NHE3 inhibits PKA-dependent functional expression of CFTR by NHERF2 PDZ interactions

www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 347 (2006) 452–459

BBRC

NHE3 inhibits PKA-dependent functional expression of CFTRby NHERF2 PDZ interactions

M. Favia a,1, T. Fanelli a,1, A. Bagorda a,2, F. Di Sole b, S.J. Reshkin a, P.G. Suh c,L. Guerra a, V. Casavola a,*

a Department of General and Environmental Physiology, University of Bari, Via Amendola 165/A, 70126 Bari, Italyb Institute for Cell and Molecular Biosciences, School of Biomedical Sciences, The Medical School, University of Newcastle, UK

c Department of Life Science, Pohang University of Science and Technology, Pohang, Kyungbuk 790-784, Republic of Korea

Received 22 May 2006Available online 28 June 2006

Abstract

It has been shown that when CFTR and NHE3 are co-expressed on the apical membrane of the A6-NHE3 cell monolayers, the twotransporters interact via a shared regulatory complex composed of NHERF2, ezrin, and PKA. We observe here that co-expression ofNHE3 reduced both PKA-dependent apical CFTR expression and its activation once in place by approximately 50%. To analyze the roleof NHERF2 in this process, we transfected NHE3 expressing and non-expressing A6 monolayers with NHERF2 cDNA in which itsbinding domains had been deleted. When only CFTR is expressed on the apical membrane, deletion of any of the NHERF2 bindingdomains inhibited both PKA-dependent apical CFTR expression and its activation, while when NHE3 was co-expressed with CFTRPDZ2 deletion was without effect on CFTR sorting and activity. This suggests that when the PDZ2 domain is ‘‘sequestered’’ by inter-acting with NHE3 it can no longer participate in CFTR functional expression.� 2006 Elsevier Inc. All rights reserved.

Keywords: CFTR; NHE3; NHERF2; PKA; PDZ domains; Ezrin

The cystic fibrosis transmembrane conductance regula-tor (CFTR) protein is responsible for the cAMP/PKA reg-ulated chloride conductance. In addition, CFTR has beenreported to modify both the function of different epithelialtransport proteins, such as the amiloride-sensitive sodiumchannel [1,2] and ROMK channels [3], and their interac-tions [4]. In normal cells, newly synthesized wild-type(wt) CFTR, after passing ER quality control, is exportedfrom the Golgi to the apical membrane as fully glycosylat-ed protein. The regulation of the half-life of wt CFTR inthe plasma membrane is complex and not yet completely

0006-291X/$ - see front matter � 2006 Elsevier Inc. All rights reserved.

doi:10.1016/j.bbrc.2006.06.112

* Corresponding author. Fax: +39 080 5443388.E-mail address: [email protected] (V. Casavola).

1 These authors contributed equally to this work.2 Present address: Laboratory of Cellular and Molecular Biology,

National Cancer Institute, National Institutes of Health, Bethesda, MD20892-4256, USA.

clear. It is known to depend on both clathrin-dependentendocytosis from the plasma membrane [5] occurring inRab5-specific endosomes and on the endocytotic recyclingof CFTR from these endosomes to the plasma membranethrough Rab11-specific recycling vesicles [6].

Growing evidence indicates that interactions of wtCFTR with PDZ containing proteins, such as NHERF,CAL, and CAP70 [7–9], compartmentalize CFTR in a mul-tiprotein complex which is important not only for the reg-ulation of the CFTR chloride channel activity but also forits targeting and retention in the apical membrane [10,11].Indeed, deletion of the CFTR domain that interacts withPDZ domains has been demonstrated to eliminate thepolarized expression of CFTR in the apical membrane ofhuman airway epithelial cells, 16HBE14o� [12].

NHERF1 (NHE exchanger regulatory factor) [13] andits related homologous NHERF2 [14] are proteins initiallyidentified as the cofactor required for cAMP-dependent

M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459 453

inhibition of NHE3 [15]. Both NHERF1 and NHERF2contain two PDZ domains and are distributed in the apicalregion of polarized epithelial cells where they interact witha wide variety of channels, transporters, and receptors [16].NHERF was the first PDZ protein found to bind the C-ter-minus target domain of CFTR [7] and this interaction wasproposed to have a central role in the apical expression ofCFTR that is essential for vectorial chloride transport.Additionally, NHERF has been demonstrated to associateby its C-terminus with the cytoskeletal adaptor protein,ezrin [13], which is also a Protein Kinase A anchoring(AKAP) protein. In this way, PKA is compartmentalizedin the vicinity of CFTR [11,17] permitting the PKA-dependent regulation of apical CFTR activity.

The capacity of both NHERF1 and NHERF2 to oligo-merize further extends the potential of NHERF to seques-ter interactive proteins within membrane microdomainsand, in this way, promotes the interaction between CFTRand other functionally related ion transporters or recep-tors. In this context, CFTR has been demonstrated to reg-ulate the activity and expression of the Na+/H+ exchanger,NHE3 [18], and, in a polarized renal cell model, we showedthat CFTR and NHE3 reciprocally interact via a sharedregulatory complex comprised of NHERF2, ezrin, andPKA [19]. These observations raise the possibility thatthe binding of CFTR and NHE3 in a common PDZdomain-based scaffold complex could provide a means bywhich epithelial cells can modulate net NaCl transport.

Here, we have evaluated this possibility by analyzing thedynamic regulatory interaction between CFTR and NHE3for binding with NHERF and show how NHE3 co-expres-sion can indirectly regulate CFTR activity and expressionon the apical membrane. Our results indicate that the co-expression of NHE3 with CFTR decreases both the sortingand the activity of CFTR possibly by modifying the com-position and stoichiometry of the multiprotein complexand/or switching CFTR function from chloride channelto conductance regulator.

Materials and methods

Cell culture. Experiments were performed with the epithelial cell line,A6, and its stable transfected variants. A6-NHE3 and A6-NHE3S552A celllines were generated by stable transfection in A6 cells of the cDNAsencoding wild-type ratNHE3 and ratNHE3 mutated at single endogenousserine position on the cytoplasmic tail of NHE3 as previously described[19]. Cells were grown in 0.8% concentrated Dulbecco’s modified Eagle’smedium supplemented with 25 mM NaHCO3, 10% heat-inactivated fetalbovine serum and penicillin/streptomycin at 28 �C in 5% CO2, and sub-cultured weekly by trypsinization using a Ca2+/Mg2+-free salt solutioncontaining 0.25% (w/v) trypsin and 1 mM EGTA. For CFTR-dependentchloride efflux, Na+/H+ exchanger activity, and biotinylation experiments,cells were seeded on 0.4-lm pore size PET filter inserts (Falcon BD Bio-sciences Labware) coated with collagen. Cells generally reached conflu-ency between 7 and 8 days after seeding when the culture medium waschanged three times a week.

Transfection of NHERF2 cDNAs. At 70–80% confluence, cells weretransiently transfected with His-tagged NHERF2 cDNAs deleted in dif-ferent domains and inserted into the pcDNA3.1/Hygro+ vector. DPDZ1and DPDZ2 are cDNAs encoding NHERF2 in which PDZ1 and PDZ2

domains have been deleted and DERM is a cDNA encoding NHERF2deleted of the ERM region (Fig. 1S) and generated as previously described[20]. All these clones behave as dominant mutants over the endogenousprotein. Cells were transiently transfected using Escort IV reagent (Sigma)according to the manufacturer’s protocol and the experiments were con-ducted 48 h later.

Fluorescence measurements of apical chloride efflux. Confluent cellmonolayers were grown on permeable collagen-coated filters, transientlytransfected and 48 h later were loaded overnight in culture medium con-taining 5 mM of the Cl�-sensitive dye, MQAE, at 28 �C in a CO2 incu-bator, and then inserted into a perfusion chamber that allowedindependent perfusion of apical and basolateral cell surfaces. Chlorideefflux was measured with a Cary Eclipse Varian spectrofluorometer asrecently described [21]. All experiments were performed in Hepes-bufferedbicarbonate-free media (Cl� medium (in mM): NaCl 110, KCl 3, CaCl2 1,MgSO4 0.5, Hepes 10, KH2PO4 1, and glucose 5, and Cl�-free medium:NaNO3 105, KNO3 3.2, MgSO4 0.8, NaH2PO4 1, Hepes 10, CaNO3 5, andglucose 5).

Detection of CFTR-dependent chloride efflux. A6 and A6-NHE3polarized monolayers exhibited a basal chloride efflux under baselineconditions when chloride was replaced by apical nitrate (0.018 ± 0.003D(F/F0)/min and 0.014 ± 0.0015 D(F/F0)/min, in A6 and A6-NHE3 cellmonolayers, respectively, n = 8, n.s.). Stimulation of PKA by addition of10 lM forskolin (FSK) significantly increased this apical chloride efflux.The addition of the CFTR inhibitor, 100 lM glibenclamide [22], to theapical perfusion fluids 5 min before and during the next FSK stimulationinhibited this PKA-dependent increase to basal levels. CFTR-dependentchloride efflux is, therefore, defined as the difference between the rate ofFSK stimulated chloride efflux before and after apical glibenclamidetreatment [19] (Fig. 2S). At the end of each experiment a two-point cali-bration procedure was performed: the maximal intensity of fluorescence(F0) was determined by perfusing the cells with the Cl�-free medium onboth sides of the monolayer and the minimal fluorescence was obtained bythen exposing the cells to a solution containing KSCN, as reported [19].For data analysis, the value for minimal fluorescence was subtracted fromthe experimentally measured fluorescence and the resulting fluorescencewas divided by the value of F0. The rate of Cl� efflux was determined bylinear regression analysis of 30 points taken at 4-s intervals and expressedin arbitrary slope changes in D(F/F0)/min.

Measurement of Na+/H+-exchange activity. Cytoplasmic pHi wasmeasured spectrofluorimetrically with the fluorescent pH-sensitive probeBCECF-AM (5 lM) trapped intracellularly in confluent A6-NHE3 mon-olayers grown on collagen-coated permeable filters as we previouslydescribed [19]. The cell monolayers were inserted into a perfusion chamberthat allowed independent perfusion of apical and basolateral surfaces.Na+/H+ exchange activity was measured by monitoring pHi recovery afteran acid load produced with the NH4Cl prepulse technique under contin-uous perfusion using a gravity-driven system [19] and the rate of pHi

recovery was determined by linear regression analysis of 15 points taken at4-s intervals. pHi was estimated from the ratio of BCECF fluorescencecalibrated by using the K+ nigericin approach [23]. All experiments wereperformed in Hepes-buffered bicarbonate-free media (Na+ medium (inmM): NaCl 110, KCl 3, CaCl2 1, MgSO4 0.5, Hepes 10, KH2PO4 1, andglucose 5, and Na+-free medium: TMACl 110, KCl 3, CaCl2 1, MgSO4

0.5, Hepes 10, KH2PO4 1, and glucose 5; pH 7.5).Protein extraction and Western blotting. Cells grown on collagen-

coated permeable filters were transfected at 70–80% confluence and 48 hlater washed with NaCl medium, lysed in lysis buffer A (0.4% sodiumdeoxycholate, 1% Igepal CA-630 (Sigma), 50 mM EGTA, 10 mM Tris–HCl, pH 7.4, with added protease inhibitor mixture), sonicated for 10 s,centrifuged for 10 min (16,000g), and the pellet was discarded. Superna-tant protein concentration was measured by Bradford method [24] and analiquot of 30 lg of protein was diluted in Laemmli buffer, heated at 100 �Cfor 5 min, and separated by 4–12% SDS–PAGE in a Criterion XT precastgel (Bio-Rad). The separated proteins were transferred to Immobilon P(Millipore) in a Trans-Blot semidry electrophoretic transfer cell (Amer-sham Biosciences) for immunoblotting. The primary antibodies used wereanti-hCFTR monoclonal antibody against the C-terminus (R&D Systems,

Fig. 1. Role of NHERF2 domains in the regulation of CFTR activity inpolarized A6 and A6-NHE3 cell monolayers. Polarized monolayers of A6and A6-NHE3 cells were transfected with vectors containing NHERF2cDNA deleted of various domains (PDZ1, PDZ2, and ERM) asrepresented in Supplemental Fig. 1S. After 48 h, the Cl� transport activityof CFTR was determined as described in Materials and methods. CFTR-dependent chloride transport was calculated from the difference inalterations of forskolin-stimulated fluorescence measurements in theabsence and presence of 100 lM of the CFTR inhibitor, glibenclamide(see Supplemental Fig. 2S). Each bar represents the mean ± SE of thecalculated differences. Statistical comparisons were made using unpairedStudent’s t test with respect to the values obtained in non-transfectedmonolayers.

454 M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459

MAB25031, dilution 1:500), anti-hNHERF2 polyclonal antibody (AlphaDiagnostic International, dilution 1:1000), and anti-His-tag monoclonalantibody (Cell signaling, dilution 1:1000). The secondary antibodies wereanti-mouse IgG for monoclonal antibody and anti-rabbit IgG for poly-clonal antibody (Sigma, dilution 1:2000). Immunocomplexes weredetected with ECL plus reagent (Amersham Biosciences), and densito-metric quantification and image processing were carried out using AdobePhotoshop and the ‘‘Image’’ software package (version 1.61, NationalInstitutes of Health, Bethesda, MD).

Biotinylation of apical membrane proteins. Cells grown on collagen-coated permeable filters were transfected at 70–80% confluence and 48 hlater washed with NaCl medium and then basolaterally treated or not for10 min with 10 lM FSK in NaCl medium at room temperature. Afterwashing rapidly, the cells were incubated with 2 mg/ml sulfo-NHS-biotin(Sigma) in NaCl medium for 30 min at 4 �C. All further steps wereperformed in a cold room. Free sulfo-NHS-biotin was removed bywashing cells twice with 0.1 M glycine in NaCl medium. Cells were lysedin Lysis Buffer A, sonicated, centrifuged, and the pellet was discarded.Volumes of supernatant, containing equal amounts of protein, wereincubated overnight at 4 �C with gentle mixing with the same amount ofstreptavidin–Agarose beads (Pierce) (50 ll of streptavidin/mg of biotin).Streptavidin-bound complexes were pelleted (16,000g) and after twowashes with buffer lysis, biotinylated proteins were eluted in Laemmlibuffer. The eluted proteins were subjected to SDS–PAGE and Westernblotting as described above.

Immunofluorescence analysis. To control the level of transfection ofcells with plasmids containing the various 6-His-tagged constructs ofNHERF2, cells were grown on round glass coverslips, were transfectedand 48 h later washed, fixed in 4% paraformaldehyde, and permeabilizedin 0.1% Triton X-100. The cells were then blocked in 0.1% gelatin for10 min and incubated in a wet environment with primary monoclonalantibody anti-His-tag (dilution 1:100) for 1 h, treated with 0.1% gelatin,and then incubated with the secondary antibody, goat anti-mouse IgGconjugated to FITC (Molecular Probes, dilution 1:1000) for 1 h. Thecoverslips were mounted onto slides with Vectashield mounting mediumwith DAPI. Fluorescence data were collected with a Nikon Eclipse TE2000-S with a 40· oil immersion objective and processed by using theMetamorph Imaging System. Collected images were exported to AdobePhotoshop for subsequent analyses.

Data analysis. Data are presented as means ± SE for the number ofsamples indicated (n). Statistical comparisons were made using eitherunpaired or paired data Student’s t test. Differences were considered sig-nificant when p < 0.05.

Results and discussion

Using two renal cell culture models derived from thenephron of Xenopus laevis, A6 cells, where endogenousCFTR is expressed on the apical membrane, andA6-NHE3 cells, where the stably transfected NHE3 is co-expressed on the apical membrane with CFTR [25], wepreviously found that CFTR, ezrin, and NHE3 were recog-nized by NHERF2 suggesting that CFTR and NHE3could reciprocally interact via a common multiproteincomplex comprised of NHERF2, ezrin, and PKA [19]. InA6-NHE3 cells the inhibition of CFTR expression by anti-sense treatment resulted in a strong decrease of PKA-dependent regulation of NHE3 activity and, reciprocally,the PKA-dependent regulation of CFTR was negativelymodulated by the presence and activation of NHE3.Indeed, the PKA-dependent regulation of CFTR-mediatedCl� secretion was lower in A6-NHE3 cells than in A6 cells[19]. To elucidate the mechanism by which NHE3 co-expression influences CFTR regulation and analyze the

role of the different NHERF2 PDZ and ERM bindingdomains in regulating CFTR sorting to the apical mem-brane and its activity once in place, we transiently transfec-ted both A6 and A6-NHE3 cells with 6His-taggedNHERF2 cDNA in which either the PDZ1, PDZ2 orERM domain had been deleted (Fig. 1SA). Western blotanalysis of the monolayers transfected with each NHERF2construct with anti-6His tag demonstrated that the trans-fection occurred correctly (Fig. 1SB). Immunofluorescenceanalysis of the transfected monolayers demonstrated that60.62 ± 2.72% (n = 24) of the cells were transfected.

Role of PDZ domains of NHERF in regulating CFTR

activity in both A6 and A6-NHE3 cells

In the first set of experiments, we measured the effect ofthe various NHERF2 constructs on PKA-regulatedCFTR-dependent chloride efflux across the apical mem-brane of both A6 and A6-NHE3 polarized cell monolayersas described in Materials and methods. Importantly, treat-ment with either the transfection vehicle, Escort IV, ortransfection with the empty vector, pcDNA3.1/Hygro+,did not significantly alter CFTR-dependent chloride trans-port (data not shown). In Fig. 1 the data on CFTR-depen-dent chloride transport obtained in the A6 and A6-NHE3cell lines are summarized. The transfection of A6 cells (leftpanel) with cDNA encoding the PDZ1 domain deletedNHERF2 completely abolishes the PKA-regulated activity

Fig. 2. Role of NHERF2 domains in the regulation of NHE3 activity inpolarized A6-NHE3 cell monolayers grown on permeable filters. (A)Representative traces showing the effect of FSK on NHE3 exchangeactivity in non-transfected polarized A6-NHE3 monolayers. NHE3activity was assayed as the initial rate of apical Na+-dependent pHi

recovery after intracellular acidification by perfusion with NH4+ medium

followed by Na+-free tetramethylammonium (TMA) medium before andafter 10 min of preincubation with FSK (10 lM). The Y-axis is the pHi

corresponding to the nigericin calibration. (B) Representative tracesshowing the effect of FSK on NHE3 exchange activity in A6-NHE3monolayers transfected with NHERF2-DPDZ2 cDNA. (C) Percentages ofPKA-dependent inhibition of NHE3 activity induced by FSK inA6-NHE3 cell monolayers transfected with vectors containing NHERF2cDNA deleted of different domains (PDZ1, PDZ2, and ERM). Each barrepresents the mean ± SE. Statistical comparisons were made usingunpaired Student’s t test with respect to the values obtained innon-transfected monolayers.

M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459 455

of CFTR, while a 50% inhibition of CFTR activity wasobtained in A6 cells transfected with the PDZ2 domaindeleted NHERF2. These data suggest that CFTR interactswith NHERF2 via both its two PDZ domains althoughwith a higher affinity for the PDZ1 as has been previouslyobserved in other cell lines [26]. Further, over-expression ofERM-binding domain deleted NHERF2 (DERM) thatretains the two PDZ domains caused an almost completeinhibition of PKA-regulated CFTR-dependent chloridetransport. This was expected since ezrin serves as an anchorfor protein kinase A (AKAP) and tethers PKA directly toCFTR for efficient and specific phosphorylation [17].

As can be observed comparing the bars representingCFTR-dependent chloride transport in control A6 and inA6-NHE3 cells in Fig. 1, the co-expression of NHE3 withCFTR in A6-NHE3 cells down-regulates PKA-dependentCFTR activation by approximately 50%, as already shownin our prior studies [19]. As observed in A6 cells, inA6-NHE3 cells (right panel) the deletion of either thePDZ1 or the ERM domain of NHERF2 strongly inhibitedthe PKA-mediated regulation of CFTR activity. However,PDZ2 domain deletion, differently from what happens inA6 cells, had no effect on CFTR activity suggesting thatwhen the PDZ2 domain is already ‘‘sequestered’’ by itsinteraction with NHE3, its deletion no longer has any sig-nificant effect on CFTR activity. An indirect confirmationof this hypothesis came from the similarity of CFTR-dependent chloride efflux observed in A6 cells transfectedwith PDZ2 deleted NHERF2 and in the not transfectedA6-NHE3 cells (0.013 ± 0.0005, n = 4, vs 0.012 ± 0.0011D(F/F0)/min, n = 8, in A6 cells transfected with DPDZ2NHERF2 and in not transfected A6-NHE3 cells,respectively).

These findings suggest that CFTR functions more effi-ciently when it can interact with both PDZ domains ofNHERF2 and confirm a previous study showing thatCFTR association with both PDZ1 and PDZ2 domainsof a single NHERF protein could promote CFTR dimer-ization with an increase in the open probability, Po [26].Additionally, Raghuram demonstrated that the phosphor-ylation of residues in the PDZ2 domain shifts theNHERF–CFTR equilibrium and causes CFTR to bebound to the PDZ1 domains of two different NHERF mol-ecules [27]. In this way, the CFTR channel is not cross-linked and, therefore, has a reduced Po. Our data providean analogous mechanism: the NHE3 interaction withPDZ2 could alter CFTR’s interaction with NHERF2 andswitch CFTR’s function from a Cl� channel to a conduc-tance regulator.

Role of NHERF2 PDZ and ERM domains in regulating

NHE3 activity in A6-NHE3 cells

NHE3 has been shown to interact through an internalmotif with the second PDZ domain of NHERF [28] thatbehaves as an adaptor protein between NHE3 and ezrinfacilitating the PKA-mediated phosphorylation and

subsequent inhibition of NHE3 [14]. We analyzed the roleof NHERF2 and its different PDZ domains in altering thePKA-dependent NHE3 regulation in the A6-NHE3 cellmonolayers. Na+/H+ exchange activity was measured asthe rate of Na+-induced recovery of cytosolic pH (pHi)after an acid load in A6-NHE3 cell monolayers and theeffect of PKA stimulation was analyzed by adding FSKon the basolateral side (Figs. 2A and B). In not transfectedA6-NHE3 cells, PKA activation inhibited NHE3 activity(from 0.062 ± 0.006 to 0.040 ± 0.004 DpH/min, n = 7,p < 0.01 before and after FSK incubation, respectively;�35.14 ± 2.39% reduction in activity) and the treatmentof A6-NHE3 monolayers with the transfection vehicleEscort IV did not alter this PKA-dependent NHE3 inhibi-tion (�36.99 ± 2.34%, n = 6, n.s.). As shown in Fig. 2C,

Fig. 3. PKA-dependent regulation of the expression of biotinylatedCFTR in polarized A6, A6-NHE3, and A6-NHE3S552A cell monolayersgrown on permeable filters. Expression level of biotinylated apical CFTRwas analyzed in A6, A6-NHE3, and A6-NHE3S552A monolayers byWestern blotting using anti-hCFTR antibody as described in Materialsand methods. (A) Representative Western blot of a typical experiment,showing that in resting conditions the expression level of CFTR is almostthe same in all three cell lines (42.05 ± 6.63 OD, 44.47 ± 3.34 OD, and42.49 ± 5.32 OD, n = 7 in A6, A6-NHE3, and A6-NHE3S552A, respec-tively, n.s.) while the FSK stimulation induced an increase in expressionlevels of apical CFTR that is summarized in (B) and expressed as thepercentage of the increase in the expression levels in respect to unstim-ulated cells for each cell line. Data represent means ± SE.

456 M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459

while PDZ1 deletion did not significantly change the PKA-dependent regulation of NHE3 (�38.32 ± 6.56%, n = 4),PDZ2 deletion almost completely reversed the inhibitoryeffect of PKA (�6.04 ± 4.25%, n = 4), without alteringthe basal NHE3 activity (0.061 ± 0.013 DpH/min, n = 4)(Fig. 2B) indicating that, differently from that happens inother cellular systems [29], the PDZ2 domain of NHERF2plays a fundamental role in PKA-dependent regulation ofNHE3 activity. Moreover, the finding that ERM deletionalso prevented the PKA-dependent inhibition of NHE3suggests that, as reported [30], cAMP/PKA regulation ofNHE3 requires the constitution of a complex comprisedof NHE3, NHERF, and the AKAP protein, ezrin, plusan intact actin cytoskeleton for optimal function.

Role of NHE3 in regulating PKA-dependent CFTR apical

sorting

PKA-dependent CFTR sorting to the apical membraneis a complex and highly regulated process that contributesto the fine tuning of CFTR activation. Stimulation ofCFTR by PKA causes inhibition of CFTR endocytosis[31] with a concomitant decrease of the endosomal CFTRpool and a consequent increase in the plasma membrane.Moreover, PKA activation increases CFTR traffickingfrom a sub-apical domain to the plasma membrane[32–34]. However, this observation remains controversialas other studies have observed that cAMP did not stimu-late CFTR translocation to the apical membrane [35,36],suggesting that PKA-dependent trafficking and membranelocalization is cell-type specific and may reflect the need forexpression of the right set of interacting proteins [37].

We first decided to determine if the decrease in CFTRactivity upon NHE3 expression observed in Fig. 1 couldbe due to a reduction in its PKA-dependent apical mem-brane expression. Apical CFTR expression was measuredby surface biotinylation in A6 and A6-NHE3 polarizedmonolayers grown on permeable filters. As shown inFig. 3A, in resting conditions the expression level of apicalfully mature CFTR was similar in A6 and A6-NHE3 cells(42.05 ± 6.63 vs 44.47 ± 3.34 OD, n = 7, in A6 andA6-NHE3 cells, respectively, n.s.), while activation ofPKA by FSK treatment induced an increase of apicalCFTR expression in A6 cells that was twice that observedin A6-NHE3 cells (62.29 ± 7.03% vs 31.58 ± 2.25%, n = 7,in A6 and A6-NHE3 monolayers, respectively). These datasuggest that NHE3 co-expression not only down-regulatesthe PKA-dependent activation of CFTR (Fig. 1), but alsodecreases PKA-dependent CFTR expression on the apicalmembrane. As our prior studies demonstrated that block-ing the PKA-dependent phosphorylation of serine 552 bymutating it to alanine (A6-NHE3S552A cells) was sufficientto block the inhibitory influence of NHE3 on CFTR activ-ity [19], we also examined if blocking the PKA-dependentphosphorylation of NHE3 can influence the PKA-depen-dent sorting of CFTR to the apical membrane. As can beseen in Figs. 3A and B, in A6-NHE3S552A cell monolayers

the extent of CFTR polarized expression was not signifi-cantly different from that found in A6 cells. This suggestseither that NHE3 can compete with CFTR for associationwith the regulatory NHERF2–ezrin–PKA complex or thatCFTR functions to direct PKA-dependent regulation toNHE3 when the two transporters are functionally co-expressed. Another possibility is that since the NHERF2–NHE3 interaction occurs in the same domain (amino acids475–711) in which phosphorylation occurs [14], it is possi-ble that the mutation of Ser-552 simply impairs the bindingof NHERF2 with NHE3 and, therefore, allows the interac-tion between CFTR and both of the two PDZ domains ofNHERF2.

Role of NHERF2 binding domains in regulating

PKA-dependent CFTR apical sorting

While NHERF’s interaction with CFTR is essential forCFTR retention on the apical membrane of polarized cells[10], the potential role of NHERF in regulating PKA-dependent CFTR sorting to the apical membrane of

M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459 457

polarized cells is still ill-defined. To evaluate whether theabove-observed PKA-dependent increase in CFTR expres-sion to the apical membrane of A6 and A6-NHE3 cells isinfluenced by NHERF2 and to analyze if the greaterPKA-dependent CFTR sorting in A6 compared toA6-NHE3 cells was related to a different NHERF2-medi-ated mechanism, we measured apical CFTR expressionby surface biotinylation of polarized monolayers in bothresting conditions and after PKA activation in A6 andA6-NHE3 cell monolayers that had been transfected withPDZ1, PDZ2 or ERM domain deleted NHERF2 con-structs as above reported. Treatment with either the trans-fection vehicle, Escort IV, or transfection with the emptyvector, pcDNA3.1/Hygro+, did not significantly alter theapical CFTR expression (data not shown).

We found that while transfection with the PDZ1, PDZ2or ERM domain deletion constructs did not significantlychange apical CFTR expression in resting conditions ineither the A6 or A6-NHE3 monolayers (data not shown),the PKA-dependent CFTR expression on the apical mem-brane was greatly altered in both cell lines demonstratingthat the NHERF2-CFTR interaction is required for theproper PKA-dependent sorting of CFTR to the plasmamembrane. In particular, Fig. 4 shows that in A6 cell mon-olayers any of the three domain PDZ1, PDZ2 or ERMdeletions greatly decreased apical CFTR expression.PDZ1 domain deleted NHERF2 completely abolished thePKA-dependent CFTR sorting, while a lower level of inhi-

Fig. 4. Role of NHERF2 domains in PKA-dependent regulation of theexpression of biotinylated CFTR in polarized A6 and A6-NHE3 cellmonolayers. Monolayers of A6 and A6-NHE3 cells, grown on permeablefilters, were transfected with vectors containing NHERF2 cDNA deletedof different domains (PDZ1, PDZ2, and ERM) and after 48 h the apicalexpression of CFTR was determined in the presence or absence of 10 lMFSK. Cell surface membrane proteins were biotinylated and analyzed byWestern blotting using anti-hCFTR antibody as described in Materialsand methods. The expression level of apical CFTR is expressed as thepercentage change of the biotinylated CFTR measured in A6 andA6-NHE3 cell monolayers in each transfection condition after FSKstimulation with respect to the levels in un-stimulated conditions.

bition was obtained transfecting A6 cells with the DPDZ2NHERF2 indicating that in A6 cells both NHERF2 PDZdomains are implicated not only in the regulation of CFTRactivity (see Fig. 1) but also in PKA-dependent CFTR sort-ing, although with different affinity. The same relationshipbetween regulation of CFTR activity and sorting was alsoobtained in A6-NHE3 cell monolayers; indeed only PDZ1deletion was able to significantly impair PKA-dependentCFTR sorting while PDZ2 deletion was in-effective, con-firming once again that in A6-NHE3 cell the PDZ2 domaindoes not participate in regulation of CFTR functionalexpression. Lastly, the deletion of ERM domain stronglydecreased the PKA-dependent apical sorting of CFTR tothe membrane in both cell lines confirming the hypothesisthat ezrin may facilitate the expression of CFTR at the api-cal membrane via interactions with the actin cytoskeleton.Indeed, a recent observation [38] that ezrin, by sequentialprotein–protein interaction events, is necessary to enablethe two NHERF PDZ domains to bring two CFTR C-ter-minals into proximity to each other raises the possibilitythat ezrin could directly regulate CFTR sorting by dynam-ically modulating the composition and the stoichiometry ofNHERF complexes.

In conclusion, the findings reported here indicate thatNHERF2 serves as a linker protein in a regulatory complexcomprised of CFTR, NHERF2, ezrin, and PKA. Further,we provide evidence for a role for the PDZ domains ofNHERF2 and the ERM proteins in regulating the efficien-cy of PKA-dependent CFTR functional expression. Inaddition, we show that NHERF2 regulates CFTR sortingon the apical membrane. Importantly, the co-expressionof NHE3 with CFTR decreases both the sorting and theactivity of CFTR possibly by modifying the compositionand stoichiometry of the multiprotein complex and switch-ing CFTR function from chloride channel to conductanceregulator.

Acknowledgments

This work was supported by the Italian Cystic FibrosisResearch Foundation and from the CEGBA: ‘‘Centro diEccellenza di Genomica in campo Biomedico e Agrario.’’

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.bbrc.2006.06.112.

References

[1] K. Kunzelmann, R. Schreiber, R. Nitschke, M. Mall, Control ofepithelial Na+ conductance by the cystic fibrosis transmembraneconductance regulator, Pflugers Arch. 440 (2000) 193–201.

[2] R. Schreiber, P. Kindle, T. Benzing, G. Walz, K. Kunzelmann,Control of the cystic fibrosis transmembrane conductance regulatorby alphaG(i) and RGS proteins, Biochem. Biophys. Res. Commun.281 (2001) 917–923.

458 M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459

[3] D. Yoo, T.P. Flagg, O. Olsen, V. Raghuram, J.K. Foskett, P.A.Welling, Assembly and trafficking of a multiprotein ROMK (Kir 1.1)channel complex by PDZ interactions, J. Biol. Chem. 279 (2004)6863–6873.

[4] A.A. Konstas, J.P. Koch, S.J. Tucker, C. Korbmacher, Cystic fibrosistransmembrane conductance regulator-dependent up-regulation ofKir1.1 (ROMK) renal K+ channels by the epithelial sodium channel,J. Biol. Chem. 277 (2002) 25377–25384.

[5] L.S. Prince, K. Peter, S.R. Hatton, L. Zaliauskiene, L.F. Cotlin, J.P.Clancy, R.B. Marchase, J.F. Collawn, Efficient endocytosis of thecystic fibrosis transmembrane conductance regulator requires atyrosine-based signal, J. Biol. Chem. 274 (1999) 3602–3609.

[6] M. Gentzsch, X.B. Chang, L. Cui, Y. Wu, V.V. Ozols, A. Choudhury,R.E. Pagano, J.R. Riordan, Endocytic trafficking routes of wild typeand DeltaF508 cystic fibrosis transmembrane conductance regulator,Mol. Biol. Cell. 15 (2004) 2684–2696.

[7] R.A. Hall, L.S. Ostedgaard, R.T. Premont, J.T. Blitzer, N. Rahman,M.J. Welsh, R.J. Lefkowitz, A C-terminal motif found in the beta2-adrenergic receptor, P2Y1 receptor and cystic fibrosis transmembraneconductance regulator determines binding to the Na+/H+ exchangerregulatory factor family of PDZ proteins, Proc. Natl. Acad. Sci. USA95 (1998) 8496–8501.

[8] J. Cheng, B.D. Moyer, M. Milewski, J. Loffing, M. Ikeda, J.E.Mickle, G.G. Cutting, M. Li, B.A. Stanton, W.B. Guggino, AGOLGI associated PDZ domain protein modulates cystic fibrosistransmembrane regulator plasma membrane expression, J. Biol.Chem. 13 (2001) 13.

[9] S. Wang, H. Yue, R.B. Derin, W.B. Guggino, M. Li, Accessoryprotein facilitated CFTR–CFTR interaction, a molecular mecha-nism to potentiate the chloride channel activity, Cell 103 (2000)169–179.

[10] A. Swiatecka-Urban, M. Duhaime, B. Coutermarsh, K.H. Karlson, J.Collawn, M. Milewski, G.R. Cutting, W.B. Guggino, G. Langford,B.A. Stanton, PDZ domain interaction controls the endocyticrecycling of the cystic fibrosis transmembrane conductance regulator,J. Biol. Chem. 277 (2002) 40099–40105.

[11] D.B. Short, K.W. Trotter, D. Reczek, S.M. Kreda, A. Bretscher, R.C.Boucher, M.J. Stutts, S.L. Milgram, An apical PDZ protein anchorsthe cystic fibrosis transmembrane conductance regulator to thecytoskeleton, J. Biol. Chem. 273 (1998) 19797–19801.

[12] B.D. Moyer, M. Duhaime, C. Shaw, J. Denton, D. Reynolds, K.H.Karlson, J. Pfeiffer, S. Wang, J.E. Mickle, M. Milewski, G.R.Cutting, W.B. Guggino, M. Li, B.A. Stanton, The PDZ-interactingdomain of cystic fibrosis transmembrane conductance regulator isrequired for functional expression in the apical plasma membrane, J.Biol. Chem. 275 (2000) 27069–27074.

[13] D. Reczek, M. Berryman, A. Bretscher, Identification of EBP50: APDZ-containing phosphoprotein that associates with members of theezrin-radixin-moesin family, J. Cell Biol. 139 (1997) 169–179.

[14] C.H. Yun, G. Lamprecht, D.V. Forster, A. Sidor, NHE3 kinase Aregulatory protein E3KARP binds the epithelial brush border Na+/H+ exchanger NHE3 and the cytoskeletal protein ezrin, J. Biol.Chem. 273 (1998) 25856–25863.

[15] E.J. Weinman, D. Steplock, S. Shenolikar, CAMP-mediated inhibi-tion of the renal brush border membrane Na+–H+ exchanger requiresa dissociable phosphoprotein cofactor, J. Clin. Invest. 92 (1993)1781–1786.

[16] J.W. Voltz, E.J. Weinman, S. Shenolikar, Expanding the role ofNHERF, a PDZ-domain containing protein adapter, to growthregulation, Oncogene 20 (2001) 6309–6314.

[17] F. Sun, M.J. Hug, N.A. Bradbury, R.A. Frizzell, Protein kinase Aassociates with cystic fibrosis transmembrane conductance regulatorvia an interaction with ezrin, J. Biol. Chem. 275 (2000) 14360–14366.

[18] W. Ahn, K.H. Kim, J.A. Lee, J.Y. Kim, J.Y. Choi, O.W. Moe, S.L.Milgram, S. Muallem, M.G. Lee, Regulatory interaction between thecystic fibrosis transmembrane conductance regulator and HCO3-salvage mechanisms in model systems and the mouse pancreatic duct,J. Biol. Chem. 276 (2001) 17236–17243.

[19] A. Bagorda, L. Guerra, F. Di Sole, C. Hemle-Kolb, R.A. Cardone, T.Fanelli, S.J. Reshkin, S.M. Gisler, H. Murer, V. Casavola, Reciprocalprotein kinase A regulatory interactions between cystic fibrosistransmembrane conductance regulator and Na+/H+ exchanger iso-form 3 in a renal polarized epithelial cell model, J. Biol. Chem. 277(2002) 21480–21488.

[20] J.I. Hwang, K. Heo, K.J. Shin, E. Kim, C. Yun, S.H. Ryu, H.S.Shin, P.G. Suh, Regulation of phospholipase C-beta 3 activity byNa+/H+ exchanger regulatory factor 2, J. Biol. Chem. 275 (2000)16632–16637.

7[21] L. Guerra, T. Fanelli, M. Favia, S.M. Riccardi, G. Busco, R.A.Cardone, S. Carrabino, E.J. Weinman, S.J. Reshkin, M. Conese,V. Casavola, Na+/H+ exchanger regulatory factor isoform 1overexpression modulates cystic fibrosis transmembrane conduc-tance regulator (CFTR) expression and activity in humanairway 16HBE14o- cells and rescues DeltaF508 CFTR func-tional expression in cystic fibrosis cells, J. Biol. Chem. 280(2005) 40925–40933.

[22] B.D. Schultz, A.K. Singh, D.C. Devor, R.J. Bridges, Pharmacologyof CFTR chloride channel activity, Physiol. Rev. 79 (1999)S109–S144.

[23] J.A. Thomas, R.N. Buchsbaum, A. Zimniak, E. Racker, IntracellularpH measurements in Ehrlich ascites tumor cells utilizing spectroscopicprobes generated in situ, Biochemistry 18 (1979) 2210–2218.

[24] M.M. Bradford, A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding, Anal. Biochem. 72 (1976) 248–254.

[25] F. Di Sole, V. Casavola, L. Mastroberardino, F. Verrey, O.W. Moe,G. Burckhardt, H. Murer, C. Helmle-Kolb, Adenosine inhibits thetransfected Na+–H+ exchanger NHE3 in Xenopus laevis renalepithelial cells (A6/C1), J. Physiol. 515 (Pt. 3) (1999) 829–842.

[26] V. Raghuram, D.D. Mak, J.K. Foskett, Regulation of cystic fibrosistransmembrane conductance regulator single-channel gating bybivalent PDZ-domain-mediated interaction, Proc. Natl. Acad. Sci.USA 98 (2001) 1300–1305.

[27] V. Raghuram, H. Hormuth, J.K. Foskett, A kinase-regulatedmechanism controls CFTR channel gating by disrupting bivalentPDZ domain interactions, Proc. Natl. Acad. Sci. USA 100 (2003)9620–9625.

[28] E.J. Weinman, Y. Wang, F. Wang, C. Greer, D. Steplock, S.Shenolikar, A C-terminal PDZ motif in NHE3 binds NHERF-1 andenhances cAMP inhibition of sodium-hydrogen exchange, Biochem-istry 42 (2003) 12662–12668.

[29] E.J. Weinman, D. Steplock, S. Shenolikar, NHERF-1 uniquelytransduces the cAMP signals that inhibit sodium-hydrogen exchangein mouse renal apical membranes, FEBS Lett. 536 (2003) 141–144.

[30] J.B. Wade, P.A. Welling, M. Donowitz, S. Shenolikar, E.J. Weinman,Differential renal distribution of NHERF isoforms and their colo-calization with NHE3, ezrin, and ROMK, Am. J. Physiol. CellPhysiol. 280 (2001) C192–C198.

[31] G.L. Lukacs, G. Segal, N. Kartner, S. Grinstein, F. Zhang,Constitutive internalization of cystic fibrosis transmembrane conduc-tance regulator occurs via clathrin-dependent endocytosis and isregulated by protein phosphorylation, Biochem. J. 328 (Pt. 2) (1997)353–361.

[32] T. Jilling, K.L. Kirk, Cyclic AMP and chloride-dependent regulationof the apical constitutive secretory pathway in colonic epithelial cells,J. Biol. Chem. 271 (1996) 4381–4387.

[33] A. Tousson, C.M. Fuller, D.J. Benos, Apical recruitment of CFTR inT-84 cells is dependent on cAMP and microtubules but not Ca2+ ormicrofilaments, J. Cell Sci. 109 (Pt. 6) (1996) 1325–1334.

[34] M. Howard, X. Jiang, D.B. Stolz, W.G. Hill, J.A. Johnson, S.C.Watkins, R.A. Frizzell, C.M. Bruton, P.D. Robbins, O.A. Weisz,Forskolin-induced apical membrane insertion of virally expressed,epitope-tagged CFTR in polarized MDCK cells, Am. J. Physiol. CellPhysiol. 279 (2000) C375–C382.

[35] B.D. Moyer, J. Loffing, E.M. Schwiebert, D. Loffing-Cueni, P.A.Halpin, K.H. Karlson, Ismailov II, W.B. Guggino, G.M. Langford,

M. Favia et al. / Biochemical and Biophysical Research Communications 347 (2006) 452–459 459

B.A. Stanton, Membrane trafficking of the cystic fibrosis geneproduct, cystic fibrosis transmembrane conductance regulator, taggedwith green fluorescent protein in madin-darby canine kidney cells, J.Biol. Chem. 273 (1998) 21759–21768.

[36] L.S. Prince, A. Tousson, R.B. Marchase, Cell surface labeling ofCFTR in T84 cells, Am. J. Physiol. 264 (1993) C491–C498.

[37] C.A. Bertrand, R.A. Frizzell, The role of regulated CFTR trafficking inepithelial secretion, Am. J. Physiol. Cell Physiol. 285 (2003) C1–C18.

[38] J. Li, Z. Dai, D. Jana, D.J. Callaway, Z. Bu, Ezrin controls themacromolecular complexes formed between an adapter proteinNa+/H+ exchanger regulatory factor and the cystic fibrosis transmem-brane conductance regulator, J. Biol. Chem. 280 (2005) 37634–37643.