A sequence upstream of canonical PDZ-binding motif within CFTR COOH-terminus enhances NHERF1 interaction

Abstract

The development of cystic fibrosis transmembrane conductance regulator (CFTR) targeted therapy for cystic fibrosis has generated interest in maximizing membrane residence of mutant forms of CFTR by manipulating interactions with scaffold proteins, such as sodium/hydrogen exchange regulatory factor-1 (NHERF1). In this study, we explored whether COOH-terminal sequences in CFTR beyond the PDZ-binding motif influence its interaction with NHERF1. NHERF1 displayed minimal self-association in blot overlays (NHERF1, K_d = 1,382 ± 61.1 nM) at concentrations well above physiological levels, estimated at 240 nM from RNA-sequencing and 260 nM by liquid chromatography tandem mass spectrometry in sweat gland, a key site of CFTR function in vivo. However, NHERF1 oligomerized at considerably lower concentrations (10 nM) in the presence of the last 111 amino acids of CFTR (20 nM) in blot overlays and cross-linking assays and in coimmunoprecipitations using differently tagged versions of NHERF1. Deletion and alanine mutagenesis revealed that a six-amino acid sequence ¹⁴¹⁷EENKVR¹⁴²² and the terminal ¹⁴⁷⁸TRL¹⁴⁸⁰ (PDZ-binding motif) in the COOH-terminus were essential for the enhanced oligomerization of NHERF1. Full-length CFTR stably expressed in Madin-Darby canine kidney epithelial cells fostered NHERF1 oligomerization that was substantially reduced (∼5-fold) on alanine substitution of EEN, KVR, or EENKVR residues or deletion of the TRL motif. Confocal fluorescent microscopy revealed that the EENKVR and TRL sequences contribute to preferential localization of CFTR to the apical membrane. Together, these results indicate that COOH-terminal sequences mediate enhanced NHERF1 interaction and facilitate the localization of CFTR, a property that could be manipulated to stabilize mutant forms of CFTR at the apical surface to maximize the effect of CFTR-targeted therapeutics.

cystic fibrosis (CF) is a lethal autosomal recessive disorder caused by loss of function variants in the gene encoding CF transmembrane conductance regulator (CFTR) (11, 61, 72). It has been demonstrated that overexpression of the scaffold protein sodium/hydrogen exchange regulatory factor-1 (NHERF1) can aid in the apical localization of CFTR bearing the most common CF-causing mutation p.Phe508del (7, 14, 16, 51). Furthermore, the Food and Drug Administration-approved corrector molecule (VX-809) significantly increased interaction between p.Phe508del-CFTR and NHERF1, thereby increasing its stability at the cell surface (3). These studies suggest that manipulation of interactions between mutant forms of CFTR and the apical membrane scaffold proteins could be used to optimize small-molecule therapy for CF.

NHERF1 [a.k.a. EBP50 (ERM-binding phosphoprotein of 50 kDa)] has two PDZ domains (52) that form a macromolecular scaffold on binding to PDZ ligands linked to the actin cytoskeleton (6, 21, 22). This macromolecular complex is responsible for the stabilization, sorting, recycling, and localization of cell surface proteins, such as CFTR (68), platelet derived growth factor receptor (44), type IIa sodium-dependent phosphate cotransporter (27), PDZK1 (34), and parathyroid hormone I receptor (71). The PDZ domains of NHERF1 recognize and bind through a canonical conserved PDZ peptide-binding pocket at the COOH-terminal motif of PDZ ligands (26). Apart from canonical ligand-binding ability, NHERF1 PDZ domains increase their scaffolding capacity through PDZ oligomerization with the same or different PDZ-containing proteins. It has been previously reported that oligomerization of NHERF1 can be enhanced by the COOH-terminal sequences of two PDZ-binding motif containing proteins, the β₂-adrenergic receptor, and the platelet-derived growth factor receptor (18, 44). However, it is unknown whether the COOH-terminus of CFTR can stimulate NHERF1 oligomerization and, in doing so, modulate its own stability at the cell surface.

It has been proposed that NHERF1 oligomerization and its affinity for PDZ ligands is mediated by phosphorylation (17, 18, 24, 36, 60). In contrast, more recent work revealed that wild-type (WT) and mutant NHERF1 is an elongated monomer in solution with no tendency to oligomerize, regardless of its phosphorylation state (19, 39, 40, 64). In light of these differences, we estimated the intracellular concentration of NHERF1 and CFTR in dissected human eccrine sweat glands so that in vitro oligomerization studies could be performed at or below physiological concentrations. After demonstrating that the CFTR COOH-terminus fosters NHERF1 oligomerization, we performed comprehensive mutagenesis to identify specific residues that confer this property. To confirm that these sequences were relevant in vivo, we show that the newly identified motif, in combination with the PDZ-binding motif in full-length CFTR, affects NHERF1 oligomerization in cell-based studies. Finally, using these two sequence motifs, we show that NHERF1 oligomerization contributes to localization of CFTR, thereby supporting the concept that these interactions provide a viable approach to stabilizing mutant forms of CFTR at the cell surface.

MATERIALS AND METHODS

Immunohistochemistry.

Frozen sections of lung and skin obtained from the Division of Surgical Pathology, Johns Hopkins University, Baltimore, MD, were embedded in optimal cutting temperature compound and held at −70°C before sectioning. Six-micrometer cryosections were mounted onto uncoated microscope slides. Sections were fixed for 10 min in precooled acetone, followed by 5-min peroxidase block at room temperature, to quench the endogenous peroxidase activity. Sections were further incubated in serum-free protein block (Dako, no. X0909) for 20 min at room temperature and incubated overnight at 4°C with anti-mouse NHERF1 antibody (LSBio). Mouse universal negative control antibody (Dako, no. N1698) was used to ascertain nonspecific staining. After washing, staining was performed using EnVision+System-HRP (AEC) kit (Dako, no. K4004). Sections were covered with peroxidase-labeled polymer for 30 min. For visualization of the reaction, sections were developed in AEC+substrate-chromogen for 5–20 min. After washing, the sections were counterstained with hematoxylin (Dako, no. S3309) for 30 s, cleared, and mounted on Faramount aqueous mounting medium (Dako, no. S3025). Slides were analyzed under a Olympus BX51 microscope.

Mass spectrometry.

De-identified skin was obtained according to institutional protocol from otherwise healthy, non-CF adult patients undergoing unrelated reconstructive surgery at the Department of Plastic and Reconstructive Surgery, Johns Hopkins University School of Medicine. Eccrine sweat glands were dissected out from the skin dermis under the stereoscopic dissection microscope, as previously described (59). Sweat gland proteins were digested with trypsin at the enzyme-to-protein ratio of 1:50 at 37°C overnight. The peptides were fractionated into 24 fractions by basic pH reverse-phase liquid chromatography and analyzed on LTQ-Orbitrap Elite mass spectrometers (Thermo Scientific, Bremen, Germany), coupled with Easy-nLC II nanoflow liquid chromatography systems (Thermo Scientific). Tandem mass spectrometry (MS/MS) data obtained from the mass spectrometry (MS) were searched against human RefSeq database (version 65) using Sequest search algorithm through Proteome Discoverer 1.4 platform (Thermo Scientific).

RNA extraction.

The sweat glands (∼50) were collected in 0.5-ml RNAlater (Qiagen, no. 76104), as they were being dissected out from the skin. The collection tube was centrifuged to remove RNAlater solution. TRIzol (500 μl) was added to the sample and mixed. The sample was processed in the bead-beater (FastPrep-24, MP) at a setting of 4 m/s for 20 s and then placed on ice for 5 min. This was repeated three times. Chloroform (200 μl) was added to the sample and incubated for 5 min at room temperature. The sample was centrifuged at 8,000 g for 5 min at 4°C. The aqueous phase was separated from the organic phase containing the DNA and protein. Ethanol (500 μl; 70%) was added to the sample for precipitation and immediately added to RNeasy columns (RNeasy Kit, Qiagen, no. 74104). The manufacturer's cleanup procedure was followed, except that the RNA sample was treated with the Turbo DNA-Free kit to remove residual DNA contamination (ThermoFisher, no. AM1907). The RNA was stored at −80°C. The quality of RNA was measured by RNA integrity number using Agilent 2100 Bioanlayzer. An RNA integrity number value of 7.8 was obtained.

Library preparation and RNA-sequencing.

Using 1.0 μg of total RNA isolated from sweat glands, library preparation and RNA-sequencing (RNA-Seq) were performed at the Johns Hopkins Medical Institutions Deep Sequencing and Microarray Core Facility. RNA-Seq library was constructed using the TruSEQ RNA Sample Prep Kit version 2 (Illumina, San Diego, CA), and 200 million paired end reads were obtained from 75 cycles run on Illumina Next Seq 500 platform. Raw reads were aligned to the reference genome (hg19) using the Bowtie2 algorithm (35), and splice junctions were identified via Tophat2 (version 2.0.13) (70) from the Tuxedo software suite. CuffQuant and Cuffdiff (Cufflinks version 2.2.1) (69) were then used to assemble transcripts, estimate their abundances, and test for differential expression among samples.

Estimation of NHERF1 and CFTR protein concentration per cell of eccrine sweat gland.

We used histone as a “proteomic ruler” for measuring mass of the protein in the measured sample. It has been demonstrated that ratio of histone MS signal to total MS signal allows the estimation of the total cellular protein mass without any additional measurements (74). To determine the total cellular protein mass in the measured sample, we have included a factor of 6.5 pg as the DNA mass of diploid human cells, since it correlates with cellular protein mass within a factor of 1.24 ± 0.29 (74).

$$\text{Protein}\hspace{0.17em}\text{mass}=\frac{\text{Protein}\hspace{0.17em}\text{MS}\hspace{0.17em}\text{signal}\hspace{0.17em}\left(\text{PSM}\right)\times \text{DNA}\hspace{0.17em}\text{mass}\hspace{0.17em}\left(6.5\hspace{0.17em}\text{pg}\right)}{\text{Total}\hspace{0.17em}\text{Histone}\hspace{0.17em}\text{MS}\hspace{0.17em}\text{signal}\hspace{0.17em}\left(\text{PSM}\right)}$$

Protein spectral match (PSM) reflected total number of identified peptide spectra matched for the protein. To calculate concentration, the cell volume was set to 2,000 μm³, as previously described (50, 75). We used fragments per kilobase per million mapped (FPKM) data from RNA-Seq to predict concentrations of CFTR that liquid chromatography (LC)-MS/MS was unable to detect. The correlation between the transcriptomics and proteomics data was used to predict the PSM from the FPKM value of their transcripts.

Plasmids.

The plasmids pGEX-6P-C111 (1370–1480), pGEX-6P-C111-DelTRL (1370–1477), pGEX-6P-C26 (1455–1480), and pGEX-6P-Del2 (1370–1394/1404–1424/1477–1480) were generated by amplifying CFTR COOH-termini cDNA fragments from previously described pRK5-SK (48) and subcloning into pGEX-6P-1 (Amersham Biosciences). Constructs C111 and C111-DelTRL had alanine substitutions at 1402 and 1403 to eliminate aggregation (49). Similarly, pGEX-6P-C55 (1426–1480) and pGEX-6P-C77 (1404–1480) were constructed by subcloning CFTR COOH-termini cDNA fragment into pGEX-6P-2 (Amersham Biosciences). NHERF1, PDZ1, and PDZ2 cDNAs cloned into the pET-30a (+) bacterial expression vector were a generous gift from E. Weinman at the University of Maryland. pGEX-6P-HA-NHERF1 was generated by subcloning hemagglutinin (HA)-NHERF1 from pCGN-HA-tagged NHERF1 (a generous gift from R. A. Hall at Emory University School of Medicine) into the SmaI site of pGEX-6P-2. pCMV-SC-NH₂-terminal myc-tagged NHERF1 was created using StrateClone mammalian expression vector systems instruction manual (Agilent Technologies).

Site-directed mutagenesis.

Quickchange XL site-directed mutagenesis kit (Stratagene, La Jolla, CA; no. 200516) was utilized to introduce alanine substitution mutations on pGEX6p-C111 at ¹⁴⁰⁴IEAMLECQQF¹⁴¹³, ¹⁴¹⁴LVI¹⁴¹⁶, ¹⁴¹⁷EEN¹⁴¹⁹, ¹⁴²⁰KVR¹⁴²², and ¹⁴²³QYD¹⁴²⁵ residues. ¹⁴¹⁷EEN¹⁴¹⁹ and ¹⁴²⁰KVR¹⁴²² were mutated in tandem to alanine to generate EENKVR-AAAAAA mutant (referred to as EENKVR-alanine). Mutants were verified by sequencing.

Purification of recombinant proteins.

NH₂-terminal glutathione S-transferase (GST)-tagged CFTR COOH-terminal constructs, NH₂-terminal HA-tagged NHERF1, and NH₂-terminal His-tagged NHERF1, PDZ1, and PDZ2, were expressed in BL21(DE3) competent cells (Stratagene, La Jolla, CA, no. 200131). GST-tagged constructs were affinity purified employing glutathione-sepharose 4B (Amersham). PreScission protease (GE HealthCare) cleavage of the GST tag on the fusion proteins expressed from pGEX6p-1 and pGEX6p-2 vectors was followed as per the manufacturer's protocol. HA-NHERF1 fusion protein was affinity purified with anti-HA affinity gel (Roche Diagnostics) and eluted with HA peptide (Roche Diagnostics). His-tagged NHERF1, PDZ1, and PDZ2 were purified with Ni-NTA agarose protocol (Qiagen).

Coimmunoprecipitation and cross-linking assays utilizing purified recombinant proteins.

Coimmunoprecipitation (Co-IP) assay was performed with HA-NHERF1 and His-NHERF1 (10 nM each) in solution incubated with 20-μl anti-HA affinity matrix (Roche Diagnostics, Germany; no. 11815016001) overnight at 4°C in the presence of C111 or GST (20 nM each). Matrix was pelleted and washed thrice with 100-μl wash buffer (50 mM Tris, 120 mM NaCl, 0.1% Nonidet P-40). The bound proteins were eluted with 50 μl of 2× Laemmli sample buffer and 50 mM DTT (BioRad, Hercules, CA; no. 161-0737) at 95°C for 5 min. Proteins in the elute were detected by Western blotting with mouse monoclonal anti-HA antibody (1:20,000, Sigma-Aldrich; no. H9658), mouse monoclonal polyhistidine antibody (1:6,000; Sigma-Aldrich; no. H1029), mouse monoclonal anti-CFTR antibody (1:1,000; Santa Cruz, no. sc-20074) and mouse monoclonal anti-GST antibody (1:10,000; no. G1160). Purified C111 (20 nM) and HA-NHERF1 (10 nM) were incubated with bis[sulfosuccinimidyl]suberate (BS³; Pierce no. 21580) at a concentration of 0.5 mM in 25 mM sodium phosphate buffer, pH 8.5, 150 mM NaCl, 1 mM DTT, 0.1 mM ATP, 0.02% NaN₃. Samples were incubated at room temperature for 5, 15, 30, 60, or 120 min before addition of ethanolamine (Sigma, no. E9508) to a final concentration of 100 mM. The 0-min sample had ethanolamine added before cross-linker. Following ethanolamine quenching, samples were incubated on ice for an additional 5 min before the addition of reducing SDS-PAGE buffer. Samples were resolved by SDS-PAGE on precast 4–20% Tris·HCl gels, and the protein bands were visualized by immunoblot analysis, as described above.

Blot overlay assays.

A standard protocol described elsewhere was followed (23). About 500 ng of purified His-tagged NHERF1 protein (full-length, PDZ1, and PDZ2) was run on 4–20% Mini-PROTEAN TGX precast gel (Bio-Rad, no. 456–10934) and electrophoretically transferred onto polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA; no. 1620174) in transfer buffer containing 25 mM Tris, pH 8.3, 192 mM glycine, 0.01% SDS, and 15% methanol using Hoefer SemiPhor semidry transfer unit (Amersham Biosciences, San Francisco, CA; no. 80621034) to which 180-mA current was applied for 1 h. The membrane was blocked overnight with 5% nonfat dried milk in PBS and washed with PBS-Tween 20 (PBST). The membrane was incubated (2 h) at room temperature in blocking buffer containing either CFTR COOH-termini alone (overlay) or CFTR COOH-termini and HA-NHERF1 (simultaneous overlay) or CFTR COOH-termini, followed by HA-NHERF1 with a PBST wash in between (1 h) and vice versa (sequential overlay). CFTR COOH-termini and HA-NHERF1 concentrations used were 20 and 10 nM, respectively. The blots were then immunoblotted using mouse monoclonal anti-HA antibody (1:20,000; Sigma-Aldrich; no. H9658). The blots were stripped in buffer containing 100 mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris·HCl, pH 6.7, and reprobed with anti-His antibody (1:30,000; Qiagen, penta-his; no. 34660).

Generation of Madin-Darby canine kidney type II-Flp recombination target stable cell lines expressing WT/mutant CFTR.

The pcDNA5/Flp recombination target (FRT)/green fluorescent protein (GFP)-WT CFTR was made by removing full-length cDNA from the existing peGFP-CFTR plasmid (46) by NheI (5′ end) and EcoRV (3′ end) digestion and ligation into the same sites in the multiple cloning sequence of the pcDNA/FRT plasmid (Invitrogen, no. K6010-02). Mutants were created using pcDNA5/FRT/GFP-WT CFTR as template. Alanine mutagenesis of ¹⁴¹⁷EEN¹⁴¹⁹, ¹⁴²⁰KVR¹⁴²², and ¹⁴¹⁷EENKVR¹⁴²² residues and deletion of terminal TRL residues were carried out as described above. The Madin-Darby canine kidney type II-FRT (MDCK II-FRT) cell line, which contains the FRT but expresses no endogenous CFTR, was the kind gift of Gregory Germino (Johns Hopkins School of Medicine, Baltimore, MD). MDCK parental cells were cotransfected using Lipofectamine 2000 (Invitrogen, no. 11668-019) with the appropriate pcDNA5/FRT/GFP-CFTR construct and pOG44 recombinase (Flp-in System kit; Invitrogen, no. K6010-02) at ratios of 1:12 or 1:15. Verification of pcDNA5/FRT/GFP-CFTR integration into the FRT site in MDCK cells was done as described earlier (33).

Co-IP of differently tagged NHERF1 expressed in MDCK II stable cells expressing WT/mutant CFTR.

MDCK II-FRT parental cells or stable cells expressing WT or mutant CFTR were cultured in Dulbecco's modified Eagle's medium (Gibco, Carlsbad, CA; no. 11965) with 10% fetal bovine serum (Gibco, no. 16140-071) in a humidified incubator at 37°C in the presence of 5% CO₂. MDCK parental cells with no CFTR integrant were selected on Zeocin (100 μg/ml; Invitrogen, no. 45-0430). MDCK stable cells stably expressing either CFTR WT or mutant CFTR were selected on Hygromycin (100 μg/ml; Invitrogen, no. 10687-010). To assess myc-NHERF1 Co-IP, cotransfections were performed in MDCK-FRT parental and MDCK-FRT stably expressing different forms of CFTR. Briefly, MDCK-FRT cells (100-mm dishes, ∼90% confluence) were cotransfected with 24 μg each of pCGN-HA-NHERF1 and pCMV-SC-N-myc-NHERF1 using Lipofectamine 2000 (Invitrogen, Carlsbad, CA; no. 11668-027). In the control dish, cells were treated with reagent only. Twenty-four hours later, the cells were washed twice with ice-cold PBS buffer and then scraped into 1 ml of ice-cold lysis buffer (120 mM NaCl, 50 mM Tris, pH 8.0, 0.5% Nonidet P-40) containing protease inhibitors cocktail (Sigma-Aldrich; no. P8340) and phosphatase inhibitors (100 mM sodium fluoride, 0.2 mM sodium orthovanadate). The lysate was rocked for 30 min at 4°C, and the insoluble material was removed by centrifugation at 13,000 g for 15 min at 4°C. The protein (300 μg) was immunoprecipitated with 100 μl of anti-HA affinity matrix (Roche Diagnostics, Germany; no. 11815016001) overnight at 4°C, the matrix was washed four times with 1 ml of cold lysis buffer and the bound proteins eluted with 100 μl of 2× Laemmli sample buffer/50 mM DTT (Bio-Rad, Hercules, CA; no. 161-0737) at 95°C for 5 min. Co-IP was detected by Western blotting with mouse monoclonal anti-c-myc antibody (Sigma-Aldrich; no. M4439). Full-length CFTR was detected by mouse monoclonal CFTR antibody 596 (CF Foundation). Anti-GAPDH (Sigma-Aldrich; no. G9545) was used as control.

Immunostaining and confocal microscopy.

MDCK cells type II were grown on noncoated glass coverslips, and cotransfected at 5:1 ratio for plasmids encoding the myc-tagged COOH-terminal constructs (C111, EENKVR-alanine mutant, and DelTRL) and GFP-tagged full-length CFTR using Lipofectamine 2000 (Invitrogen, Carlsbad, CA; no. 11668-027). Forty-eight hours posttransfection, cells were fixed with 4% paraformaldehyde for 20 min, washed extensively with PBS, and permeabilized with 0.25% Triton X-100 for immunostaining. Nonspecific binding sites were blocked with 2.5% normal goat serum (Invitrogen; no. 16201). The cells were incubated for 1.5 h with mouse anti-CFTR (1:1,500, UNC CFTR antibody distribution program by CFFT, no. 596) and rabbit anti-c-myc (1:500, Sigma, no. C3956) primary antibodies, washed with PBS, and thereafter incubated for 30 min with Alexa Fluor 488 anti-mouse (Invitrogen, no. A10684) and Alexa Fluor 647 anti-rabbit (Invitrogen, no. A21246) secondary antibodies, respectively. Coverslips were mounted with Prolong Gold antifade reagent with 4,6-diamidino-2-phenylindole (Invitrogen; P36931). The specimens were visualized on a Zeiss 510 confocal microscope (Zeiss, Thornwood, NY; microscope located at the Johns Hopkins School of Medicine Microscope Facility) using ×63 magnification. Images were generated using 16-fold line averaging. The xz cross sections were produced using a 0.2-μm motor step. At least 10 transfected cells were tested for the localization of CFTR or myc-tagged C111 constructs. Images were prepared for publication with LSM Carl Zeiss Software and Adobe Photoshop.

RESULTS

NHERF1 and CFTR are expressed at nanomolar concentrations in human cells.

To determine an appropriate concentration of NHERF1 to be employed in the evaluation of NHERF1 oligomerization in vitro, we estimated the intracellular concentration of NHERF1 in dissected sweat glands. We chose this tissue because CFTR is expressed in sweat gland and dysfunction of CFTR in sweat gland, resulting in elevated sweat chloride, is a diagnostic feature of CF. Immunohistochemistry of a formalin-fixed skin section stained with NHERF1 antibody revealed apical expression of NHERF1 in the reabsorptive ducts of sweat glands, coincident with the localization of CFTR (13) (Fig. 1A). Using the same antibody, NHERF1 was localized to the apical surface of bronchial epithelial cells, as previously reported (43) (Fig. 1B). Next, sweat glands were dissected from freshly collected skin and subjected to LC-MS/MS and RNA-Seq. NHERF1 in the sweat gland was detected from its LC-MS/MS spectrum (Fig. 1C). NHERF1 concentration per cell in the sweat gland was estimated to be 259.1 nM and 242.2 nM by LC-MS/MS and RNA-Seq, respectively (Table 1). While LC-MS/MS and RNA-Seq estimates of NHERF1 concentration in the sweat gland were similar, LC-MS/MS did not detect CFTR in the sweat gland, even though its predicted concentration was 5.2 nM by RNA-Seq (Table 1). To investigate this discrepancy, we queried the human proteome map to determine CFTR levels in tissues expected to express CFTR (32). In accordance with RNA-Seq data (Fig. 1D), pancreas displayed the highest CFTR protein expression across 30 different tissues (http://www.humanproteomemap.org/batch.php). Notably, the concentration of CFTR in the pancreas estimated from MS (∼20 nM) was comparable to the prediction from RNA-Seq (16.5 nM). RNA-Seq predicted CFTR concentration in lung to be 5.2 nM, however, two prior studies did not detect CFTR in the proteome of whole lung (32, 73). The discrepancy between MS and RNA-Seq estimates of CFTR concentration in the sweat gland and lung may be due to its low expression level, varying from a few to a few hundred of protein copies per cell (9, 10, 15, 57). MS may not detect very low abundance proteins if significant ion suppression occurs when peptides originating from the protein of interest coelute, with peptides originating from high-abundance proteins (58). Although RNA, immunohistochemical, and clinical evidence indicates that CFTR is present in the sweat gland and lung, its concentration in these tissues appears to be below the detection threshold of current MS methods.

Fig. 1.

Expression of NHERF1 and CFTR in human eccrine sweat gland. A: immunohistochemistry of a human skin showing apical membrane immunoreactivity of NHERF1 in the reabsorptive ducts of the eccrine sweat glands (original magnification ×20). B: immunohistochemistry of a human lung tissue showing apical membrane immunoreactivity of NHERF1 in the bronchiolar epithelial cells (original magnification ×20). Arrow indicates NHERF1 staining. Inset shows magnification ×40. C: representative MS/MS spectrum for NHERF1 protein in eccrine sweat gland dissected out from human skin. D: heat map of gene expression of 37 genes (rows) in 6 tissues (columns) based on RNA-Seq data. *CFTR and NHERF1 (SLC9A3R1). RNA-Seq data were obtained from The Human Protein Atlas, from the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database for the pancreas (accession numbers: ERR315466, ERR315429, ERR315436), skin (ERR315372, ERR315376, ERR315339), lung (ERR315424, ERR315444, ERR315487), colon (ERR315348, ERR315403, ERR315462), and kidney (ERR315443, ERR315494, ERR315468). Eccrine sweat gland RNA-Seq data were obtained in house. Color scale represents log10 transformed fragments per kilobase of transcript per million mapped reads (FPKM) + 1. Darker colors indicate higher expression.

Table 1.

Concentrations of NHERF1 and CFTR per human cell calculated from the RNA-Seq and proteomic data

	NHERF1 Concentration per Cell		CFTR Concentration per Cell
Tissue Type	RNA-Seq	LC-MS/MS	RNA-Seq	LC-MS/MS
Sweat gland	242.2	259.1	5.2	Not detected
Pancreas	49.9	75.3	16.5	20.3
Lung	35.45	60.82	3.81	Not detected

Values are in nM. RNA-Seq and proteomic data on eccrine sweat gland were obtained in house. Human proteome map was accessed to obtain proteomic data on pancreas and lung (http://www.humanproteomemap.org/batch.php). RNA-Seq data on pancreas and lung were obtained from The Human Protein Atlas, from the National Center for Biotechnology Information (NCBI) Sequence Read Archive database.

The CFTR COOH-terminus promotes oligomerization of NHERF1 at physiological concentration.

To determine the protein concentration at which NHERF1 self-associated in an overlay assay, full-length NHERF1 or sequences encompassing the PDZ1 or PDZ2 domains of NHERF1 with the His epitope fused to the NH₂-terminus (500 ng of each) were immobilized on membrane and then incubated with HA-NHERF1 (HA epitope fused to the NH₂-terminus) at concentrations ranging from 1 nM to 3,000 nM. After washing, HA-NHERF1 interacting with the immobilized His-NHERF1 was detected with an anti-HA antibody. Low-affinity binding of HA-NHERF1 with His-NHERF1 (K_d 1,256-1,508 nM; 95% confidence interval) and His-PDZ1 (K_d 1,182-1,467 nM; 95% confidence interval) were observed (Fig. 2A). HA-NHERF1 displayed minimal binding to PDZ2 that is consistent with reports that full-length NHERF1 can bind to its own PDZ1 domain with low affinity, but does not bind to the PDZ2 domain (5, 18, 40, 67). Based on these results, we performed all subsequent in vitro experiments using 10 nM of HA-NHERF1, a concentration at which oligomerization with immobilized His-NHERF1 was not evident and well below that observed in cells of the human sweat gland.

Fig. 2.

CFTR COOH-terminus enhances formation of NHERF1 oligomers at nanomolar concentrations. A: binding of HA-tagged NHERF1 with full-length His-tagged NHERF1, PDZ1 (a.a. 11–101), and PDZ2 (a.a. 152–358) detected with anti-HA antibody (1:20,000). His-tagged NHERF1, PDZ1, and PDZ2 (500 ng each) were run on SDS-PAGE and transferred to the polyvinylidene difluoride membrane, and blot was overlaid with increasing concentrations of HA-tagged NHERF1 (1 nM, 10 nM, 100 nM, 500 nM, 1 μM, 1.5 μM, 2.0 μM, 2.5 μM, and 3 μM). Values are means ± SE (n = 3). B: effect of C111 on the binding of HA-NHERF1 with His-tagged NHERF1, PDZ1, and PDZ2. C111 refers to last 111 amino acids (a.a. 1370–1480) of human CFTR. In simultaneous overlay, HA-NHERF1 (10 nM) + C111 (20 nM) were overlaid at the same time onto the immobilized His-tagged NHERF1, PDZ1, and PDZ2. In sequential overlay, incubation with C111 (20 nM) was followed by HA-NHERF1 (10 nM) with a PBST wash in between (C111>>>>HA-NHERF1) or vice versa (HA-NHERF1>>>>C111). Incubation time was 2 h for both simultaneous and sequential overlays. Values are means ± SE (n = 5). ***P < 0.0001, one-way ANOVA (comparison with NHERF1 oligomerization by itself). C: coimmunoprecipitation of His-NHERF1 by HA-NHERF1 (10 nM each) in the absence or presence of C111 (20 nM). Graph illustrates fold increase in His-NHERF1 binding to HA-NHERF1 by C111. Values are means ± SE (n = 3). ***P < 0.001, one-way ANOVA. D: analysis of HA-NHERF1 oligomers in the presence of C111 or C111 mutants (C26, C111ΔTRL, and alanine substitution of ¹⁴¹⁷EENKVR¹⁴²²) using a cross-linker N-hydroxysuccinimide ester (BS³). The reaction was quenched with ethanolamine (100 nM) at the time indicated. The 0-min sample had ethanolamine added before cross-linker. Oligomers of NHERF1 were detected by immunoblotting with anti-HA antibody.

To assess whether the CFTR COOH-terminus could enhance NHERF1 oligomerization, overlay experiments were performed in the presence of a 111-amino acid (a.a.) peptide from the CFTR COOH-terminus (a.a. 1370–1480; termed C111). C111 has two alanine substitutions (H1402A and R1403A) that eliminates its propensity to aggregate (48, 49). Addition of C111 (20 nM) greatly enhanced HA-NHERF1 binding with full-length His-NHERF1 and His-PDZ1 compared with the level of interaction using HA-NHERF1 alone (Fig. 2B). Minimal complex formation was observed between HA-NHERF1 and His-PDZ2 even in the presence of C111. To confirm that C111 had to be present to enhance NHERF1 interaction, protein overlay experiments were performed in sequential fashion. Oligomerization was observed when the membrane was overlaid first with C111, washed, and then incubated with HA-NHERF1 (C111>>>>HA-NHERF1). However, incubation with HA-NHERF1 alone, followed by a wash and subsequent incubation with C111, did not result in NHERF1 oligomerization (HA-NHERF1>>>>C111).

Next, we performed Co-IP and cross-linking experiments to demonstrate that C111 can enhance NHERF1 oligomerization in solution. We tested whether HA- and His-tagged NHERF1 proteins (10 nM each) coimmunoprecipitate in the presence of C111 (20 nM). The complexes were immunoprecipitated with anti-HA affinity beads, separated by electrophoresis and detected on Western blot using anti-His antibody (Fig. 2C). While His-NHERF1 displayed minimal Co-IP with HA-NHERF1, addition of C111 increased His-NHERF1 Co-IP with HA-NHERF1 by 5 ± 0.5-fold (P < 0.001, Fig. 2C). GST used as a negative control showed absence of nonspecific immunoprecipitation. We used N-hydroxysuccinimide ester (BS³) as a chemical cross-linker to demonstrate HA-NHERF1 (10 nM) oligomers formed in the presence of C111 (20 nM). Oligomerization was detected by immunoblotting with anti-HA antibody. After 5 min of exposure to cross-linker, protein complexes of discrete molecular mass corresponding to dimers, trimers, and tetramers of the NHERF1 were observed (Fig. 2D). Increased duration of exposure to cross-linker did not alter the pattern or intensity of oligomerization.

A six-amino acid (¹⁴¹⁷EENKVR¹⁴²²) sequence in the CFTR COOH-terminus fosters the formation of NHERF1 oligomers.

To determine which regions in the CFTR COOH-terminus are involved in enhancing NHERF1 oligomerization, we performed overlay experiments with a series of CFTR COOH-terminal peptides (Fig. 3A). The CFTR COOH-terminus lacking the last three amino acids (C111-DelTRL) did not foster HA-NHERF1 oligomerization, demonstrating that enhancement of NHERF1 oligomerization in the overlay experiments requires the presence of an intact PDZ-binding motif. The last 77 amino acids of CFTR (C77) enhanced NHERF1 oligomerization, while peptides containing the last 55 amino acids (C55) or the last 26 amino acids (C26) did not. These results indicated that the regions of CFTR responsible for NHERF1 oligomerization consisted of the PDZ-binding motif (a.a. 1477–1480) and COOH-terminal sequences between a.a. 1404 and 1425 (shown in shaded bars). Next, we wanted to test whether this is a result of specific interactions in the CFTR COOH-terminus, rather than a peptide length requirement. Therefore, we created a COOH-terminal construct, C111Del2, composed of residues 1370–1394, 1404–1424, and 1477–1480. The 1370–1394 sequence was not expected to affect NHERF1 oligomerization, but was included to create a construct of comparable length to C55 (48 a.a. vs 55 a.a.). As predicted, the Del2 peptide enhanced NHERF1 oligomerization to a level comparable to that of C77.

Fig. 3.

Identification of a six-amino acid sequence in the CFTR COOH-terminus involved in enhancing NHERF1 oligomers. A: diagram of C111 (a.a. 1370–1480) deletions showing regions 1404–1425 and 1477–1480 (shaded) as essential for enhancing NHERF1 oligomers. Blots are the overlays of 10 nM HA-NHERF1 and 20 nM of each CFTR COOH-terminal deletion fragments C111-DelTRL (deletion a.a. 1478–1480), C77 (a.a. 1404–1480), C55 (a.a. 1426–1480), C26 (a.a. 1455–1480), and Del2 (a.a. 1370–1394, 1404-24, and 1477–1480) onto the immobilized His tagged NHERF1, PDZ1, and PDZ2 (500 ng each). Binding of HA-NHERF1 with His-NHERF1 and PDZ fragments was detected using anti-HA antibody (1:20,000). The graph illustrates the fraction of NHERF1 binding to the corresponding His tagged constructs in the presence of C111 deletion mutants. Values are means ± S.E. (n = 5). ***P < 0.0001, one-way ANOVA, as in Fig. 2B. B: diagram of C111 constructs with various alanine substitutions (white regions with A) used to identify a.a. 1417–1422 as essential for NHERF1 oligomers. Blots are the overlay assays as described above using the CFTR COOH-terminus (C111) with the following residues mutated to alanine: IE-QF (¹⁴⁰⁴IEAMLECQQF¹⁴¹³), ¹⁴¹⁴LVI¹⁴¹⁶, ¹⁴¹⁷EEN¹⁴¹⁹, ¹⁴²⁰KVR¹⁴²², and ¹⁴²³QYD¹⁴²⁵. The graph illustrates quantitation of the blots as described above. Values are means ± SE (n = 5). ***P < 0.0001. C: diagram of C111 constructs with single alanine amino acid substitutions at the corresponding position, as indicated by the labels on the figure. Graph represents the quantitation of the blot overlay assays as described above. Values are means ± SE (n = 3). ***P < 0.0001.

To identify the specific amino acids that promote NHERF1 oligomerization, we performed overlays with a series of alanine substitutions within residues 1404–1425 in the COOH-terminal peptide C111 (Fig. 3B). Substitution of the ¹⁴¹⁷EEN¹⁴¹⁹ or ¹⁴²⁰KVR¹⁴²² residues resulted in a C111 peptide that was unable to promote oligomerization of NHERF1. Mutagenesis of other residues in this region (¹⁴⁰⁴IEAMLECQQF¹⁴¹³, ¹⁴¹⁴LVI¹⁴¹⁶, and ¹⁴²³QYD¹⁴²⁵) did not affect NHERF1 oligomerization. Single amino acid mutagenesis revealed that mutation of either glutamate (¹⁴¹⁷E or ¹⁴¹⁸E) was sufficient to eliminate enhancement of NHERF1 complex formation, while substitution of residues ¹⁴¹⁹N, ¹⁴²⁰K, or ¹⁴²²R individually did not affect the ability of these sequences to enhance NHERF1 oligomerization (Fig. 3C). The latter observation suggests that the two glutamate residues at 1417 and 1418 are critical contributors to the enhancement of NHERF1 oligomerization, although our results could not exclude a role for the adjacent ¹⁴²⁰KVR¹⁴²² residues. Thus a parsimonious conclusion is to implicate the entire six-amino acid region ¹⁴¹⁷EENKVR¹⁴²² in the promotion of NHERF1 oligomerization.

¹⁴¹⁷EENKVR¹⁴²² sequence acts in concert with PDZ-binding motif to form NHERF1 oligomers.

To test whether the novel motif (¹⁴¹⁷EENKVR¹⁴²²) required the presence of the terminal PDZ-binding motif (¹⁴⁷⁸TRL¹⁴⁸⁰) to foster NHERF1 oligomerization, we performed a series of competition experiments. We hypothesized that COOH-terminal sequences would promote the formation of HA-NHERF oligomers in solution, thereby depleting HA-NHERF1 available to interact with immobilized His-NHERF1 and C111 complexes. Indeed, we observed a concentration-dependent decrease in HA-NHERF1 interaction when C111 was incubated with HA-NHERF1 before being overlaid on immobilized His-NHERF1-C111 complexes (Fig. 4A). NHERF1 interaction was reduced by >95% at concentrations of C111 exceeding 100 nM.

Fig. 4.

Formation of NHERF1 oligomers by ¹⁴¹⁷EENKVR¹⁴²² sequence in CFTR COOH-terminus is not independent of terminal PDZ-binding motif. A: effect of HA-NHERF1 binding to His-tagged NHERF1, PDZ1, and PDZ2 by preincubation with increasing concentration of C111. His-tagged NHERF1, PDZ1, and PDZ2 on the membrane were overlaid with C111 (20 nM) for 2 h. After PBST wash, the membrane was overlaid with HA-NHERF1 (10 nM) preincubated with C111 at 1 nM, 10 nM, 100 nM, or 200 nM for 2 h. Values are means ± SE (n = 3). B: effect of HA-NHERF1 binding to the His-tagged NHERF1, PDZ1, and PDZ2 by preincubation with C111 mutants. His-tagged NHERF1, PDZ1, and PDZ2 on the membrane were overlaid with C111 (20 nM) for 2 h. After PBST wash, the membrane was overlaid with HA-NHERF1 (10 nM) preincubated with C111, C111-Del2, C111-EENKVR-alanine mutant, and C111-DelTRL (100 nM each) for 2 h. Values are means ± SE (n = 3). ***P < 0.0001, one-way ANOVA.

Next, we asked whether C111 with alanine substitution of EENKVR or TRL deletion were able to deplete HA-NHERF1 and prevent interaction with His-NHERF1-C111 complexes. However, neither C111-EENKVR-alanine nor C111-DelTRL (100 nM each for 2 h) prevented the formation of HA-NHERF1 complexes (Fig. 4B). Conversely, the minimal construct C111-Del2, containing both TRL and the EENKVR region, inhibited complex formation when preincubated with HA-NHERF1 at 100 nM for 2 h (Fig. 4B).

¹⁴¹⁷EENKVR¹⁴²² and PDZ-binding motif in full-length CFTR enhance NHERF1 interaction in polarized epithelial cell lines.

To establish the role of EENKVR and TRL in full-length CFTR, we constructed MDCK II-FRT cell lines stably expressing either wild-type CFTR (WT-CFTR) or one of four CFTR mutants. Three cell lines expressed CFTR with all six residues (CFTR-1417EENKVR1422), the first three residues (CFTR-1417EEN1419), or next three residues (CFTR-1420KVR1422) of EENKVR substituted with alanines; and a fourth expressed CFTR with a deletion of PDZ-binding motif in the COOH-terminus (CFTR-DelTRL). Mature fully glycosylated protein was the predominant form of CFTR for all mutants, except CFTR-EENKVR-alanine mutant, where core glycosylated form of CFTR (band B) was the major translation product (Fig. 5, top cell line). To test for interaction, each of the five cell lines was cotransfected with two versions of NHERF1: one tagged with HA, and the other with myc. Whole cell lysates probed with anti-HA or anti-c-myc confirmed that both NHERF1 constructs were coexpressed in each cell line (Fig. 5, middle cell lines). Immunoprecipitation of HA-NHERF1 followed by immunoblotting with anti-c-myc revealed Co-IP of myc-NHERF1 in cell lines stably expressing CFTR-WT (Fig. 5, bottom cell line). In contrast, parental MDCK-FRT cells demonstrated minimal interaction between the two forms of NHERF1. Furthermore, the level of interaction between HA-NHERF1 and myc-NHERF1 in cells expressing CFTR-EENKVR, -EEN, and -KVR-alanine mutants or DelTRL mutant was comparable to that observed in parental cells. These results were reproducible in multiple experiments (Fig. 5, bottom graph). Together, these results indicated that WT CFTR enhanced oligomerization of the HA- and myc-tagged forms of NHERF1, and that both ¹⁴¹⁷EENKVR¹⁴²² and ¹⁴⁷⁸TRL¹⁴⁸⁰ sequences were required for enhanced oligomerization.

Fig. 5.

¹⁴¹⁷EENKVR¹⁴² and PDZ-binding motif in full-length CFTR foster NHERF1 interaction in polarized epithelial cell line. Coimmunoprecipitation of myc-NHERF1 by HA-NHERF1 from the protein lysate of MDCK II-FRT cells stably expressing full-length GFP-CFTR-wild type, GFP-CFTR alanine mutants (¹⁴¹⁷EENKVR¹⁴²², ¹⁴¹⁷EEN¹⁴¹⁹, ¹⁴²⁰KVR¹⁴²²), and GFP-CFTR-DelTRL. Both myc-NHERF1 and HA-NHERF1 were transiently cotransfected. Graph illustrates fold increase in myc-NHERF1 binding to HA-NHERF1 in CFTR expressing MDCK cells compared with MDCK parentals. Values are means ± SE (n = 3). ***P < 0.0001, one-way ANOVA.

Both ¹⁴¹⁷EENKVR¹⁴²² sequence and PDZ-binding motif act in concert to facilitate localization of CFTR in MDCK II cells.

In a previous study, we demonstrated that C111 was able to disrupt the apical localization of full-length CFTR-WT in polarized (MDCK II) cells. Presence of the PDZ-binding motif in the COOH-terminal peptide was required to displace CFTR from the apical membrane (48). We, therefore, asked if coexpression of C111 with the EENKVR sequence substituted with alanines could affect apical localization of CFTR expressed in polarized MDCK II cells. C111 redistributed full-length CFTR to the cytosol, while C111-DelTRL failed to alter localization (Fig. 6). Importantly, C111-EENKVR-alanine mutant did not alter the localization of WT CFTR, even though it contained an intact PDZ-binding motif (Fig. 6). These results demonstrate that preferential apical localization of full-length CFTR requires the presence of both the internal EENKVR region and the terminal PDZ-binding motif.

Fig. 6.

¹⁴¹⁷EENKVR¹⁴²² and PDZ-binding motif are required for apical localization of full-length CFTR in polarized MDCK II cells. Confocal images (×63 magnification) of immunostained MDCK-FRT cells imaged in the xz-plane (scanned apical to basal membrane) are shown. First column shows MDCK FRT cells transiently expressing full-length GFP wild-type CFTR as control. Second, third, and fourth columns show the effect of coexpression of myc-tagged C111, C111-EENKVR-alanine mutant, and C111-DelTRL, respectively, on the localization of full-length GFP wild-type CFTR. Expression of the myc-tagged C111 constructs is shown in red detected by Alexa Fluor 647, and GFP-tagged WT CFTR in green detected by Alexa Fluor 488. All cells were counterstained with DAPI (blue) to detect the nuclei. Scale bar, 5 μm.

DISCUSSION

This study provides new insight into the biochemical and cellular mechanistic basis of the complex dynamic interactions between NHERF1 and CFTR COOH-terminus. A systematic search of the CFTR COOH-terminus identified a six-amino acid region ¹⁴¹⁷EENKVR¹⁴²² that promotes the oligomerization of NHERF1 at concentrations that are well below its physiological levels in multiple tissues that express CFTR. To validate the importance of this region in the context of full-length CFTR, we show that the newly identified region, in combination with the PDZ-binding motif, affects NHERF1 oligomerization in MDCK stable cells expressing full-length CFTR. Finally, we show that the six-amino acid region, along with the PDZ-binding motif, contribute to the preferential apical localization of full-length CFTR. These results inform our understanding of the interactions between CFTR and NHERF1 beyond the PDZ-binding motif and suggest potential strategies to increase the stability of mutant CFTR at the apical surface.

Earlier studies that demonstrated NHERF1 self-association (18, 67) have been recently questioned (19, 64). One argument put forth is that blot overlay assays are insufficient to firmly conclude that oligomers are formed in solution and in the cellular environment (19, 64). Our data corroborate studies reporting that NHERF1 self-association is an extremely low-affinity process. Higher molecular mass complexes corresponding to dimers are not observed until protein concentrations reach 120 μM, as reported previously (40). A comparative gene expression profile generated by RNA-Seq in multiple tissues, including sweat gland in our study, indicate that intracellular concentration of NHERF1 is in the nanomolar range. At such a low concentration, NHERF1 is unlikely to oligomerize in any of these tissues without a catalyst. Here, we have addressed the issue by demonstrating that CFTR COOH-terminus fosters NHERF1 oligomerization at physiological levels. We believe that our data from binding experiments are reliable because evidence of interaction was as follows: 1) consistent across three different methods, blot overlay, in vitro pull-down, and Co-IP of differently tagged NHERF1 proteins; and 2) shown to be sequence specific by all three methods.

The studies presented here refine the region conferring oligomerization to ¹⁴¹⁷EENKVR¹⁴²² in the CFTR COOH-terminus. This six-amino acid sequence identified in this study does not conform to the preferred COOH-terminal peptide binding motif ([S/T]-x-Φ or D-[S/T]-x-[V/I/L]) for the class I PDZ domain of NHERF1. Therefore, we have considered that it may act as an internal PDZ ligand (28). Internal ligands have a consensus PDZ-binding sequence within a stable β-finger structure with a hairpin region that mimics a free carboxyl group, as demonstrated in neuronal NOS and ETA endothelial receptor (26, 55). The EENKVR region of CFTR is predicted to lie between two β-strands capable of forming an anti-parallel β-sheet (62). Internal consensus PDZ motifs without a β-finger [e.g., D-(S/T)-F-L in EGFR (37), K-T-x-x-x-Φ in Frizzled (38) and Idax (42), I-G-Y-G-Y-L in Grasp65 (65), and a stable β-finger without a PDZ motif (H-R-E-M-A-V) in PALs1 (56)] have also been identified, but none match the EENKVR sequence or even the shorter sequences ¹⁴¹⁷EEN¹⁴¹⁹ or ¹⁴²⁰KVR¹⁴²² that were found to enhance NHERF1 oligomerization. Interestingly, an acidic/dileucine motif (E-K-E-N-K-L-L) situated 36–42 to amino acids from the COOH-terminus has been shown to affect endocytic recycling modes of the β₂-adrenergic receptor (25). We have shown that alanine mutagenesis of either glutamate (¹⁴¹⁷E or ¹⁴¹⁸E) is sufficient to eliminate enhancement of NHERF1 oligomer formation. Multiple sequence alignment of the amino acid sequences of CFTR carboxyl-terminal tails across diverse species reveals that ¹⁴¹⁷E is strongly conserved (2). From these observations, we speculate that ¹⁴¹⁷E and ¹⁴¹⁸ E within ¹⁴¹⁷EENKVR¹⁴²² motif provide negatively charged groups that are critical for the forma tion of NHERF1 oligomers. Although single alanine substitution of the remaining amino acids did not affect NHERF1 oligomerization, replacement of three residues ¹⁴²⁰KVR¹⁴²² did. Therefore, we speculate that ¹⁴¹⁹NKVR¹⁴²² sequence forms a structural feature essential for the activity of the ¹⁴¹⁷EE¹⁴¹⁸ residues.

PDZ domains have evolved to form multiprotein complex assembly by the simultaneous interaction with more than one ligand (8, 30). The contribution of the PDZ-binding ligand in decreasing the threshold concentration for oligomerization of NHERF1 seems to be a possible mechanism for the formation of multiprotein complexes at the appropriate location in the cell to facilitate interaction with membrane proteins, i.e., near or at the cell membrane. Yang et al. have illustrated an important role of NHERF2/NHERF3 heterodimerization in the regulation of NHE3 activity by promoting assembly of macrocomplexes (76). Since NHERF proteins are involved in the regulation of many other transporters or receptors, macrocomplexes mediated by NHERF interactions could play other important physiological functions (76). Based on the above results, we theorized that ¹⁴¹⁷EENKVR¹⁴²² mediated NHERF1 oligomerization contributes to the preferential CFTR localization at the apical surface. Initial reports suggested that the PDZ-binding motif was responsible for apical distribution of CFTR in MDCK II and bronchial epithelial cells (53, 54). However, later studies provided evidence for NHERF1-independent CFTR localization. When the PDZ-binding motif was deleted by COOH-terminal truncations (6, 26, or 40 residues) or blocked by proteins fused to the COOH-terminus (HA, enhanced GFP, and enhanced GFP-DTRL), the localization of variant CFTR was indistinguishable from WT CFTR (4). These findings suggested that apical targeting of CFTR was independent of NHERF1 interaction and involved sorting signals outside of the PDZ-binding motif of CFTR. Indeed, subsequent studies from our group indicated that PDZ-binding motif alone was essential but not sufficient for apical localization, and that additional sequences in the COOH-terminus mediate the process (47, 48).

Here we provide evidence that the ¹⁴¹⁷EENKVR¹⁴²² sequence, in concert with the COOH-terminal PDZ-binding motif, plays a role in the preferential apical localization of CFTR. Coexpression of CFTR-C111-WT, but not C111-EENKVR-alanine or C111-delTRL, caused redistribution of full-length WT CFTR into the cytoplasm. We speculate that the introduction of the WT COOH-terminal construct containing both oligomerization motif (¹⁴¹⁷EENKVR¹⁴²²) and terminal PDZ-binding motif (¹⁴⁷⁸TRL¹⁴⁸⁰) compete with the full-length WT CFTR for NHERF1 molecules to displace it from the apical surface. A similar phenomenon has been reported for TIP-1 and CASK; these two PDZ proteins compete with each other to dislodge the ligand, Kir2.3, from its membrane anchoring complex and confer a shift toward endosomal targeting of the channel (1). It is also plausible that coexpression of WT CFTR COOH-terminal construct sequestered NHERF1, thereby altering CFTR localization. This process may be akin to viral hijacking that provides high-affinity COOH-terminal motifs, resulting in sequestration of PDZ proteins and rewiring of cell signaling (12, 31). One example is the hijacking of polarity protein PALS1 by the SARS coronavirus E protein, which alters the tight junction formation and epithelial morphogenesis (31). Of note, three COOH-terminal motifs (¹⁴³⁰LL, ¹⁴¹³FLVI, and ¹⁴²⁴YDSI) have been implicated in endocytic trafficking of CFTR (20, 29). We propose that the ¹⁴¹⁷EENKVR¹⁴²² region is a new addition to the motifs at the CFTR COOH-terminus involved in its CFTR localization.

Finally, there is evidence that manipulation of NHERF1 interactions with CFTR has clinical and therapeutic implications. An infant carrying a COOH-terminal truncation mutation E1418X that deletes the ¹⁴¹⁷EENKVR¹⁴²² motif (paired with F508del) exhibits CF based on absent chloride transport on nasal potential difference tracings (66). In contrast, two truncation mutations that leave the ¹⁴¹⁷EENKVR¹⁴²² intact (S1455X and Q1476X) have been associated with elevated sweat chloride, but no lung disease (45, 63). Arora et al. have shown that F508del can be stabilized at the cell surface with the use of VX-809 (a corrector molecule) by potentiation of its interaction with NHERF1 (3). More recently, it has been demonstrated that exchange protein directly activated by cAMP (EPAC1) interacts with NHERF1 through CFTR, and its activation is additive to F508del CFTR rescue with VX-809 (41). Taken together, the characterization of the ¹⁴¹⁷EENKVR¹⁴²² motif provides a new potential mechanism to manipulate the interaction between CFTR and NHERF1 to amplify the effect of small-molecule treatments.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-044003 to G. R. Cutting, and Grant CUTTIN15XX1 from Cystic Fibrosis Foundation Therapeutics to N. Sharma.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

N.S., J.L., L.B.G., A.L., M.J.P., T.E., E.D., C.-H.N., G.D.R., D.B., and M.M. performed experiments; N.S., J.L., P.S., and M.A. analyzed data; N.S., J.L., A.P., and G.R.C. interpreted results of experiments; N.S. and J.L. prepared figures; N.S., J.L., and G.R.C. drafted manuscript; N.S. and G.R.C. edited and revised manuscript; G.R.C. approved final version of manuscript.