CK19 stabilizes CFTR at the cell surface by limiting its endocytic pathway degradation

Abstract

Protein interactions that stabilize the cystic fibrosis (CF) transmembrane conductance regulator (CFTR) at the apical membranes of epithelial cells have not yet been fully elucidated. We identified keratin 19 (CK19 or K19) as a novel CFTR‐interacting protein. CK19 overexpression stabilized both wild‐type (WT)‐CFTR and Lumacaftor (VX‐809)–rescued F508del‐CFTR (where F508del is the deletion of the phenylalanine residue at position 508) at the plasma membrane (PM), promoting Cl^– secretion across human bronchial epithelial (HBE) cells. CK19 prevention of Rab7A‐mediated lysosomal degradation was a key mechanism in apical CFTR stabilization. Unexpectedly, CK19 expression was decreased by ~40% in primary HBE cells from homogenous F508del patients with CF relative to non‐CF controls. CK19 also positively regulated multidrug resistance–associated protein 4 expression at the PM, suggesting that this keratin may regulate the apical expression of other ATP‐binding cassette proteins as well as CFTR.—Hou, X., Wu, Q., Rajagopalan, C., Zhang, C., Bouhamdan, M., Wei, H., Chen, X., Zaman, K., Li, C., Sun, X., Chen, S., Frizzell, R. A., Sun, F. CK19 stabilizes CFTR at the cell surface by limiting its endocytic pathway degradation. FASEB J. 33, 12602–12615 (2019). www.fasebj.org

ABBREVIATIONS

ABC

ATP‐binding cassette

Ad

adenovirus

BSA

bovine serum albumin

CF

cystic fibrosis

CFTR

CF transmembrane conductance regulator

CK18

keratin 18

CK19

keratin 19

co‐IP

coimmunoprecipitation

CXCR2

C‐X‐C chemokine receptor type 2

EE

early endosome

ER

endoplasmic reticulum

F508del

deletion of the phenylalanine residue at position 508

FBS

fetal bovine serum

GFP

green fluorescent protein

HBE

human bronchial epithelial

HEK

human embryonic kidney

IBMX

isobutylmethylxanthine

IP

immunoprecipitation

Lamp1

lysosomal‐associated membrane protein 1

LE

late endosome

MESNA

mercaptoethanesulfonic acid

MRP4

multidrug resistance–associated protein 4

NMDG‐Cl

N‐methyl‐d‐glucamine chloride

PM

plasma membrane

RE

recycling endosome

rF508del

rescued F508del

Scr

scrambled

shRNA

short hairpin RNA

shScr

Scr shRNA

siRNA

small interfering RNA

SPQ

6‐methoxy‐N‐(3‐sulfopropyl) quinolinium

UPS

ubiquitin‐proteasome system

VX‐809

Lumacaftor

WT

wild type

The cystic fibrosis (CF) transmembrane conductance regulator (CFTR), a member of the ATP‐binding cassette (ABC) transporter superfamily, is a cAMP/PKA‐regulated chloride channel at the apical membranes of epithelial cells lining the airways and other epithelial tissues (1). Mutations of the cftr gene led to CF, a life‐threatening disease in Caucasian populations (2). In patients with CF, the most common mutation is the deletion of the phenylalanine residue at position 508 (F508del). F508del‐CFTR is a folding/processing mutant found in over 90% of patients with CF on at least 1 allele (3, 4). Instead of trafficking to the plasma membrane (PM), more than 99% of this CFTR mutant protein is degraded by the ubiquitin‐proteasome system (UPS) (5, 6). A fraction of F508del can return to its folding pathway with the help of correctors [e.g., endoplasmic reticulum (ER)‐targeted corrector of Lumacaftor (VX‐809)] (7, 8). If it escapes degradation, F508del can reach the PM and function as a chloride channel (3, 9, 10). However, the rescued F508del (rF508del) is also impaired by a reduced half‐life at the PM because of cell surface quality control involving its modification by ubiquitin, internalization from the cell surface, and degradation by the lysosome (11–13). A fraction of wild‐type (WT)‐CFTR is also degraded in the ER by the UPS; however, most WT‐CFTR traffics to the PM. Following its endocytosis (14), WT‐CFTR enters early endosomes (EEs) and is transferred to recycling endosomes (REs) for return to the PM, thus stabilizing the WT protein in the cell periphery (15).

Rab GTPases are involved in the intercellular trafficking and apical recycling of CFTR. The initial internalization of WT‐CFTR from the PM into the EE is Rab5 dependent, and this is followed by Rab11‐dependent transport to the RE or Rab7‐dependent transport to late endosomes (LEs) (14, 16). However, it is difficult for the internalized rF508del to recycle back to the PM because of its continuing unstable structure and rapid ubiquitination (17–19). Thus, in addition to its chloride channel function, membrane stability is also needed to maintain a requisite number of rescued (rF508del) CFTRs for anion secretion (20, 21). Unfortunately, the protein interactions that control the stabilization of CFTR at the apical membrane are not adequately elucidated (22–24).

Keratin 19 (CK19 or K19), the smallest class I intermediate filament protein, is expressed predominantly in epithelial cells of lung, intestine, and liver (25–27). In recent years, there have been few published studies of the cellular functions of CK19, which is usually cited as a biomarker for cellular differentiation and cancer (25, 26, 28–31). Although CK19′s relative, keratin 18 (CK18), a type II keratin, has been previously identified as a CFTR‐binding protein, which stabilizes PM CFTR by accelerating its Rab11‐mediated recycling rate, as we show here, CK19′s regulation of CFTR is also significant and distinguishable from that of CK18. CK19 and CK18 usually are reported to be altered in parallel in disease development (32, 33). Together, these findings suggest the hypothesis that CK19 is involved in the regulation of apical CFTR channel density.

Here, we identify CK19 as a novel CFTR‐interacting protein in human airway epithelial cells. CK19 improved the membrane stability of both WT‐CFTR and VX‐809–rescued F508del by reducing its Rab7‐mediated lysosomal degradation. We further show that CK19 expression was decreased significantly in primary human bronchial epithelial (HBE) cells from homogenous F508del patients with CF. In addition, CK19 increased the expression of multidrug resistance‐associated protein 4 (MRP4) at the PM, indicating that CK19 may be a common regulator of the apical expression of ABC proteins.

MATERIALS AND METHODS

Cell culture, antibodies, and chemicals

Calu‐3 cells were cultured in DMEM/Ham's F‐12 medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 15% fetal bovine serum (FBS). Human embryonic kidney (HEK) 293 and HeLa cells were cultured in DMEM with 10% FBS. CFBE41o‐parental cells were cultured using minimum essential medium (MEM; Thermo Fisher Scientific) medium with 10% FBS. CFBE‐WT and CFBE‐F508del cells were cultured in CFBE41o‐ parental medium with 0.5 or 1 μg/ml puromycin. All cells were maintained in a humidified atmosphere containing 5% CO2 in air at 37°. Primary epithelial cell lysates from individuals with and without CF were obtained from the CF Research Center Airway Cell Core at the University of Pittsburgh. CK19 knockdown and scrambled (Scr) short hairpin RNA (shRNA) (referred to as shScr)‐expressing HEK293 cells were selected by using HEK medium with 1 μg/ml puromycin and maintained in HEK medium with 0.5 μg/ml puromycin. Calu‐3 and CFBE‐F508del cells were grown on Transwell filters (3470; Corning, Corning, NY, USA) as previously described by Marozkina et al. (34). CFTR antibodies M3A7 and L12B4 were purchased from MilliporeSigma (Burlington, MA, USA), 24–1 was from R&D Systems (Minneapolis, MN, USA), 596 and 217 were from the CFTR antibody service maintained at the University of North Carolina at Chapel Hill (Chapel Hill, NC, USA). Actin and myc antibodies were purchased from MilliporeSigma, hemagglutinin from Abcam (Cambridge, MA, USA), and CK19, CK18, Rab5, Rab7, Rab11, lysosomal‐associated membrane protein 1 (Lamp1), syntaxin 6, cytochrome c oxidase complex IV (COXIV), MRP4, calnexin, and ubiquitin were purchased from Cell Signaling Technology (Danvers, MA, USA). Leupeptin was from MilliporeSigma, and VX‐809, GlyH101, forskolin, isobutylmethylxanthine (IBMX), and mercaptoethanesulfonic acid (MESNA) were purchased from MilliporeSigma. Small interfering RNAs (siRNAs) were from Dharmacon (Lafayette, CO, USA).

Plasmid constructs, DNA constructs

WT‐CFTR and F508del in pcDNA3.1 vector were previously described in Sun et al. (35). The CK19 cDNA sequence was amplified by PCR and inserted into pcDNA3.1 (+) at EcoR I and EcoRV sites. MRP4 and triple HA‐tagged C‐X‐C chemokine receptor type 2 (CXCR2) plasmids were a gift from Dr. Chunying Li (36, 37). Myc‐ubiquitin was purchased from Addgene (Watertown, MA, USA).

Recombinant lentiviral shRNAs

Oligos were cloned into pLKO.1 vector as previously described, which were sequenced and are shown in Table 1 (35).

Table 1.

shRNAs targeting CK19

No.	Sequence, 5′–3′
1	CGAGAAGCTAACCATGCAGAA
2	GCGAAGCCAATATGAGGTCAT
3	CGAACCAAGTTTGAGACGGAA
4	CAGGAAGATCACTACAACAAT
5	CCAGCGGCTCATGGACATCAA

The shRNA sequences of CK19.

Recombinant adenovirus

Recombinant adenoviruses were made as previously described by Howard et al. (38). Briefly, the viruses were generated using the ViraPower Adenoviral Expression System (Thermo Fisher Scientific). Crude viral stocks were then amplified twice, and recombinant adenoviruses were purified with CsCl ultracentrifugation and titrated by serial dilution.

Liquid chromatography tandem mass spectrometry analysis

The peptides were first separated on a reverse‐phase C18 column with a 45‐min gradient using the Dionex UltimateTM HPLC system. Mass spectrometry (MS) and tandem MS (MS/MS) spectra were then acquired on an QStar XL Mass Analyzer (Applied Biosystems, Foster City, CA, USA) using the information‐dependent acquisition mode. An MS scan was performed from m/z 300‐1, 500 for 1 s followed by production scans on the 2 most‐intense multiply charged ions. Peaklists were submitted to the Mascot server (Matrix Science, Boston, MA, USA) to search against the NCBInr database for Homo sapien proteins (https://www.ncbi.nlm.nih.gov/refseq/about/nonredundantproteins/) with carbamidomethyl (C) as a fixed modification and oxidation (M) and N‐acetylation (protein N terminus) as variable modifications (39).

Coimmunoprecipitation and Western blot analysis

Cell lysate in lysis buffer [50 mM 4‐(2‐hydroxyethyl)‐1‐piperazineethanesulfonic acid (HEPES) (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P‐40, 10% glycerol] was precleared with protein A‐ and G‐Sepharose beads (Thermo Fisher Scientific). The precleared cell lysates were mixed with 10 μg of the indicated antibodies and 25 μl of washed protein A‐ and/or G‐Sepharose beads and incubated overnight at 4° with gentle rotation. The immunocomplexes were resuspended in SDS sample buffer and subjected to SDS‐PAGE gel and immunoblotting.

Cell surface biotinylation

Cell surface proteins were biotinylated with EZ‐linked sulfosuccinimidyl‐20(biotinamido)ethyl‐1,3‐dithiopropionate (Sulfo‐NHS‐SS‐Biotin; Thermo Fisher Scientific). Cells were chilled on ice and washed with ice‐cold PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄, 1.4 mM KH₂PO₄, pH 7.4) containing 0.1 mM CaCl₂ and 1 mM MgCl₂ (PBS‐CM) and incubated with 1 mg/ml EZ‐Link Sulfo‐NHS‐SS‐Biotin in PBS (pH 8.6) for 60 min. The cells were rinsed with cold PBS‐CM 3 times, and free biotin molecules were quenched with PBS containing 1% bovine serum albumin (BSA) for 10 min before cell lysis and streptavidin‐agarose precipitation. The precipitants were subjected to immunoblotting.

CFTR internalization assay

Cell surface proteins were biotinylated as described in the previous section. The cells were warmed to 37° for specified periods. At the end of the indicated time points, the cells were rapidly cooled by rinsing with cold PBS‐CM and replaced on ice. Biotin molecules at the cell surface were cleaved off with a membrane impermeant MESNA (50 mM) in stripping buffer (150 mM NaCl, 1 mM EDTA, 0.2% BSA, and 20 mM Tris, pH 8.6) and washed with the buffer 2 additional times. The cell lysates were precipitated with streptavidin‐agarose followed by immunoblotting.

Lysosome enrichment fractionation

CK19 and Scr knockdown HEK293 cells transfected with WT‐CFTR were cultured for 72 h and then with leupeptin for about 24 h before harvest. The cells were then washed in PBS and were scraped into 1 ml cold PBS with a rubber policeman and pelleted in a polystyrene tube [300 g, 5 min in a Kubota centrifuge (Kubota, Osaka, Japan)]. Lysosomes were purified using the Lysosome Enrichment Kit (89839; Thermo Fisher Scientific).

6‐Methoxy‐N‐(3‐sulfopropyl) quinolinium assays

Cells were cultured on glass coverslips for 2 wk, and cAMP‐stimulated anion efflux was determined from the intensity response of the halide‐sensitive fluorophore, 6‐methoxy‐N‐(3‐sulfopropyl) quinolinium (SPQ), to cAMP agonists (10 μM forskolin plus 100 μM IBMX). Cells were loaded with the fluorophore and exposed sequentially to solutions containing NaI, NaNO₃, NaNO₃ plus agonists, and again to NaI, as previously described in refs. 15 and 40. The time course of fluorescence intensity, relative to that at zero time, F₁/F₀, provided normalization for dye loading. The data were obtained from 10 to 15 cells on each of 2 coverslips from each patient; each experiment was repeated at least twice.

Short‐circuit current recordings

Short‐circuit currents were measured as previously described by Rasgado‐Flores et al. (41). In brief, CFBE‐F508del cells were infected with adenovirus (Ad)‐CK19 or Ad–green fluorescent protein (GFP) as the control. After 24 h of infection, cells were cultured for about 5 d and then mounted in modified Ussing chambers until monolayer resistance was more than 800 mΩ.cm². Before use, cells were incubated with 3 μM corrector VX‐809 or vehicle (DMSO) for 24 h. The cultures were continuously short‐circuited.

Whole‐cell current recordings

Whole‐cell current recordings were performed using transfected CFBE41o‐ cells as previously described (42). Briefly, currents were sampled by a 200B Axopatch amplifier controlled by Clampex 8.1 software through a Digidata 1322A acquisition board (Axon Instruments, Union City, CA, USA). Solutions were maintained at 37° at a flow rate of 2 ml/min. The imaging chamber was mounted on the stage of a Nikon Diaphot microscope (Nikon, Tokyo, Japan) equipped with standard illumination, and a xenon lamp (Nikon, Tokyo, Japan) with GFP filter cube (Ex 485/Em 550) permitted identification of expressing cells. Seal resistances exceeded 8 GΩ, and pipette capacitance was compensated; all experiments used a standard voltage‐clamp protocol with a holding potential of –40 mV. N‐methyl‐d‐glucamine chloride (NMDG‐Cl) was used for both the bath and pipette solutions to isolate chloride currents. The bath solution was (mM) 140 NMDG‐Cl, 10 HEPES, 1 MgCl₂, 1.5 CaCl₂, 5 glucose, pH 7.3, and the pipette solution was (mM) 140 NMDG‐Cl, 10 HEPES, 1 MgCl₂, 5 glucose, 1 EGTA, pH 7.2, and it also contained 1 mM Mg‐ATP and 100 μM GTP. NMDG‐glutamate was added to the bath solution to generate a 25 mOsm gradient that was needed to obviate cell swelling. Pipettes using thin‐wall borosilicate glass were pulled to tip diameters of 1–2 μm (access resistance <4 MΩ).

Confocal microscopy

Immunofluorescence staining of filter‐grown cells was performed as previously described in Sun et al. (35). Briefly, cells were fixed in 4% paraformaldehyde and permeabilized with a mixture of 4% paraformaldehyde and 0.1% Triton X‐100. After blocking with purified goat serum, the monolayers were incubated in the appropriate primary antibodies and subsequently incubated with FITC (green) or rhodamine (red)‐labeled secondary antibodies (1:1000; Molecular Probes, Eugene, OR, USA) for 1 h. The cells were washed 3 times with buffer A (0.5% BSA and 0.15% glycine at pH 7.4 in PBS) after each step. The filters were mounted on glass coverslips and imaged by confocal microscopy. Images were acquired with a laser scanning confocal microscope [SP8 Harmonic Compound (HC) Field Planarity (PL) Apochromat Objective (APO) Confocal Scanning 2 (CS2) ×63, numerical aperture 1.4, oil; Leica Microsystems, Wetzlar, Germany] and exported to ImageSpace (Molecular Dynamics, Sunnyvale, CA, USA) for subsequent reconstruction and processing.

Pulse chase

HEK293 cells were cotransfected with WT‐CFTR with or without CK19. Cells were labeled for 30 min with [³⁵S]‐Met and ‐Cys (120 μC_i/ml) and chased in isotope‐free medium for the indicated times followed by CFTR immunoprecipitation (IP). The signals were detected by autoradiography.

Statistics analysis

The data were analyzed and compared using the unpaired, 2‐tailed Student's t test. Statistical tests were statistically significant when P < 0.05.

RESULTS

CK19 is a CFTR‐interacting protein that facilitates apical membrane CFTR expression in airway epithelial cells

To investigate CFTR‐interacting proteins in human airway epithelial (HBE) cells, we precipitated CFTR from the total cellular pool of CFBE‐WT and CFBE‐F508del cell lines, human airway epithelial cell lines that stably express WT‐CFTR or F508del (3, 40). We used CFTR antibody 24‐1 or a combination of CFTR antibodies L12B4 and M3A7 for IP. Input of CFBE‐WT and CFBE‐F508del was shown in the lower panel of Fig. 1A . A protein of ~40 kDa was precipitated in both cell lines (Fig. 1A , upper panel). The ~40 kDa protein band was excised and analyzed by liquid chromatography MS/MS after in‐gel digestion with trypsin. This protein was identified as CK19.

Figure 1.

CK19 is a CFTR‐interacting protein in human airway cells with effects on CFTR expression. A) Lysates from 2 different airway cell lines, CFBE‐F508del and CFBE‐WT, were mixed with different CFTR antibodies (L12B4 and M3A7), 24‐1, or IgG and precipitated with protein A and G beads. Bound proteins were fractionated on 4–20%‐gradient SDS‐PAGE and stained with Coomassie blue. The labeled proteins were excised from the gel and subjected to liquid chromatography MS/MS analysis. Proteins from cell lysates (100 mg) were subjected to immunoblot with M3A7. Lower panel shows the input (20 μg each sample) of upper panel; the left lane is CFBE‐WT cell, and the right one is CFBE‐F508del cell. B) Lysates from Calu‐3 cells were subjected to IP with M3A7 antibody or IgG and examined by immunoblot with CK19 antibody. C) Polarized Calu‐3 cells were stained with antibodies against CFTR (596) (a) or CK19 (b) and examined by confocal microscopy in the apical plane of the cells (left panels of a, b). Digitally rendered x‐z plane reconstructions of a set of 10 sections through each x‐y view were obtained (right panels of c). Scale bars, 5 μm. D) WT‐CFTR or F508del plasmids were cotransfected into HEK293 cells with GFP or CK19. Cells were lysed 2 d post‐transfection and subjected to immunoblot with the indicated antibodies. E) Quantification of the CFTR signals in D (means ± sem, n = 3). HC, antibody heavy chain.

To confirm the interaction between CK19 and CFTR, we performed a similar coimmunoprecipitation (co‐IP) in Calu‐3 cells, a human lung cancer epithelial cell line exhibiting WT‐CFTR expression (40), with the M3A7 CFTR antibody. Endogenous CK19 from Calu‐3 cells was also pulled down with CFTR (Fig. 1B ), demonstrating the interaction between CK19 and CFTR in these human airway cell lines.

Using confocal immunofluorescence microscopy, we found that CFTR expression was not homogenous in polarized Calu‐3 cells (Fig. 1Ca ) (40), yet CK19 was expressed with an apical distribution in Calu‐3s, and CK19 and CFTR were expressed in the same cells (Fig. 1Cb ). The reconstructed x‐z and y‐z sections of these polarized Calu‐3 cells were consistent with CK19 colocalization with CFTR in the apical membrane (Fig. 1Cc ); see also Fig. 2A, B .

Figure 2.

CK19 stabilizes WT‐CFTR at the PM. A) HEK293 cells were transfected with WT‐CFTR and CK19 or GFP, and after 48 h, proteins at the cell surface were labeled with 1 mg/ml Sulfo‐NHS‐SS‐Biotin. CFTR was determined by immunoblot with CFTR antibody 217. Equal volumes of the supernatant were immunoprecipitated with CK19 and actin antibodies. B) Cell surface CFTR from A was quantified (means ± sem, n = 3). C) Pulse‐chase experiments performed using HEK293 cells cotransfected with WT‐CFTR and CK19 or GFP. The cells were labeled with [³⁵S]‐Met and ‐Cys and chased at the indicated times. The labeled proteins were immunoprecipitated for CFTR (M3A7 + 24‐1 antibodies) and detected by autoradiography. D, E) Quantitation of the pulse‐chase experiments from C; density values are expressed as bands C/C + B (percent maturation, D) or the percent of band B density at chase time = 0 (100%, E). F) HEK293 cells stably expressing WT‐CFTR were transfected with CK19 or GFP. Forty‐eight hours post‐transfection, proteins at the cell surface were labeled with biotin. Cells were collected at the indicated times, and biotin remaining on the cell surface was cleaved off by MESNA. Biotinylated proteins in cell lysate were precipitated with streptavidin beads, fractionated on SDS‐PAGE, and CFTR identified by immunoblot with ab 217. G) HeLa cells were infected with the indicated combinations of Ad‐CFTR, Ad‐CK19, or Ad‐GFP at multiplicity of infection = 50. Forty‐eight hours postinfection, cAMP‐stimulated halide efflux was monitored by SPQ dequenching assays; cAMP/PKA stimulation was performed using 10 μM forskolin plus 1 mM IBMX. The relative fluorescence intensity, F₁/F₀, is plotted as a fraction of the mean intensity obtained from 2 time points just after recordings were initiated. Each trace (means ± sem) represents data from at least 2 coverslips, with 10–15 cells recorded from each. H) Peak cAMP‐stimulated whole‐cell currents recorded from CFBE‐WT after infection with CK19‐ or GFP‐expressing adenovirus, normalized for apparent whole‐cell surface area (pA/pF) (means ± sem, n = 3).

To understand the implications of the CK19‐CFTR interaction, we initially coexpressed WT‐CFTR or F508del with CK19 or GFP (negative control) in HEK293 cells and analyzed protein expression by immunoblotting. Overexpression of CK19 protein resulted in about 2.5‐ and 1.7‐fold increments in the mature and immature expression of WT‐CFTR, respectively, and a 1.5‐fold increase in that of F508del (Fig. 1D, E ). These data suggest that CK19 supports apical membrane CFTR expression. CFTR's colocalization with CK19 is consistent with the observed physical interactions between these proteins, as shown in our IP experiments (Fig. 1A, B ).

CK19 stabilizes WT‐CFTR at the cell surface and augments its cAMP‐dependent anion efflux

The increase in mature CFTR found in cells after WT‐CFTR and CK19 coexpression (Fig. 1D ) suggests that CK19 may stabilize CFTR at the PM. To address this hypothesis, we generated an HEK293 cell line transiently overexpressing WT‐CFTR and CK19 and then analyzed cell surface CFTR using a surface biotinylation assay. Overexpression of CK19 elicited an ~4.5‐fold increase in mature CFTR (band C) at steady state (Fig. 2A, B ).

To explore the mechanism of enhanced PM CFTR, first we performed pulse‐chase experiments in which HEK293 cells were transfected with WT‐CFTR and CK19 or GFP. After labeling with [³⁵S]‐Met and ‐Cys, the cells were collected at designated time points and subjected to CFTR IP. The maturation ratio (bands C/C + B) of WT‐CFTR coexpressed with CK19 was clearly higher than that of WT‐CFTR coexpressed with GFP; from 4 to 22 h, the difference was about 2.8‐fold (Fig. 2C, D ). However, no difference in immature B band of WT‐CFTR was observed with CK19 overexpression (Fig. 2C, E ). These results indicate that CK19 improved the membrane stability of WT‐CFTR without affecting the kinetics of immature CFTR degradation by the UPS. To further support the conclusion that CK19 prolonged CFTR expression on the PM, we measured the WT‐CFTR internalization rate. CFBE41o‐parental cells were infected with recombinant adenovirus carrying WT‐CFTR (Ad‐CFTR) or GFP (Ad‐GFP) and subjected to surface biotinylation. Ad‐GFP was used to monitor the transduction efficiency and as the negative control in those experiments (Supplemental Fig. S1A). Total WT‐CFTR expression was increased about 2.5‐fold in cells with Ad‐CK19 at the steady state compared with cells infected with Ad‐GFP (Supplemental Fig. S1B). Next, those cells were subjected to surface biotinylation. After collecting biotinylated cells at the designated time points, biotin remaining at the cell surface was cleaved by incubating the cells with a membrane impermeant‐reducing agent, MESNA. The internalization of PM WT‐CFTR was slowed in Ad‐CK19–infected cells vs. Ad‐GFP–infected cells (Fig. 2F ). At the 10 h time point, CFTR internalized by Ad‐GFP–infected cells was greater than that in the Ad‐CK19–infected cells (39 vs. 21%; n = 3, P < 0.01). The difference was even more pronounced at the 15‐h time point, when 55% of CFTR was cleared from the cell surface of control cells compared with 30% of CK19‐overexpressing cells (Fig. 2F ). These results suggest that CK19 stabilized WT‐CFTR on the PM by inhibiting its internalization.

We also found that the membrane stability of WT‐CFTR was reduced in cells after CK19 knockdown using RNA interference. We made 5 independent lentivirus vectors containing shRNA targeting CK19 (Table 1) and stably expressed these shRNAs in HEK293 cells. Depletion of endogenous CK19 in these cells was detected by immunoblotting. Compared with the shScr control cells, shRNA targeting CK19 expression (shCK19 #1, #3, and #5) reduced CK19 by ~37, 53, and 55%, respectively (Supplemental Fig. S1C). Stable CK19 knockdown and Scr control HEK293 cell lines were produced with shCK19 #1, #5, or shScr lentivirus infection. After WT‐CFTR transfection, the cell surface pool of CFTR was biotinylated and quantified. As illustrated in Supplemental Fig. S1D, reduced CK19 expression elicited 80 and 55% decreases in expression relative to that in control cells.

Next, we measured cAMP‐stimulated anion efflux using the halide‐sensitive fluorophore SPQ, which was loaded into HeLa cells expressing CFTR with or without CK19, expressed using adenoviral vectors; GFP was expressed separately as a control. The peak fluorescence intensity from HeLa cells infected with Ad‐CFTR plus Ad‐CK19 increased by 2.2‐fold compared with the cells transduced with Ad‐CFTR plus Ad‐GFP (Fig. 2G,H ).

Finally, whole‐cell patch‐clamp recordings were performed under ionic gradient conditions designed to isolate choride currents across the PMs of CFBE airway cells. Under these conditions, cells expressing WT‐CFTR with CK19 showed a 2.5‐fold increase in Cl current during stimulation by forskolin to raise PKA/cAMP levels (Fig. 2H ). These results indicate that CK19 not only improves the membrane stability of PM‐localized WT‐CFTR but also elicits the PM stabilization of functional CFTR.

CK19 stabilizes VX‐809–rescued F508del at the cell surface and enhances chloride secretion

We showed that CK19 stabilized PM WT‐CFTR expression and function in model cell lines (Fig. 2). To determine whether these findings would translate to rF508del, we first tested HEK293 cells coexpressing F508del‐CFTR with CK19 or GFP and treated with 3 μM VX‐809 for 24 h. On average, increased expression of CK19 resulted in the accumulation of both immature and mature forms of F508del with VX‐809 treatment (1.8‐ and 2.3‐fold, respectively). When CK19 was overexpressed alone, the mature form of F508del increased to levels comparable to those observed with VX‐809 treatment alone (Fig. 3A, B ). Furthermore, CK19 improved the expression of both mature and immature forms of F508del in a dose‐dependent manner. The increase of the mature form of F508del by CK19 was amplified by VX‐809 treatment (Supplemental Fig. S2).

Figure 3.

CK19 facilitates cell surface expression and chloride secretion by rF508del. A) HEK293 cells were transfected with F508del and CK19 or GFP. Forty‐eight hours post‐transfection, cells were incubated for 24 h with VX‐809, and then CFTR expression was determined by immunoblot from 40 μg of loaded protein. B) Quantified CFTR bands from A (means ± sem, n = 3). C) Pulse chase of F508del performed and analyzed as in Fig. 2 C, E. D) Quantitation of the composite pulse‐chase data from C (mean ± sem, n = 2). E) CFBE‐F508del cells were infected with Ad‐CK19 or Ad‐GFP. Twenty‐four hours post‐transfection, cells were incubated at 27° for an additional 24 h to amplify F508del expression level; then, proteins at the cell surface were labeled with biotin. Cells were collected at indicated times, and CFTR identified by immunoblot with ab 217. Comparable horizontal panels shown were from the same film. F) Quantitation of the time course of mature F508del from E (means ± sem, n = 3). G) CFBE‐F508del cells were infected with Ad‐CK19 or Ad‐GFP for 24 h, seeded on Transwell filters for 24 h, and then incubated with 3 μM VX‐809 or vehicle (DMSO) for 24 h. Short‐circuit currents (I_sc ) were measured using transepithelial Ussing chamber assays. H) Quantified currents from experiments as in G, taken at peak current values (means ± sem, n =3). I) HBE cells were derived from airways from homozygous F508del patients with CF obtained following lung transplantation. WT‐CFTR (non‐CF) primary epithelial cells were usually from patients with idiopathic pulmonary fibrosis or from nondisease donor lungs. Cells were cultured in flasks prior to collection for keratin immunoblotting. J) Normalized expression of CK19 and CK18 from primary CF and non‐CF HBE cells (means ± sem, n =6).

Next, F508del was analyzed by pulse chase, but no differences were found between cells coexpressing F508del with CK19 or GFP. This result was consistent with the lack of CK19 effect on the immature form of WT‐CFTR (Figs. 2C, E and 3C, D ), indicating that CK19 is not involved in CFTR proteasomal degradation of the immature forms of either WT or mutant forms of CFTR (10, 43). In addition, the dynamic change of membrane rF508del was estimated using biotinylation. CFBE‐F508del cells were infected with Ad‐CK19 or Ad‐GFP on d 1 and then were cultured at 27° for the following day. As shown in Fig. 3E, F , rF508del was more stable in cells after Ad‐CK19 infection, especially after 2 h. This finding is in accordance with that in CFBE‐WT cells (Fig. 2C, F ), indicating that CK19 improved the PM stability of both WT‐CFTR and rF508del.

To evaluate the functional consequences of F508del recovery by CK19 and VX‐809, we measured transepithelial currents of CFBE‐F508del cells, which can produce high levels of F508del and form a polarized monolayer with high electrical resistance (44). Cells were infected with Ad‐CK19 or Ad‐GFP, and after 24 h, they were incubated with VX‐809 or DMSO (vehicle control) for an additional 24 h before mounting in Ussing chambers. Overexpressing CK19 generated an ~12‐fold increase in forskolin‐stimulated Cl^– current in cells that were also treated with VX‐809 (Fig. 3G, H ). Taken together, our data illustrate a significant role for CK19 in both rF508del membrane stability and its ability to support chloride secretion.

CK19 expression is reduced in primary CF HBE cells

As CK19 increased corrector rF508del in HBE cell lines, we analyzed CK19 expression in primary HBE cells from 6 homozygous F508del patients with CF and donors without CF (Fig. 3I, J ) (45). Patient information in Table 2 includes age at transplant, gender, and disease basis of transplant (donors without CF). Immunoblots from these lines show that CK19 protein expression was attenuated by ~40% in HBE cells from patients with CF compared with individuals without CF. Another keratin, CK18, was not significantly reduced in CF HBE cells (Fig. 3I, J ).

Table 2.

Information from donors with and without CF

Donors	Age a	Gender	Genotype
Non‐CF
HBE‐980	64	F	COPD
HBE‐983	59	M	IPF
HBE‐985	50	F	Scleroderma
HBE‐986	66	M	COPD
HBE‐944	40	F	LAM
HBE‐968	67	M	COPD
CF
CF‐153	29	M	dF508/dF508
CF‐181	25	F	dF508/dF508
CF‐184	31	M	dF508/dF508
CF‐196	42	F	dF508/dF508
CF‐205	47	M	dF508/dF508
CF‐209	37	F	dF508/dF508

Patient data for CK19 expression. COPD, chronic obstructive pulmonary disease; F, female; IPF, idiopathic pulmonary fibrosis; LAM, lymphangioleiomyomatosis; M, male. ^a Age at transplant.

CK19 attenuates Rab7‐mediated CFTR lysosomal degradation

Our results show that CK19 improved CFTR membrane stability by reducing the kinetics of its internalization. As Rab GTPases are known to modulate the endocytic trafficking of CFTR (14, 16, 46–49), we checked the interaction of CFTR with Rab GTPases. For this purpose, we performed co‐IPs in CFBE‐WT cells and determined the ability of WT‐CFTR to co‐IP–associated endosomal markers and vice versa. The data indicate that under these conditions, CFTR interacted with Rab7A from CFBE‐WT cells but not with Rab5C or Rab11 (Supplemental Fig. S3A). It has been reported that Rab7A facilitates CFTR trafficking from EEs to LEs/lysosome for degradation (16); therefore, our results suggest that CK19 abrogates the endocytic trafficking of CFTR toward LE/lysosome compartments.

To evaluate the role of Rab7A in CFTR handling following its internalization from the PM, we performed WT‐CFTR co‐IPs in CK19 knockdown and Scr HEK293 cell lines using the same amounts of WT‐CFTR protein as the input and blotted for Rab isoforms (Fig. 4A lowest panel). Interestingly, more Rab7A was coprecipitated by CFTR antibody 217 from CK19 knockdown cells than that from the Scr controls (Fig. 4A , uppermost panel). These results suggest that interaction of Rab7A and WT‐CFTR was enhanced by CK19 knockdown. We also compared the interaction of Rab7A with rF508del vs. WT‐CFTR. To obtain comparable amounts of mature CFTR in the input, the lysates of CFBE‐WT and VX‐809–treated CFBE‐F508del cells were adjusted in a ratio of 1:5 (Fig. 4B , uppermost and middle panels). Comparable amounts of Rab7A were precipitated by CFTR antibody from both cell lysates shown in the figures (1:1.2), whereas ~5‐fold more Rab7A was precipitated by Rab7A antibody from VX‐809–treated CFBE‐F508del cells than that from CFBE‐WT cells, which was in accordance with the amount of precipitant in the input (Fig. 4B , lowest panel). Rab7A precipitated by the CFTR antibody was dependent on the quantity of the mature form of CFTR. Thus, the interaction between mature CFTR and Rab7A was an indication of internalized CFTR's fate.

Figure 4.

CK19 blocks CFTR degradation by lysosomes. A) HEK293 cells were cotransfected with WT‐CFTR, shCK19, or shScr. After normalization to the same amount of CFTR, cell lysates were subjected to IP with ab 217 or IgG. Precipitated proteins were examined by immunoblot with antibodies to Rab7A, Rab5C, or Rab11. B) CFBE‐F508del cells were incubated with 3 μM VX‐809 for 24 h. Before IP, lysates from CFBE‐WT and rescued CFBE‐F508del cells were adjusted to contain the same amount of mature WT or rF508del (total cell lysate = 1:5); then, they were precipitated with CFTR ab 217 or Rab7 antibody. Bound proteins were separated on 4–20%‐gradient SDS‐PAGE, transferred to PVDF membrane, and detected by immunoblotting with 217 or Rab7 antibody. C) Cell surface or cytosolic expression of WT‐CFTR in HEK293 cells was detected by Western blot after Rab7A was reduced by its siRNA for 48 h; siScr as control. Comparable horizontal panels shown were from the same film. D) CK19 knockdown or Scr control HEK293 cell lines were transfected with WT‐CFTR; after 24 h, cells were incubated with 2 μg/ml leupeptin for another 24 h. Lysosome fragments were purified as described in Materials and Methods and detected by immunoblot with LAMP1 antibody as lysosome marker. E) The lysosomes from D were subjected to IP with CFTR ab 217 and examined by immunoblot with ubiquitin (Ub) antibody. F) CFBE parental cells were cotransfected with WT‐CFTR and shCK19 or shScr, and after 24 h, they were incubated with leupeptin for another 24 h. The cells were stained with antibodies to CFTR (596, mouse monoclonal) and LAMP1 (rabbit polyclonal) and subsequently incubated with anti‐mouse rhodamine or anti‐rabbit FITC‐labeled secondary antibodies. The cells were examined by confocal microscopy. Scale bars, 5 μm.

To further confirm the specific role of Rab7A in CFTR degradation, Rab7A expression was reduced using an siRNA (siRab7A). The results showed that WT‐CFTR protein was dramatically increased, 7.8‐ and 13.6‐fold, in the whole‐cell lysates of CK19 knockdown vs. control (shScr) cells with reduced Rab7A expression, respectively (Fig. 4C , second panel). In addition, the cell surface WT‐CFTR pool with Rab7A knockdown was detected by biotinylation. The PM WT‐CFTR increased by 5.6‐ and 10.7‐fold in the CK19 knockdown cells vs. the control cells with decreased Rab7A expression, respectively (Fig. 4C , uppermost panel). CFTR expression increased dramatically in both membrane and cytoplasmic fractions with Rab7A inhibition, but the increment of CFTR in CK19 knockdown cells was less than that in Scr control cells (Fig. 4C ). These results strongly suggest that Rab7A promoted the lysosomal degradation of CFTR internalized from the PM and that the impact of Rab7A could be offset by CK19.

To further confirm that CK19 attenuated WT‐CFTR lysosome‐mediated degradation, we analyzed the quantitative change in WT‐CFTR in lysosomes induced by CK19 down‐regulation. The lysosomes were isolated from WT‐CFTR overexpressing CK19 knockdown and Scr HEK293 cells treated with the lysosome inhibitor leupeptin for 24 h. The purity of the lysosomes was supported by immunoblot with the lysosome marker (LAMP1, Fig. 4D , lower panel) and membrane organelle markers shown in Supplemental Fig. S3B. Compared with Scr cells, WT‐CFTR was reduced to ~72% of control in lysosomes from CK19 knockdown cells, whereas the reduction in WT‐CFTR was reversed, for the most part, when the cells were treated with leupeptin. These findings indicate that reduced CK19 expression accelerated the lysosomemediated degradation of WT‐CFTR (Fig. 4D , upper panel). The total WT‐CFTR in the CK19 knockdown and scramble cells is shown in Supplemental Fig. S3C (uppermost panel). CK18 expression was not affected by CK19 knockdown (Supplemental Fig. S3C, middle panel). Finally, ubiquitylation of WT‐CFTR in lysosomes was analyzed using equal amounts of lysosomal fragments (shown in Fig. 4D , lower panel). Ubiquitylation of WT‐CFTR was dramatically increased in the lysosomes from CK19 knockdown cells treated with leupeptin (Fig. 4E ). These data indicate that ubiquitin‐tagged WT‐CFTR encountered rapid degradation in lysosomes with reduced CK19 expression and demonstrated that CK19 protected WT‐CFTR from degradation by quality control mechanisms in the cell periphery.

A change in the subcellular location of WT‐CFTR was also associated with inhibition of CK19 expression and detected using confocal microscopy. CFBE41o‐ parental cells were cotransfected with WT‐CFTR and either shCK19 or shScr for 48 h, before treatment with leupeptin for an additional 24 h. CFTR and the lysosome marker LAMP1 were immunostained with #596 CFTR antibody and LAMP1 antibody, respectively. In control (shScr) cells, WT‐CFTR was mainly located in perinuclear regions and on the PM (Fig. 4Fa, c ). WT‐CFTR increased in both the lysosomes and at the PM after leupeptin treatment (Fig. 4Fd, f ). However, in CK19 knockdown cells, the expression of WT‐CFTR at the PM was minimal, and less WT‐CFTR was present in the lysosomes (Fig. 4Fg, i ) compared with control cells (Fig. 4Fa, c ). In CK19 knockdown cells treated with leupeptin, membrane WT‐CFTR was partially recovered (Fig. 4Fj, l ) but was still reduced compared with that in control cells (Fig. 4Fa, c ). Notably, most WT‐CFTR was localized in the lysosomes under these conditions (Fig. 4Fl ). These observations underscore the finding that PM CFTR is unstable in the absence of CK19 and the likely explanation that CFTR is rapidly degraded in lysosomes after its internalization. These observations further emphasize the impact of CK19‐mediated stabilization of WT‐CFTR at the PM and its ability to obviate WT‐CFTR's lysosomal degradation in the cell periphery.

CK19 stabilizes MRP4 at the PM

To determine whether CK19 can have similar effects on other membrane proteins, the MRP4 and GPCR CXCR2 were examined. Both the cytosolic and membrane pools of MRP4 decreased in CK19 knockdown cells (Fig. 5 ); however, CXCR2 did not reproduce this outcome under the same conditions (Supplemental Fig. S4). MRP4 and CFTR are members of the ABC transporters, which is one of the most abundant membrane protein superfamilies extant from prokaryotes to humans (46). Thus, other than CFTR, these results suggest that CK19 may generally stabilize some ABC transporters at the PM.

Figure 5.

CK19 stabilizes the membrane expression of MRP4. A) MRP4 was transiently transfected into CK19 knockdown or Scr HEK393 cells. After 48 h, MRP4 expression was determined by immunoblot, and its cell surface expression was determined by biotinylation, as in Fig. 2 for CFTR. B) Quantified total and surface expression levels of MRP4 from A (means ± sem, n = 3).

DISCUSSION

In the present study, we identified an intermediate filament protein, CK19, that stabilizes both WT‐CFTR and VX‐809–rescued F508del on the cell surface by interfering with their Rab7‐mediated lysosomal degradation.

Membrane stability is crucial for the function of the CFTR chloride channel (12, 17, 20). As previously described, internalized CFTR, is packaged into vesicles having 2 destinations: recycling to the PM via EEs and REs or transfer to the LEs and subsequently to the lysosome for degradation (16, 46–49). The balance between recycling and degradation of endocytosed CFTR is a key determinant of CFTR's membrane stability (24). The endocytic cargo adapter, Rab7A, regulates internalized CFTR trafficking from EE to LE/lysosome, and its activity leads to a decline of CFTR expression on the cell surface (14, 16, 48). Our data suggest that CK19 negatively impacts the interaction between Rab7A and CFTR, subsequently sparing CFTR from lysosomal degradation and, therefore, leading to the accumulation of PM CFTR and enhanced Cl^– secretion. Our findings also support conclusions drawn from previous studies that implicated Rab7 in lysosomal degradation of membrane‐localized CFTR (48).

The majority of F508del is prematurely degraded by the proteasome (3). However, F508del can function as a chloride channel if rescued from degradation by the UPS and trafficked to the PM. Nevertheless, a limiting factor impacting rF508del is its short PM half‐life, which results from its destabilization at the PM (11, 18, 19, 22 50). Instead of recycling to the cell surface, internalized F508del traffics from REs to the lysosome (18, 19, 51, 52).

We demonstrated that CK19 prolonged rF508del PM half‐life by preventing Rab7A‐mediated lysosomal degradation, and we observed a similar mechanism for WT‐CFTR. Recovery of relatively small amounts of the PM F508del is sufficient to support the chloride secretory function of HBE cells (53, 54). We found that CK19 dramatically increased F508del expression at the surface membrane when HBE cells were treated with the corrector, VX‐809. CK19 enhanced the membrane stability of both rF508del and WT‐CFTR, thus increasing their contribution to macroscopic chloride transport activity. The functional instability of F508del elicits its rapid degradation; conversely, chloride channel activity is a parallel manifestation of CFTR's membrane stability (55).

We also identified that CK19 increased band B expression of both WT‐CFTR and F508del; these changes were amplified by VX‐809 treatment in F508del, suggesting the possibility that CK19 and CFTR perhaps meet in the ER and traffic together to the PM. Unfortunately, our data are not sufficient to explain the underlying molecular mechanism at this time.

Notably, our study showed decreased expression of CK19 in primary HBE cells from homogenous F508del patients with CF. CK19 expression was reduced ~40% in CF HBE cells. This finding was in accordance with our experimental data and supported our conclusion that CK19 modulates CFTR expression in airway epithelia. The synergy between CK19 and F508del corrector VX‐809, which acted by enhancing rF508del‐mediated chloride secretion, indicates that a better understanding of the CK19 pathway may offer a therapeutic approach to CF caused by class II CFTR folding mutations.

To determine whether CK19 regulates other membrane proteins, we assessed its effect on MRP4 and the GPCR CXCR2. The results show that CK19 stabilized MRP4 at the cell surface in a way similar to CFTR but that CK19 did not affect the stability of CXCR2. As members of the ABC transporter protein family, CFTR and MRP4 play important physiologic roles, and their mutation relates to different diseases types: CF and multidrug resistance (28, 56–60). Therefore, our findings may be of significance for a better understanding of the impact of CK19 on the apical density of the ABC transporter proteins.

CK19 belongs to the keratin superfamily of filament proteins, which play roles in the maintenance of cell and tissue integrity related to their roles in organelle and protein targeting (25–27, 31, 61). Keratin filaments form complex cytoskeletal networks that interact with numerous proteins (62, 63). However, relatively few reports have focused on their cellular functions in recent years. CK19 has been recognized as a biomarker for cellular differentiation in cancer cells (25, 26, 28–31). Here, we uncovered a new biologic role for CK19 in interacting with CFTR and positively regulating its membrane stability. Cytoskeletal protein‐mediated CFTR expression is further supported by other reports. Actin and filamin A were shown to improve CFTR expression at the apical membranes of epithelial cells by interacting with the cellular cytoskeletal network (63–65). Numerous trafficking and regulatory proteins are known to mediate CFTR endocytosis, such as endocytic adaptor molecule disabled‐2 and lemur tyrosine kinase 2; however, it is not clear whether they are involved in the CK19 network (21, 65).

As a type II keratin, CK18 has been identified as a CFTR‐binding protein that stabilizes membrane CFTR by accelerating its Rab11‐mediated recycling. In addition, CK18 binds CFTR by competing with AP2, part of the endocytic adapter complex (66–68). Apparently, the regulation of CFTR trafficking by CK18 is directed at a different step or steps from that of CK19, which regulates the membrane stability of CFTR at the LE/lysosome stage. Premchandar et al. (69) reported that CK18 hampers F508del biogenesis and insertion into the PM by binding to the nucleotide‐binding domain 1 of CFTR. We found that CK18 did not participate in the CK19‐mediated regulation of PM CFTR stability/expression in primary HBE cells from patients with CF and the F508del mutation, which suggests that CK18 may interact with CFTR at an earlier stage of endocytic trafficking than CK19. Keratins are a large protein superfamily with multiple functions (70); thus, it is likely that the roles of CK19 and CK18 are different. In fact, CK19 and CK18 are often reported as changing together during disease development (32, 33), raising the possibility that these keratins influence CFTR at different stages of its biogenesis.

Rab7 is a well‐characterized and key regulator of LE trafficking and lysosome biogenesis and maturation (71–73). Many proteins, including many disease‐related membrane proteins (e.g., TLRs, transferrin receptor, and activated epidermal growth factor receptor) are degraded by the Rab7‐mediated lysosomal pathway after their endocytosis (74–78). To accomplish its role in protein trafficking, Rab7 recruits other proteins, such as Rab7‐interacting protein, while Rab7 is under the regulation of its regulatory proteins, such as some ubiquitin E3 ligases and deubiquitinase (79–82). Further elucidation of these pathways can assist in our understanding of CFTR handling in the cell periphery.

In summary, we have revealed a novel mechanism by which CK19 stabilizes both WT‐CFTR and VX‐809‐rescued F508del on the cell/apical surface through diminished Rab7‐mediated lysosomal degradation (Fig. 6 ). Acting through this pathway, CK19 facilitated rF508del's chloride channel function combined with the trafficking corrector, VX‐809. The impact of CK19 on CFTR offers a more detailed understanding of the behavior of WT‐CFTR and rF508del at the PM. The proteins involved in CFTR membrane stability, including CK19, Rab7, and its effectors and regulators, may provide new avenues to CF therapy in the future.

Figure 6.

Model showing pathway of CK19 stabilization of PM WT‐CFTR and rF508del. Cell surface CFTR is internalized into EEs then transferred to REs for return to the PM. The rescued but misfolded mutant rF508del CFTR is recognized by Rab7A for transfer to LEs for lysosomal degradation, a process that can be reduced by CK19 to hinder peripheral rF508del degradation. Ub, ubiquitin.

AUTHOR CONTRIBUTIONS

X. Hou, R A., Frizzel, and F. Sun were responsible for the project and designed the experiments; X. Hou, Q. Wu, and S. Chen produced the CK19 knockdown cells; X. Hou, Q. Wu, C. Rajagopalan, C. Zhang, K. Zaman, and X. Sun performed cell and molecular biological experiments; X. Hou, Q. Wu, and H. Wei contributed biotinylation experiments; F. Sun designed and produced recombinant adenovirus and shRNAs; F. Sun performed pulse chase; X. Hou, Q. Wu, C. Rajagopalan, C. Zhang, and F. Sun analyzed the experiments; X. Chen did the liquid chromatography MS/MS analysis; C. Li performed the MRP4 and CXCR2 experiments; X. Hou, Q. Wu, C. Rajagopalan, M. Bouhamdan, R. A. Frizzell, and F. Sun wrote the paper; and all authors discussed the results and commented on the manuscript.

Supporting information

Supplemental Data

Supplemental Data

Supplemental Data

Supplemental Data

Hou, X. , Wu, Q. , Rajagopalan, C. , Zhang, C. , Bouhamdan, M. , Wei, H. , Chen, X. , Zaman, K. , Li, C. , Sun, X. , Chen, S. , Frizzell, R. A. , Sun, F. CK19 stabilizes CFTR at the cell surface by limiting its endocytic pathway degradation. FASEB J. 33, 12602–12615 (2019). www.fasebj.org