 would be expected to reduce the endocytic retrieval and, thus, increase the expression and the half-life of rescued ⌬F508-CFTR in the apical plasma membrane. To test this hypothesis, we transiently expressed the GFP-tagged ⌬F508-CFTR or the GFP-⌬F508-CFTR Y517A mutant in parental CFBE41oϪ cells. Because the transfection efficiency is low in CFBE41oϪ cells grown on permeable supports, 3 the transient expression studies were performed in CFBE41oϪ cells grown on plastic tissue culture plates. As demonstrated in Fig. 6, CFBE41oϪ cells form tight junctions and thereby polarize when grown on non-permeable growth support such as glass or plastic tissue culture plates. The apical membrane expression of GFP-⌬F508-CFTR and GFP-⌬F508-CFTR Y517A was similar at steady state (Fig. 7A). Furthermore, as illustrated in Figs. 7, B and C, the apical membrane half-life of GFP-⌬F508-CFTR Y517A (1.0 h) did not differ from that of GFP-⌬F508-CFTR (1.1 h). These data demonstrate that the putative endocytic motif, YRSV (517-520), does not contribute to the decrease in the apical membrane half-life of rescued ⌬F508-CFTR.

Endogenous Rab11a Facilitates Recycling of WT-CFTR and Rescued ⌬F508-CFTR to the Apical Plasma Membrane in Polarized CFBE41oϪ
Cells-Trafficking between apical recycling endosomes and the apical plasma membrane in epithelial cells is facilitated by Rab11a, a member of the Ras-like small GTPases (77)(78)(79)(80). Rab11a facilitates plasma membrane expression of WT-CFTR and rescued ⌬F508-CFTR in heterologous cells (24). Thus, studies were conducted to determine whether Rab11a facilitates the endocytic recycling of CFTR in human airway epithelial cells. We first examined whether WT-CFTR and ⌬F508-CFTR interact with endogenous Rab11a in polarized human airway epithelial cells. As illustrated in Fig. 8, endogenous Rab11a co-immunoprecipitated with endogenous ⌬F508-CFTR in parental CFBE41oϪ cells and with WT-CFTR and ⌬F508-CFTR in stably transduced CFBE41oϪ cells. In addition, endogenously expressed WT-CFTR and Rab11a co-immunoprecipitated in another polarized human airway epithelial cell line (Calu-3; Fig. 8D). These data demonstrate that Rab11a interacts with WT-CFTR and ⌬F508-CFTR in recycling endosomes.
The co-immunoprecipitation studies presented above suggest a role for Rab11a in the endocytic recycling of CFTR in polarized human airway epithelial cells. To examine more directly the role Rab11a plays in endocytic recycling of WT-CFTR and ⌬F508-CFTR, studies were conducted using siRNA-mediated silencing of Rab11a expression. Reduced expression of Rab11a should decrease the endocytic recycling of CFTR and, thus, reduce the apical membrane expression of WT-CFTR and rescued ⌬F508-CFTR. To test these predictions, CFBE41oϪ cells stably expressing either WT-CFTR or ⌬F508-CFTR were transfected with double-stranded small interfering RNA specific for a non-conserved region of the human Rab11a sequence (siRab11a) or with the Non-sil. siRNA) as described under "Materials and Methods." Expression of endogenous Rab11a was similar in the WT-CFTR-and ⌬F508-CFTR-transduced CFBE41oϪ cells (Fig. 9A). siRab11a decreased the expression of Rab11a in both cell lines (Fig. 9B). Neither Rab5a nor Rab4 protein expression was altered by siRab11a (Fig. 9C). By contrast, Non-sil. siRNA had no effect on the endogenous expression of Rab11a, WT-CFTR, or ⌬F508-CFTR when compared with the non-transfected cells (not shown). As predicted, siRab11a decreased the expression of WT-CFTR and ⌬F508-CFTR in the apical plasma membrane (Fig. 9D). The decreased apical membrane expression of WT-CFTR and ⌬F508-CFTR in CFBE41oϪ cells is consistent with the view that Rab11a facilitates the endocytic recycling of WT-CFTR and rescued ⌬F508-CFTR in polarized human airway epithelial cells. Trafficking of ⌬F508-CFTR in Human Airway Epithelial Cells NOVEMBER 4, 2005 • VOLUME 280 • NUMBER 44 Endocytosis and endocytic recycling determine, at least in part, the expression of CFTR in the apical membrane (28,81,82). Thus, inhibition of the endocytic recycling of ⌬F508-CFTR by siRNA-mediated silencing of Rab11a together with increased endocytosis of ⌬F508-CFTR should result in lower apical membrane expression of ⌬F508-CFTR compared with WT-CFTR. As expected, a similar degree of Rab11a silencing in the WT-CFTR-and ⌬F508-CFTR-expressing cells had a more profound inhibitory effect on the plasma membrane expression of ⌬F508-CFTR (Fig. 9D). FIGURE 6. Representative fluorescence staining demonstrating the ability of CFBE41o؊ cells to polarize on non-permeable growth supports, as determined by the distribution of the adherence junction protein, ZO-1. CFBE41oϪ cells stably expressing WT-CFTR or ⌬F508-CFTR plated on glass coverslips were grown at 37°C for 48 h and then at 27°C for 36 h, conditions required to stimulate the trafficking and expression of ⌬F508-CFTR in the plasma membrane. Cells were fixed and stained with a polyclonal anti-ZO-1 antibody and Alexa Fluor 647 goat anti-rabbit secondary antibody as described under "Materials and Methods." Z-stacks of the immunolabeled cells were acquired with the Olympus IX70 wide field microscope (40ϫ objective) fitted with Orca AG deep cooling CCD camera (Hamamatsu Photonics, K.K.; Japan) and Openlab 4.0.3 software (Improvision Inc.), and the volumes were deconvolved (iterative restoration) with Volocity 3.5.1 software (Improvision Inc.). FIGURE 7. Summary of experiments performed to determine the effects of the Y517A mutation in ⌬F508-CFTR on the apical membrane expression at steady state and the apical membrane half-life of rescued ⌬F508-CFTR. Parental CFBE41oϪ cells grown on plastic tissue culture plates were transiently transfected with either GFP-⌬F508-CFTR or the GFP-⌬F508-CFTR Y517A mutant. Reduced temperature (27°C for 36 h) was used to increase the trafficking and expression of ⌬F508-CFTR in the plasma membrane. A, summary of biotinylation experiments demonstrating that the apical membrane expression of GFP-⌬F508-CFTR and GFP-⌬F508-CFTR Y517A at steady state did not differ. B, summary of biotinylation experiments demonstrating that the apical membrane half-life of GFP-⌬F508-CFTR (1.0 h) and GFP-⌬F508-CFTR Y517A (1.1 h) did not differ in CFBE41oϪ cells. Disappearance of GFP-⌬F508-CFTR and GFP-⌬F508-CFTR Y517A from the apical membrane was monitored over time in the presence of 20 g/ml cyclohexamide (CHX) at 37°C. The half-life was calculated using the one-phase exponential decay model as described under "Materials and Methods." C, representative Western blots demonstrating the disappearance of GFP-⌬F508-CFTR and GFP-⌬F508-CFTR Y517A from the apical membrane over time in CFBE41oϪ cells. Cyclohexamide did not affect ezrin expression in cell lysates. Proteins were separated by SDS-PAGE using a 7.5% gel. Three experiments were performed per group.

Rab11a Mutants Affect the Apical Membrane Expression of Rescued
⌬F508-CFTR in Polarized CFBE41oϪ Cells-Rab GTPases alternately bind GTP and GDP and hydrolyze GTP to GDP (83). The cycling of Rab proteins between the GTP-bound (active) and GDP-bound (inactive) form, which serves as the molecular basis for their activity, has been utilized as a valuable tool to study the function of Rab proteins in the trafficking events of membrane proteins. Rab11a S25N, a mutant deficient in GTP binding (GDP-locked), has a dominant negative effect on the endocytic recycling of numerous plasma membrane proteins (80, 84 -89). Rab11a S20V, a mutant deficient in GTP hydrolysis (GTPlocked) (86), has a dominant positive effect on the endocytic recycling of several plasma membrane proteins (86,90,91).
Thus, in CFBE41oϪ cells the GDP-locked Rab11a S25N should compete with endogenous Rab11a for cargo (⌬F508-CFTR) and/or for the Rab11a effectors involved in ⌬F508-CFTR recycling and, consequently, decrease expression of ⌬F508-CFTR in the apical membrane. Furthermore, the GTP-locked Rab11a S20V, overexpressed in CFBE41oϪ cells, should compete with endogenous Rab11a for cargo (⌬F508-CFTR) and/or for the Rab11a effectors involved in ⌬F508-CFTR recycling and, consequently, increase expression of ⌬F508-CFTR in the apical membrane. Using FLAGtagged constructs of Rab11a, experiments were conducted to test these predictions and to confirm and extend our studies examining the role of Rab11a as an important regulator of ⌬F508-CFTR recycling. The ⌬F508-CFTR-transduced CFBE41oϪ cells were transiently transfected with FLAG-Rab11a WT, FLAG-Rab11a S25N, FLAG-Rab11a S20V, or GFP (unrelated cDNA control), as described under "Materials and Methods." Transfection of GFP alone had no effect on the apical membrane expression of ⌬F508-CFTR compared with non-transfected cells (not shown). FLAG-Rab11a WT had no effect on the apical membrane expression of ⌬F508-CFTR compared with the GFP control ( Fig. 10A) but, similar to endogenous Rab11a (Fig. 8), was capable of interacting with ⌬F508-CFTR (Fig. 10B). As predicted, the GDP-locked FLAG-Rab11a S25N decreased the plasma membrane expression of ⌬F508-CFTR (Fig. 10A). Moreover, the GTP-locked FLAG-Rab11a S20V increased the plasma membrane expression of ⌬F508-CFTR (Fig. 10A). Taken together, these data indicate that Rab11a facilitates the endocytic recycling of WT-CFTR and rescued ⌬F508-CFTR to the apical membrane in human airway epithelial cells.

DISCUSSION
The major novel observation in the present study is that the ⌬F508 mutation decreases the apical membrane half-life of rescued ⌬F508-CFTR in polarized human airway epithelial cells by accelerating its endocytosis from the apical membrane without inhibiting the endocytic recycling of ⌬F508-CFTR. Furthermore, our data demonstrate that the ⌬F508 mutation does not globally disrupt the apical clathrin-mediated endocytic pathway or fluid phase endocytosis in human airway epithelial cells (CFBE41oϪ). Moreover, our data demonstrate that both WT-CFTR and rescued ⌬F508-CFTR undergo trafficking in the Rab11aspecific apical recycling compartment in polarized human airway epithelial cells.
We report that the reduction of the apical membrane half-life of rescued ⌬F508-CFTR in polarized human airway epithelial cells is a consequence of increased endocytic retrieval of ⌬F508-CFTR from the apical membrane. In all published studies on non-epithelial cells (23)(24)(25) or epithelial cells (26) heterologously expressing CFTR, the ⌬F508 mutation reduced the biochemical half-life of plasma membrane CFTR. Thus, previous studies are in general agreement with our finding in polarized human airway epithelial cells in which the ⌬F508 mutation reduced the biochemical half-life of ⌬F508-CFTR. By contrast, previous studies on heterologous, non-epithelial cells suggest that the reduced half-life of rescued ⌬F508-CFTR in the plasma membrane is due to a defect in the endocytic recycling of ⌬F508-CFTR. In particular, a study in BHK-21 cells suggests that the FIGURE 8. Representative immunoprecipitation experiments demonstrating that WT-CFTR and ⌬F508-CFTR interact with endogenous Rab11a in polarized human airway epithelial cells. CFTR was immunoprecipitated from CFBE41oϪ cells stably expressing either WT-CFTR (A) or ⌬F508-CFTR (B), from parental CFBE41oϪ cells endogenously expressing ⌬F508-CFTR (⌬F508-CFTRe) (C), or from Calu-3 cells endogenously expressing WT-CFTR (D) using anti-CFTR antibody M3A7. The immunoprecipitated (IP) complexes were blotted with an antibody against Rab11a (24 kDa). The last lanes (IP IgG) demonstrate that a non-immune IgG antibody failed to co-immunoprecipitate Rab11a. Endogenous Rab11a co-immunoprecipitated with WT-CFTR and ⌬F508-CFTR in stably transduced CFBE41oϪ cells and with endogenous ⌬F508-CFTR in parental CFBE41oϪ cells. The nonspecific bands, marked with an arrow, represent the light chain of the immunoprecipitating antibody (M3A7). Proteins were separated by SDS-PAGE using 15% gels. All experiments were repeated two to three times from separate cultures with similar results. Western blot experiments demonstrating that expression of endogenous Rab11a was similar in CFBE41oϪ cells stably transduced with either WT-CFTR or ⌬F508-CFTR. CFBE41oϪ cells were transfected with double-stranded small interfering RNA specific for the Rab11a sequence (siRab11a) or with the non-Non-sil. siRNA, as described under "Materials and Methods." B, summary of Western blot experiments demonstrating that siRab11a similarly decreased the expression of endogenous Rab11a in the WT-CFTR-and ⌬F508-CFTRtransduced CFBE41oϪ cells. C, representative Western blots demonstrating that siRab11a had no effect on the expression of endogenous Rab4, Rab5a, or ezrin. Non-sil. siRNA had no effect on the endogenous expression of Rab11a, WT-CFTR or ⌬F508-CFTR when compared with non-transfected cells (not shown). D, summary of biotinylation experiments demonstrating that siRab11a decreased the plasma membrane expression of WT-CFTR and ⌬F508-CFTR compared with the Non-sil. SiRNA control (single asterisks). Furthermore, the amount of the plasma membrane ⌬F508-CFTR was significantly lower than the amount of the plasma membrane WT-CFTR in the siRab11a-transfected cells (double asterisk), consistent with increased endocytosis of ⌬F508-CFTR, in addition to the siRab11a-induced inhibition of the endocytic recycling. Asterisks indicate p Ͻ 0.05. 3 experiments/group. FIGURE 10. Experiments performed to determine the effects of Rab11a mutants on the plasma membrane expression of ⌬F508-CFTR in stably transduced CFBE41o؊ cells. A, summary of biotinylation experiments performed to determine the effects of the FLAG-Rab11a WT, the GTP-locked FLAG-Rab11a S20V, and the GDP-locked FLAG-Rab11a S25N on the plasma membrane expression of ⌬F508-CFTR. The asterisk indicates p Ͻ 0.05. Three to four experiments were performed per group. B, representative immunoprecipitation experiment demonstrating that ⌬F508-CFTR interacts with FLAG-Rab11a WT in CFBE41oϪ cells. CFBE41oϪ cells stably expressing ⌬F508-CFTR were transiently transfected with FLAG-Rab11a WT. ⌬F508-CFTR was immunoprecipitated using antibody M3A7, and the immunoprecipitated complexes were blotted with an antibody against FLAG-Rab11a WT (24 kDa). The last lane demonstrates that a non-immune IgG antibody failed to co-immunoprecipitate (IP) FLAG-Rab11a WT. The nonspecific bands marked with an arrow indicate the light chain of the immunoprecipitating antibody M3A7. Proteins were separated by SDS-PAGE using 15% gels. The experiment was repeated two times from separate cultures, with similar results. ⌬F508 mutation has no significant effect on endocytosis but profoundly inhibits the endocytic recycling of CFTR (25). Furthermore, another study in BHK-21 cells demonstrates that, unlike WT-CFTR, ⌬F508-CFTR does not recycle to the plasma membrane unless the wild type Rab11a is overexpressed (24). By contrast, we report that the ⌬F508 mutation does not affect the endocytic recycling of CFTR in polarized human airway epithelial cells. Furthermore, both WT-CFTR and ⌬F508-CFTR undergo trafficking in the Rab11a-specific apical recycling compartment. Substantial evidence indicates that membrane trafficking events are cell type-specific (42)(43)(44)(92)(93)(94). Thus, it is not surprising that the endocytic trafficking of rescued ⌬F508-CFTR differs in polarized human airway epithelial cells such as CFBE41oϪ cells and in non-polarized fibroblasts such as BHK-21 cells. Furthermore, differences in the trafficking itineraries via the Rab11-dependent pathway in epithelial and non-epithelial cells exist (84,86,95). Thus, we believe that the difference in endocytic recycling of rescued ⌬F508-CFTR in CFBE41oϪ and heterologous cells likely represents a difference in the cellular mechanisms that regulate the endocytic trafficking of plasma membrane proteins in these cells.
We conducted several studies to determine why the ⌬F508 mutation increased endocytosis of ⌬F508-CFTR. The ⌬F508 mutation did not produce a generalized defect in clathrin-mediated endocytosis or in fluid phase endocytosis and did not affect expression of Rab5a, an important regulator of clathrin-mediated and fluid-phase endocytosis, in CFBE41oϪ cells. Our data are in general agreement with another study demonstrating that in human airway epithelial cells (JME) the ⌬F508 mutation has no effect on fluid phase endocytosis (64).
Studies in this manuscript ruled out the possibility that a hypothetical cryptic endocytic motif, predicted to facilitate endocytosis of ⌬F508-CFTR (75), contributes to the decrease in the apical membrane half-life of ⌬F508-CFTR in CFBE41oϪ cells. Factors such as position of the canonical YXXØ sequence relative to the C terminus of the protein and to the plasma membrane, the sequence features (amino acids in position Yϩ1 and Yϩ2), the distance from additional endocytic signals or acidic clusters, and the posttranslational modifications at or near the motif determine the strength of the YXXØ endocytic motifs (96 -103). One or more of these factors could be responsible for absence of the effect of YRSV (517-520) on the plasma membrane expression and half-life of rescued ⌬F508-CFTR.
Because the endocytic defect in polarized CFBE41oϪ cells was specific to ⌬F508-CFTR and was not produced by the putative cryptic endocytic motif YRSV (517-520), it is conceivable that the ⌬F508 mutation increases the binding affinity of ⌬F508-CFTR to a known or yet unidentified ⌬F508-CFTR adaptor protein in the endocytic trafficking pathway. Additional studies beyond the scope of the present work are required to identify these mechanisms.
In summary, our data provide direct evidence that in human airway epithelial cells the ⌬F508 mutation reduces the apical membrane halflife of rescued ⌬F508-CFTR by specifically accelerating endocytosis without decreasing the endocytic recycling of ⌬F508-CFTR. Moreover, the ⌬F508 mutation does not disrupt the endocytosis of BCRP, another ABC transporter, or fluid phase endocytosis. Furthermore, our data demonstrate that both WT-CFTR and ⌬F508-CFTR undergo trafficking in the Rab11a-specific, apical recycling compartment. It is tempting to speculate that proteins involved in the endocytosis and/or endocytic recycling of rescued ⌬F508-CFTR may become a therapeutic target to increase the apical membrane expression of ⌬F508-CFTR in polarized human airway epithelial cells.