QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Vol. 17, Issue 6, 2661-2673, June 2006

This Article Abstract Full Text (PDF) All Versions of this Article: E05-09-0843v1 17/6/2661    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cha, B. Articles by Donowitz, M. Search for Related Content PubMed PubMed Citation Articles by Cha, B. Articles by Donowitz, M.

The NHE3 Juxtamembrane Cytoplasmic Domain Directly Binds Ezrin: Dual Role in NHE3 Trafficking and Mobility in the Brush Border

Boyoung Cha^*, Ming Tse^*, Chris Yun^*,, Olga Kovbasnjuk^*, Sachin Mohan^*, Ann Hubbard, Monique Arpin, and Mark Donowitz^*

*Departments of Physiology and Medicine, Gastroenterology Division and Department of Cell Biology, Johns Hopkins University School of Medicine, Baltimore, MD 21205; Unite Mixte de Recherche 144, Centre National de la Recherche Scientifique/Institut Curie, 75248 Paris, France; Division of Digestive Diseases, Department of Medicine, Emory University School of Medicine, Atlanta, GA 30322

Submitted September 16, 2005; Revised February 28, 2006; Accepted March 6, 2006
Monitoring Editor: Anthony Bretscher

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Based on physiological studies, the epithelial brush-border (BB) Na⁺/H⁺ antiporter3 (NHE3) seems to associate with the actin cytoskeleton both by binding to and independently of the PDZ domain containing proteins NHERF1 and NHERF2. We now show that NHE3 directly binds ezrin at a site in its C terminus between aa 475-589, which is separate from the PSD95/dlg/zonular occludens-1 (PDZ) interacting domain. This is an area predicted to be -helical, with a positive aa cluster on one side (K516, R520, and R527). Point mutations of these positively charged aa reduced (NHE3 double mutant [R520F, R527F]) or abolished (NHE3 triple mutant [K516Q, R520F, R 527F]) ezrin binding. Functional consequences of these NHE3 point mutants included the following. 1) A marked decrease in surface amount with a greater decrease in NHE3 activity. 2) Decreased surface expression due to decreased rates of exocytosis and plasma membrane delivery of newly synthesized NHE3, with normal total expression levels and slightly reduced endocytosis rates. 3) A longer plasma membrane half-life of mutant NHE3 with normal total half-life. 4) Decreased BB mobile fraction of NHE3 double mutant. These results show that NHE3 binds ezrin directly as well as indirectly and suggest that the former is related to 1) the exocytic trafficking of and plasma membrane delivery of newly synthesized NHE3, which determines the amount of plasma membrane NHE3 and partially determines NHE3 activity, and 2) BB mobility of NHE3, which may increase its delivery from microvilli to the intervillus clefts, perhaps for NHE3-regulated endocytosis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Multiple members of the nine-isoform NHE gene family interact with the actin cytoskeleton (Zachos et al., 2005). There is some understanding of the proteins involved and functional consequences for NHE1 and NHE3. NHE1 directly binds the actin binding proteins ezrin/radixin/moesin (ERM) using two positively charged amino acid-rich areas in the juxtamembrane region of the NHE1 C terminus. This association has structural and functional consequences for the fibroblast cytoskeleton, with NHE1 acting as an anchor in the lamellipodia for actin via ERM binding (Denker et al., 2000; Denker and Barber, 2002). Disruption of NHE1/ERM binding is associated with abnormal stress fibers and focal adhesions and results in lack of polar migration by fibroblasts in filling in a wound (Denker et al., 2000; Denker and Barber, 2002; Baumgartner et al., 2004).

The epithelial brush-border (BB) NHE isoform NHE3 also associates with the actin cytoskeleton. We previously showed that the association of NHE3 with the actin cytoskeleton occurs via two areas in its C-terminal regulatory domain. One area, between aa 585-689 (Yun et al., 1997, 1998), bound the PSD95/dlg/zonular occludens-1 (PDZ) domain-containing proteins, NHERF1 or NHERF2. The NHERFs simultaneously bind NHE3 and link it to the actin cytoskeleton via binding to ezrin. NHERF1 and NHERF2 are scaffolding proteins that assemble multiprotein signaling complexes, largely through their multiple PDZ domains and C-terminal ERM binding domain (Shenolikar et al., 2004; Donowitz et al., 2005). For example, NHE3 protein complexes, which include NHERF1/NHERF2, ezrin, and PKAII, facilitate the phosphorylation of NHE3 by protein kinase A, which is necessary for cAMP inhibition of NHE3 activity (Yun et al., 1998; Zizak et al., 1999; Weinman et al., 2000). Also, cGMP inhibition of BB NHE3 involves a complex of NHE3, NHERF2, and cGMP kinase II (Cha et al., 2005). The identified functional consequences of this NHERF binding include formation of multiple, large, multiprotein complexes and also limitation of mobility of NHE3 in epithelial cell BB (Cha et al., 2004; Li et al., 2004).

There is a second NHE3–actin cytoskeleton association that is separate and upstream from the NHERF binding domain. The evidence includes that NHE3 truncated to aa 585 (NHE3 585), which lacks all NHERF binding, colocalized with actin in the apical cytoskeleton and was affected by disrupting the actin cytoskeleton similarly to wild-type NHE3 in terms of exhibiting reduced mobility at the apical membrane of opossum kidney (OK) cells (Cha et al., 2004).

One of the ways NHE3 associates with the cytoskeleton involves ezrin. Ezrin has two complementary self-association domains known as N- and C-ERM association domains (ERMADs), which encompass residues 1-296 and 479-585, respectively (Bretscher et al., 2002). These N- and C-ERMADs interact strongly with each other, masking the membrane protein and F-actin binding sites and maintaining ezrin as an inactive monomer in the cytoplasm. The major F-actin binding site of ezrin is within its carboxy-terminal 30 amino acids. The N-terminal domain of ERM proteins have protein 4.1, ezrin, radixin, moesin (FERM) domains, which act as multifunctional protein and lipid binding sites. Recently the three-dimensional structure of the ezrin N-terminal FERM domain was solved (Pearson et al., 2000; Smith and Cerione, 2002; Hamada et al., 2003; Smith et al., 2003; Finnerty et al., 2004) and shown to be composed of three subdomains: FERM-I (aa 4-82), FERM-II (aa 96-195), and FERM-III (aa 204-297). These subdomains together form a compact cloverleaf-shaped structure. In addition to associating directly with the cytoplasmic tails of some membrane proteins, the FERM domain of ezrin interacts strongly with NHERF1 and NHERF2 (Finnerty et al., 2004).

In the current study, we identify the upstream cytoskeleton association domain of NHE3 as an ezrin binding domain and demonstrate its functional role in NHE3 trafficking and BB mobility. This report provides the first evidence that proteins the bind ezrin at two sites can use those interactions to regulate separate functions.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Materials
QuikChange site-directed mutagenesis kit was from Stratagene (La Jolla, CA). Nigericin and 2',7'-bis(2-carboxyethyl)-5(6)-carboxylfluorescein (BCECF) were from Invitrogen (Carlsbad, CA). EZ-Link Sulfo-NHS-SS-biotin and Sulfo-NHS-acetate were from Pierce Chemical (Rockford, IL). Glutathione-Sepharose 4B beads were from GE Healthcare (Little Chalfont, Buckinghamshire, United Kingdom). Monoclonal anti-vesicular stomatitis virus glycoprotein (VSV-G) antibodies were derived from the P5D4 hybridoma from T. Kreiss. Polyclonal anti-NHERF1 (-5199) and anti-NHERF2 (-2570) antibodies were described previously (Yun et al., 1997, 1998). Polyclonal phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558) antibody was from Cell Signaling Technology (Beverly, MA). Monoclonal anti-ezrin antibody (E8897) was from Sigma-Aldrich (St. Louis, MO). Monoclonal anti-ezrin recognizes only ezrin, but anti-p-ezrin antibodies recognize ezrin, radixin, and moesin. Thus, our studies allow demonstration of ezrin/NHE3 interaction, although studies with p-ezrin (Figure 8) could identify any ERM proteins. Monoclonal anti-VSV-G antibody immobilized on agarose was from Sigma-Aldrich and monoclonal anti-hemagglutinin (HA) affinity matrix was from Roche Diagnostics (Indianapolis, IN). Monoclonal Rab11 antibody was from BD Biosciences (San Jose, CA). Polyclonal anti-cis-Golgi (GPP130) antibody was from Covance (Princeton, NJ). Protease inhibitor cocktail [4-(2-aminoethyl)benzenesulfonyl fluoride, transepoxysuccinyl-L-leucylamido-(4-guanidino)butane (E-64), bistatin, leupeptin, aprotinin, and sodium EDTA] was from Sigma-Aldrich.

Construction of Expression Vectors for NHE3 C-Terminus Ezrin Binding Mutants
The NHE3 C-terminus ezrin binding mutations were made using the QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's protocol. The template for mutagenesis was the pcDNA3.1/Hygro⁺ vector (Invitrogen) containing rabbit NHE3, which had an epitope tag derived from the VSV-G at the COOH-terminal end, as described previously (Cha et al., 2005). The sense oligonucleotides (5'-3') used for introducing the mutations in pcDNA3.1/Hygro⁺ NHE3V were (designation used for mutations is domain mutated of NHE3 C terminus [F1 or F2] single [S], double [D], or triple [T] mutant): NHE3VF1 single (R527F), CGGCCCAGAAGTCTTTCGACCGGATT CTG; NHE3VF1 double (R520F, R527F), GCAAACTGCTCATGTTCCAGTCGGCC CAG AAGTCTTTCGACCGG ATTCTG; and NHE3VF1 triple (K516Q, R520F, R527F), GGTTCCTCAGCCAACTGCTCATGTTC CAGTCGGCCCAGAAGTCTTTCGACCG GATTCTG, and NHE3VF2 triple RRR (R604L, R605F, R606F), TCGCTGGAGCAGCTGT TCTTCAGCGTGCGCGAC, respectively (16 cycles, 95°C, 30 s; 55°C, 1 min; and 68°C, 16 min). The changed bases are underlined. These cDNAs were sequenced in full. NHE3-EGFP, NHE3F1 D-EGFP (R520F, R527F), and NHE3 F1T (K516Q, R520F, R527F) were assembled into pEGFPN3 vector (Clontech, Mountain View, CA) in frame with the C-terminal enhanced green fluorescent protein (EGFP) coding sequence.

Fusion Proteins and In Vitro Translation Products
The recombinant proteins glutathione S-transferase (GST), GST-WT-ezrin (aa 1-585), GST-N-ezrin (NE) (aa 1-309), GST-C-ezrin (aa 310-585), GST-NE1 (aa 1-97), GST-NE2 (aa 1-198), and GST-NE3 (aa 1-310) were made in the pGEX2T-1 vector (the plasmids were from M. Arpin). 6xHis-tagged fusion proteins made in the pET30a vector included the NHE3 C terminus, F1 (aa 475-589), F2 (aa 590-667), F3 (aa 668-747), F4 (aa 748-832), F_Full CT (aa 475-832), and NHERF1 and NHERF2 (Cha et al., 2004).

In vitro-translated, [³⁵S]methionine-labeled NHE3 C-terminal fragments were generated using the TnT Quick-coupled transcription/translation system (Promega, Madison, WI). The amount of each labeled protein used in binding assays was determined by the intensity of the autoradiography signal per amount of GST-labeled fusion protein. GST-fusion proteins were expressed in BL21 cells and induced with 1 mM isopropyl -D-thiogalactoside at 37°C for 4 h. Fusion proteins were purified on GST-Sepharose according to the manufacture's recommendations (GST Gene Fusion system; GE Healthcare). The concentration of each GST fusion protein was estimated by Ponceau S (Sigma-Aldrich) staining after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) transfer to nitrocellulose, using known amounts of bovine serum albumin (BSA) as an internal standard or by Bradford protein assay.

In Vitro Binding Assays
GST fusion proteins immobilized on reduced glutathione (GSH)-Sepharose beads were incubated with protein lysates or in vitro translation products in lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 10% glycerol, 1% NP-40, 200 µM Na₃VO₄, and protease inhibitors) overnight at 4°C. Complexes were pelleted at 10,000 x g for 2 min and washed three times in 1 ml of lysis buffer. Bound proteins were eluted from the beads by heating in Laemmli's buffer for 5 min at 80°C. Proteins were separated in 10% or 14% SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted.

Immunoprecipitation and Immunoblot Analysis
Coimmunoprecipitation experiments were performed using lysates from PS120/NHE3 and PS120/NHE3_585 cells. Cell lysates were in lysis buffer (60 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM KCl, 5 mM EDTA trisodium, 3 mM lysis buffer EGTA, 1 mM Na₃VO₄, and 1% Triton X-100 with protease inhibitor cocktail [Sigma-Aldrich]). Aliquots (2 mg of protein) of lysate were incubated with either monoclonal anti-VSV-G antibody immobilized to agarose or monoclonal anti-HA affinity matrix overnight at 4°C. The aliquots were washed five times (60 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM KC1, 5 mM EDTA trisodium, 3 mM EGTA, 1 mM Na₃VO₄, and 1% Triton X-100), and then bound proteins were eluted in Laemmli's buffer, separated by SDS-PAGE, and transferred to nitrocellulose. The blots were probed with monoclonal anti-VSV-G or monoclonal anti-ezrin primary antibodies and fluorescently labeled secondary antibody (IRDye 800 or Alexa Fluor 680) according to the manufacturer's protocol, and bands were visualized by the Odyssey system (LI-COR, Lincoln, NE).

Cell Culture and cDNA Transfection of NHE3 Ezrin Binding Mutants
The cDNAs of wild-type NHE3V and its ezrin binding mutants NHE3VF1 single mutant (R527F), NHE3VF1 double mutant (R520F, R527F), NHE3VF1 triple mutant (K516Q, R570F, R527F), NHE3VF2 triple mutant (R604L, R605F, R606F), and NHE3_585V (aa 1-585) in pcDNA3.1/Hygro+ vector (Invitrogen) were transfected into Na⁺/H⁺ exchanger-deficient PS120 fibroblasts using 10 µl of Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol, and resistant clones were selected with 600 U/ml hygromycin. Stably transfected PS120 fibroblasts were maintained at 37°C in a humidified atmosphere with 5% CO₂ in DMEM supplemented with 10% (vol/vol) fetal bovine serum (FBS), 100 U/ml penicillin, and 100 µg/ml streptomycin (PS120 media). In addition, some NHE3-transfected cells (NHE3 wild-type and NHE3F2 triple mutant) were selected by an H⁺-killing procedure consisting of 50 mM NH₄Cl/saline solution for 1 h, followed by an isotonic 2 mM Na⁺ solution for 1 h (Levine et al., 1995). Surviving cells were then placed in normal culture medium and allowed to reach 30–50% confluence. The H⁺-killing process was initially repeated every 2–3 d until more than 50% of the cells survived and was then repeated every week. Acid selection was not used in the NHE3F1 mutants (single, double, and triple).

Immunofluorescence
PS120 cells were grown on tissue culture coverslips to 70% confluence, fixed with 3% paraformaldehyde/phosphate-buffered saline for 20 min at 4°C, and the residual formaldehyde was neutralized with 20 mM glycine in phosphate-buffered saline (PBS) for 10 min. Cells were then permeabilized for 30 min in 0.1% saponin/PBS before being blocked for 30 min in 1% BSA/PBS supplemented with 10% FBS. Cells were incubated with primary antibody in 1% BSA/PBS for 60 min at room temperature. After three 10-min washes in 0.1% saponin/PBS, goat secondary antibodies (Alexa conjugates; Invitrogen) were added at 1:100 dilution in 1% BSA/PBS, incubated for 30 min, and again washed three times for 10 min in 0.1% saponin/PBS. Cells were then mounted on slides with Prolong (Invitrogen) antifade reagent and viewed on a Zeiss LSM 410 confocal fluorescence microscope.

Measurement of Na⁺/H⁺ Exchange Activity
The transfected PS120 cells grown to 70–80% confluence on glass coverslips were placed in serum-free medium for 4 h to arrest division. The Na⁺/H⁺ exchanger activities of transfected PS120 cells were measured using the intracellular pH-sensitive dye BCECF-acetoxymethyl ester, as described previously (Levine et al., 1995; Cha et al., 2005).

Cell Surface Biotinylation and Immunoblotting
Transfected PS120 cells were grown to 70–80% confluence in 10-cm Petri dishes. The cells were then serum starved for 4 h. All subsequent manipulations were performed at 4°C. For surface labeling of NHE3, cells were incubated with 0.5 mg/ml NHS-SS-biotin (biotinylation solution; Pierce Chemical) for 20 min and repeated once, solubilized with the lysis buffer, and then incubated for 1 h with streptavidin-agarose beads. Western analysis and the quantification of the surface fraction were performed as described previously (Akhter et al., 2002; Kim et al., 2002).

Endocytosis
Endocytosis was measured by a protocol slightly modified from the reduced GSH-resistant endocytosis assay we described previously (Lee-Kwon et al., 2003). Cells at 4°C were labeled with 1.5 mg/ml sulfo-NHS-SS-biotin for 40 min and quenched at 4°C. Cells were then incubated at 37°C for 30 min to allow endocytosis and rinsed twice with ice-cold PBS at 4°C. All subsequent steps were at 4°C. Surface biotin was cleaved by washing with 50 mM Tris-HCl and 150 mM GSH, pH 8.8. In this way, the freshly endocytosed proteins bearing biotin were protected from cleavage with GSH. Cells were then solubilized in N⁺ buffer (60 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM KCl, 5 mM EDTA trisodium, 3 mM EGTA, 1 mM Na₃VO₄, and 1% Triton X-100), and biotinylated proteins were retrieved with streptavidin-agarose beads and quantified as described above (Lee-Kwon et al., 2003).

Exocytic Insertion of NHE3 into the Plasma Membrane
To measure exocytic insertion of NHE3 (also called endocytic recycling), NHS-reactive sites on the cell surface were first blocked by pretreatment with sulfo-NHS-acetate as described above, with some slight modifications (Lee-Kwon et al., 2003). Cells were rinsed twice with PBS-Ca-Mg (PBS with 0.1 mM Ca²⁺ and 1 mM Mg²⁺) at 4°C. The apical surface was then exposed to 1.5 mg/ml sulfo-NHS-acetate in PBS-Ca-Mg for 2 h at 4°C. After quenching for 20 min at 4°C, the cells were rinsed with PBS and then incubated in normal PS120 media at 4°C for 30 min as a control and at 37°C for 10 min to allow intracellular NHE3 to reach the plasma membrane (exocytosis). Cells were then treated with 0.5 mg/ml sulfo-NHS-SS-biotin as described above and lysed with N⁺ buffer. The biotinylated fraction, which represents newly inserted surface proteins, was precipitated by streptavidin-agarose, and the precipitate was subjected to SDS-PAGE and Western blotting with anti-VSV-G antibody, as described previously (Cavet et al., 2001). Given the long half-life of total cellular NHE3 (Cavet et al., 2001), there was only a small contribution by the synthetic pathway to this pool of NHE3.

Newly Synthesized NHE3 Delivered to the Plasma Membrane
To measure the newly synthesized NHE3 that was delivered to the plasma membrane, PS120/NHE3 and NHE3 F1 double mutant cells were incubated in Met/Cys-free media for 1.5 h and then pulse labeled with Tran³⁵S-Label reagent (Met/Cys: 0.3 mCi/ml; ICN Pharmaceuticals, Costa Mesa, CA). The excel pulse-media was quickly removed by washing with DMEM media without Met and Cys for 30 min at 37°C. Cells were then chased for 0, 30, 60, 120, and 180 min in PS120 media at 37°C. Cells were then cooled to 4°C, and cell surface biotinylation was performed as described above (Akhter et al., 2002; Kim et al., 2002). Subsequently, the cells were lysed and proteins were isolated by streptavidin beads. Proteins were separated by 10% SDS-PAGE and transferred to nitrocellulose. Detection of ³⁵S-labeled NHE3V and NHE3V F1 double mutant was after 24-48 h using the Storm 860 PhosphorImager system (GE Healthcare). Intensity measurements of ³⁵S-labeled NHE3V and NHE3V F1 double mutant were quantified with MetaMorph software (Molecular Devices, Sunnyvale, CA) (top lanes). Western analysis was then performed on the same samples using anti-VSV-G antibody by the Odyssey system (bottom lanes) and quantified with MetaMorph software.

Measurement of Half-Life of Plasma Membrane NHE3 Using Cell Surface Biotinylation
To determine the half-life of plasma membrane NHE3V and NHE3V F1 double mutant, a cell surface biotinylation method was used, as described previously (Cavet et al., 2001).

Total NHE3 Half-Life Determined by Pulse-Chase Labeling of Cultured Cells. PS120 cells stably transfected with wild-type NHE3V and NHE3V F1 double mutant were grown to 70% confluence and then starved in depletion medium (DMEM without Met and Cys) for 1 h at 37°C. The cells were then incubated with Met/Cys-free DMEM containing 0.2 mCi/ml [³⁵S]Met/Cys cell labeling mix (GE Healthcare) for 4 h. To chase, the labeling solution was aspirated and the cells rinsed four times with PBS. Cells were then incubated with DMEM containing 10% serum, 2 mM Met, and 2 mM Cys for between 0 and 41 h. The details are as described previously (Cavet et al., 2001).

Fluorescence Recovery after Photobleaching (FRAP). To determine the lateral mobility of NHE3-EGFP and NHE3 F1 double mutant-EGFP at the apical surface of polarized OK cells (the endogenous NHE3 expression was previously minimized by the "acid suicide" technique; Pouyssegur et al., 1984), we used FRAP, as reported previously (Cha et al., 2004). To perform FRAP, OK cells were cultured on glass-bottomed 35-mm plastic culture dishes in DMEM media (without phenol red) supplemented with 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin at 37°C in a 5% CO₂, 95% air atmosphere until 100% confluent. The cells were then transiently transfected using Lipofectamine, as described previously (Cha et al., 2004), with NHE3-EGFP and NHE3 F1 double mutant-EGFP and studied 48 h later.

FRAP was performed on a stage heated to 37°C of a Zeiss LSM 410 confocal microscopy using the 488-nm line of a 400-mW Kr/Ar laser in conjunction with a 100x Zeiss1.4 NA Planapochromat oil immersion objective. The details were described previously (Cha et al., 2004), with signal collected in the OK cell apical membrane in the area of microvilli not over the juxtanuclear pool of NHE3, which we previously reported did not represent newly trafficked NHE3 (Cha et al., 2004).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
NHE3 Coimmunoprecipitates Ezrin in the Absence of NHERF1/NHERF2 Binding
Recently, we suggested that NHE3 might interact with the actin cytoskeleton in a manner separate from the established indirect binding via NHERF1 or 2 (Cha et al., 2004). We hypothesized that this NHERF-independent binding of NHE3 to actin might involve ezrin. This was tested here by determining an NHE3 truncation that did not bind NHERF1, and NHERF2 still coprecipitated ezrin, as did full-length NHE3 (Cha et al., 2004). We selected PS120 cells for this study because they do not express exogenous NHE3, NHERF1 (minimal amount), or NHERF2 (Yun et al., 1998; Ahn et al., 2001). PS120/null cells were stably transfected with NHE3VSV-G (NHE3V) or NHE3_585VSV-G (NHE3_585V) (subsequently all NHE3 constructs are listed without the VSV-G designation to simplify the nomenclature). Immunoprecipitation was performed with monoclonal anti-VSV-G antibody and with anti-HA antibody as a negative control. Ezrin was coprecipitated by anti-VSV-G antibody from both NHE3 and NHE3_585 cells but not with the negative control HA antibody (Figure 1) or from nontransfected PS120 cells (our unpublished data). Thus, ezrin also associates with NHE3 independently of NHERF binding. The percentage of total ezrin coprecipitated was 1.5% for NHE3 and 1.3% for NHE3_585.

View larger version (36K): [in this window] [in a new window]   Figure 1. NHE3 and NHE3 truncated to aa 585 (NHE3_585) coimmunoprecipitate ezrin in PS120 cells. Immunoprecipitation was performed with anti-HA monoclonal antibody (lane 3) as a negative control and monoclonal anti-VSV-G antibody from both PS120/NHE3 (lane 4) and PS120/NHE3_585 (lane 5) cell lysates (2 mg) (the VSV-G designation of each NHE3 cell line is omitted to simplify terminology). Immune complexes were separated by 10% SDS-PAGE and transferred to nitrocellulose membranes. The left two lanes are 40 µg of total lysate of PS120/NHE3 and PS120/NHE3_585 cells. Immunoblot above was with anti-ezrin antibody and below with anti-VSV-G. Ezrin was coimmunoprecipitated with both NHE3 and NHE3_585.

View larger version (43K): [in this window] [in a new window]   Figure 2. Cytoplasmic domain of NHE3 binds to the N-terminal domain of ezrin. (A) Pull-down assays of NHE3 C terminus (CT) by ezrin domain fusion proteins. GST fusion proteins of full-length ezrin (WT, aa 1-585), N-ezrin (aa 1-309), C-ezrin (aa 310585), and GST were incubated with 35S-labeled NHE3-CT (aa 475-832). Ponceau S staining of fusion proteins is shown above. Pulled down NHE3-CT detected by autoradiography is shown in the bottom panel. Input of [35S]NHE3-CT probe is shown in Figure 2C bottom, right lane. Only N-terminal ezrin pulled down the NHE3 C terminus. (B) Overlay assays with a probe of in vitro translation 35S-labeled NHE3-CT (475-832) and ezrin domain fusion proteins. GST-full-length ezrin, GST-N-ezrin, GST-C-ezrin, and GST were transferred to the nitrocellulose membrane and overlaid with [35S]NHE3-CT (475-832). NHE3 C-terminus only bound N-ezrin. (C) N-terminal ezrin binds to the first 115 amino acids of the juxtamembrane region of NHE3 C terminus (F1 fragment [aa 475-589]) based on pull-down assays. In vitro translation 35S-labeled 6xHis-tagged NHE3 C-terminal fragment F1 (aa 475-589), F2 (aa 590-667), F3 (aa 668-747), F4 (aa 748-832), and FFull CT (aa 475-832) were used for pull-down assays with GST-N-terminal ezrin (aa 1-310) (shown in middle panel identified with anti-ezrin antibody). The pull-down with glutathione beads detected by autoradiography is shown in the top panel (14% gel), illustrating that only full-length NHE3 C terminus and the F1 fragment bound N-ezrin. The input in vitro probes for F1–F4 and full-length C terminus are shown in the bottom panel, separated on a 10% gel.

FERM Subdomain III of N-Terminal Ezrin Binds to NHE3 C Terminus
The ezrin FERM domain is composed of three structural modules FERM-I (residues 4-82), FERM-II (residues 96-195), and FERM-III (residues 204-297) that together form a compact clover-shaped structure (Hamada et al., 2000, 2003). We tested which region of the N-terminal ezrin was needed to bind NHE3. GST fusion proteins of N-terminal ezrin made up of three domains, called NE1 (aa 1-97), NE2 (aa 1-198), and NE3 (aa 1-310), were used for in vitro pull-down binding assays with in vitro translation products of [³⁵S]NHE3 full-length C terminus (475-832). Figure 3A shows that the NHE3 C terminus bound only to NE3 and not to NE1 or NE2. This result shows that the aa 199-310 of N-terminal ezrin is necessary for NHE3 C-terminus binding, which is approximately equivalent to the entire ezrin FERM subdomain III (aa 204-297).

View larger version (31K): [in this window] [in a new window]   Figure 3. FERM III subdomain of ezrin is necessary to bind to the cytoplasmic domain of NHE3 and NHERF1/2. (A) GST-fusion proteins of N-terminal ezrin NE1 (1-97), NE2 (1-198), and NE3 (1-310) and GST were used in pull-down assays with in vitro-translated 35S-labeled 6xHis-tagged NHE3-CT (475-832). Top, input GST-fusion protein expression by Ponceau S staining. Bottom, pull-down results by autoradiography in which only NE3 ezrin bound NHE3. (B) Both NHERF1 and NHERF2 bind to FERM subdomain III (199-310) of ezrin. GST-fusion proteins of N-terminal ezrin NE1 (1-97), NE2 (1-198), and NE3 (1-310) and GST were used in pull-down binding assays with 6xHis-tagged fusion proteins of NHERF1 (B1) and NHERF2 (B2). Pulled down complexes were separated by 14% SDS-PAGE, transferred to nitrocellulose membranes, and stained with Ponceau S (top). Western analyses identified NHERF1 (-5199) (B3) and NHERF2 (-2570) (B4). Both bound only to NE3 ezrin. No positive controls were used to show NE2 interactions, as reported previously (Lozupone et al., 2004).

NHE3 Juxtamembrane C-Terminal F1 Domain Has a Positive Amino Acid Cluster (K516, R520, and R527) on One Side of a Putative -Helical Area
We next determined where ezrin bound in the F1 fragment of the C-terminal domain of NHE3. Positively charged amino acid clusters in the juxtamembrane cytoplasmic domain of CD43, CD44, intercellular adhesion molecule (ICAM)-1/2, and NHE1 are involved in ERM binding (Heiska et al., 1998; Yonemura et al., 1998; Denker and Barber, 2002). By secondary structure predictions, we found that the NHE3 F1 juxtamembrane region contains an -helical region (aa 502-543) (Gene Runner, Hastings Software, Hastings, NY; Chou-Fasman analysis) (Figure 4A). We examined aa constituents of serial putative -helical regions starting at each aa between 502 and 527. Shown in Figure 4B, a positive amino acid cluster was present on one side of a putative -helix which started at aa 511 (K516, R520, and R527). This putative -helix (aa 511-528) had an estimated isoelectric point of 12.1. In Figure 4C, a comparison of the sequences of this juxtamembrane region in NHE1-5 is shown, including comparison among species. This positively charged aa cluster (indicated by boxed aa) is preserved among human, rat, and rabbit NHE3, but all three aa are not present in other NHE isoforms (NHE1, -2, -4, and -5). Another positively charged aa residue cluster in the NHE3 C terminus is located in the F2 fragment (aa 590-667). This domain is not predicted to be -helical. This positive aa cluster includes R604, R605, and R606 (indicated by double underlines in Figure 4C). This cluster also is conserved among human, rabbit, and rat NHE3 but not in other plasma membrane NHE isoforms.

View larger version (49K): [in this window] [in a new window]   Figure 4. NHE3F1 domain has a positively charged amino acid cluster (K516, R520, and R527) on one side of an -helix (aa 511-528). (A) Secondary structure predictions of NHE3 F1 domain (Gene Runner; Chou-Fasman analysis). The C-terminal F1 domain of NHE3 contains a long -helical rich domain, has some turns, but no -sheet structure (thick lines are predicted regions of stated domains). (B) Because ezrin binding is often to an -helical area and to positive amino acid clusters, this NHE3 F1 -helical domain was examined for positive amino acid clusters. K516, R520, and R527 (boxed amino acids) form a positive amino acid cluster on one side of the putative -helix between aa 511-528. The calculated isoelectric point of the putative -helix for aa 511-528 was 12.1. Isoelectric points were calculated using a program from Gene Runner software. (C) Amino acid sequence alignment in the juxtamembrane F1 and F2 regions of NHE3, comparing species and also similar domains in NHE1, -2, -4, and -5 (aa numbers refer to full-length rabbit NHE3). The conserved NHE3 F1-positive amino acids (K516, R520, and R527) are boxed, and those in NHE3 F2 are double underlined. Accession numbers for the sequences, top to bottom, are P26432, P48764, P26433, P23791, P19634, P26431, P50482, P48763, P26434, Q14940, and Q9Z0X2. (D) Nomenclature of NHE3 mutants studied, with VSV-G designation removed for simplicity. S, single; D, double; and T, triple mutation.

View larger version (44K): [in this window] [in a new window]   Figure 5. NHE3 F1 triple mutant fails to bind N-terminal ezrin, and F1 double mutant binds markedly less. (A) 6xHis-tagged 35S-labeled in vitro translation products of NHE3 F1 (aa 475-589) and NHE3 F1 triple mutant (K516Q, R520F, and R527F) were used for pull-down assays with GST-N-terminal ezrin. Right, input of [35S]NHE3 F1 fragment and F1 triple mutant fragment. Left, pull-down results that also include, as a negative control, NHE3 C-terminal His6-F4 domain fragment (left lane). (B) Lysates of PS120, PS120/NHE3, PS120/NHE3 F1 triple mutant, and PS120/NHE3 F2 triple mutant cells were used for pull-down assays with GST-N-ezrin. The NHE3 pulled down was identified with anti-VSV-G antibody. Bottom, protein expression in lysates of NHE3, NHE3 F1 triple mutant, and NHE3 F2 triple mutant were similar. The NHE3 F1 triple mutant was not pulled down by N-ezrin, whereas NHE3 and NHE3 F2 triple mutant were pulled down in similar amounts. (C) Cell lysates of PS120/NHE3, PS120/NHE3 F1S (R527F), PS120/NHE3 F1D (R520F, R527F), and PS120/NHE3 F1T (K516Q, R520F, and R527F) cells were used for immunoprecipitation with anti-HA-antibody and anti-VSV-G antibody using 1 mg of total lysate. Immunoblot was with monoclonal anti-ezrin antibody (top). Bottom, amount of NHE3 in each lysate identified by anti-VSV-G antibody.

View larger version (24K): [in this window] [in a new window]   Figure 6. NHE3 F1 ezrin binding triple mutant has minimal transport activity in PS120 cells but NHE3 F2 triple mutant has NHE3 activity slightly less than wild-type NHE3. (A) Transport activity of PS120/NHE3 (), PS120/NHE3 F2 triple mutant (), and PS120/NHE3 F1 triple mutant (). Shown is time course of Na-dependent alkalinization. All studies used mixed pairs and were repeated at least three times. Similar amounts of expression of NHE3 in all cell lines is shown in Figure 5B, bottom. (B) Transport activity of PS120/NHE3 () (Vmax = 3389 ± 194 µM/s, K'(H+)i = 0.26 ± 0.02 µM), PS120/NHE3 F1 single mutant () (Vmax = 1318 ± 36 µM/s, K'(H+) i = 0.30 ± 0.01 µM), PS120/NHE3 F1 double mutant () (Vmax = 720 ± 20 µM/s, K'(H+)i = 0.40 ± 0.01 µM), PS120/NHE3 F1 triple mutant () (too slow to calculate). The intracellular [H+] was monitored with BCECF in a spectrofluorometer and kinetic parameters were determined as described in Materials and Methods (Levine et al., 1995). A single experiment is shown. The amount of NHE3 identified by anti-VSV-G antibody is shown in insert. Mean ± SEM were calculated from at least three separate experiments. The results shown are from at least three separate clones except for NHE3 FIT, in which results were similar for six clones.

View larger version (75K): [in this window] [in a new window]   Figure 7. Cell surface expression of wild-type NHE3 and NHE3 F1 single, NHE3 F1 double, and NHE3 F1 triple ezrin binding mutants demonstrate ezrin binding dependence of surface expression of NHE3. (A) Representative Western blots of NHE3 and ezrin binding mutants were quantified with multiple dilutions of total, surface (avidin-precipitated), and intracellular (nonavidin-precipitated) fractions probed with anti VSV-G antibody and visualized with enhanced chemiluminescence. The density of each band was determined by scanning densitometer and ImageQuant software and plotted against the sample volume (microliters), shown above each lane. Fractions matched for similar densitometric values were compared in the analysis of a single Western blot. (B) Summary of surface levels (as percentage of total) of NHE3 WT and each ezrin binding mutant. Experiments were repeated at least three times for each exchanger, and results are shown as mean ± SEM. P values at least < 0.05 are in comparison with wild-type NHE3 and indicated by asterisk (*).

View larger version (46K): [in this window] [in a new window]   Figure 8. Intracellular Location of NHE3 F1 triple mutant suggests abnormal trafficking. Confocal microscopy was used to examine the localization of NHE3 F1 triple mutant compared with wild-type NHE3 in PS120 cells. As shown previously, NHE3 in PS120 cells is 15% plasma membrane and 85% intracellular, most prominently in the recycling compartment (A1). NHE3 in these cells has prominent colocalization with Rab11 (A2 and A3). In PS120 cells, NHE3 F1T has less of a plasma membrane distribution (A4). These cells have a larger Rab11 compartment than wild-type, which again colocalizes with NHE3 (A5 and A6). NHE3 and NHE3 FIT did not colocalize with a Golgi marker (B1–3 and B4–6) and the Golgi size was not changed in NHE3 FIT cells. Phospho-ezrin in NHE3 and NHE3 FIT cells localized in the plasma membrane similarly (C2 and C5). Whereas there was some NHE3 and p-ezrin overlap (C1–3), there was virtually no overlap of NHE3 FIT and p-ezrin (C4–6). A, B, and C were from cells deprived of serum for 4 h. D was from cells exposed continually to serum. In serum containing conditions, both NHE3 and p-ezrin colocalized, especially in lamellipodia (arrows) (D1–3). In NHE3 FIT cells, p-ezrin seemed similar with a prominent distribution in lamellipodia (arrows), whereas NHE3 was primarily intracellular (D4–6). Images were taken with 40 and 100x objectives. Bars, 10 µm.

Decreased Surface NHE3 in Ezrin Binding Mutants Is Due to Decreased Endocytic Recycling Plus Decreased Plasma Membrane Delivery of Newly Synthesized NHE3
To understand the mechanisms that account for the smaller percentage of direct ezrin binding NHE3 mutants on the plasma membrane, rates of NHE3 endocytosis, endocytic recycling (exocytosis) and delivery to the plasma membrane of newly synthesized NHE3 were determined. In these studies, wild-type NHE3 was compared with the NHE3 F1 double mutant, which was used rather than the triple mutant because it has a much greater plasma membrane expression, which made measurements of trafficking more reliable. Shown in Figure 9A, the amount of biotinylated NHE3 internalized in 30 min was similar or slightly reduced in NHE3 and NHE3 F1D, indicating that endocytosis over this time was not significantly altered.

View larger version (55K): [in this window] [in a new window]   Figure 9. Decreased surface NHE3 in ezrin binding mutants is due to decreased endocytic recycling plus decreased delivery of newly synthesized NHE3. (A) Endocytosis is similar in NHE3 and NHE3 F1D. Shown is total surface biotinylation from cells always kept at 4°C and not washed with GSH (first lanes from left). Ten percent of total surface was loaded on the gel. Background of cells internalized at 4°C refers to similarly treated cells always kept at 4°C but washed with glutathione buffer (second lanes). Internalized NHE3 (third lanes) is the amount of internalized biotinylated NHE3 taken up at 30 min at 37°C after removing surface biotin with glutathione buffer and subtracting results from lane 2. Biotinylated NHE3 and NHE3 F1D were collected with avidin beads and detected with anti-VSV-G antibody. These are results of a single experiment which was repeated three times with similar results. (B) Endocytic recycling (exocytosis) is decreased in NHE3 F1D. Amount of NHE3 and NHE3 F1D that newly arrived at the plasma membrane in 10 min at 37°C is shown in lane 2. 4°C (30 min) is amount of newly arrived surface NHE3 at 4°C (lane 1). Total surface is amount of biotinylated surface NHE3 in nonsulfo-NHS acetate exposed cells kept at 4°C. After 10 min at 37°C, there was much less newly arrived surface NHE3 F1D than wild-type NHE3. Results from one experiment are shown, which was repeated three times with similar results. (C) Pulse chase/surface biotinylation to determine delivery to plasma membrane of newly synthesized NHE3/NHE3 F1D demonstrated reduced delivery of NHE3 F1D. (C1) PS120 cells expressing NHE3 and NHE3 F1D were pulse labeled with 0.3 mCi/ml [35S]Met/Cys for 30 min at 37°C, followed by chase at 37°C for several times (0, 30, 60, 120, and 180 min). After completion of each time, cells were chilled to 4°C, surface biotinylated, and lysates were precipitated with streptavidin and separated by SDS-PAGE. Then, biotinylated surface[35S]NHE3 and NHE3 F1D were detected by autoradiography (top) and IB with P5D4 (bottom), respectively, on the same gel/blot and signals quantitated. (C2) Intensities from autoradiography and immunoblot were quantitated by MetaMorph software. The intensity of [35S]surface NHE3 and NHE3 F1D (from Cl, top) were normalized to the amount of surface NHE3 or NHE3 F1D for each specimen (C1, bottom). The initial 0-min intensities of NHE3 and NHE3 F1D were set at 100, and results from subsequent times were compared with zero-time results. Amount of newly synthesized NHE3 delivered to the plasma membrane was reduced in NHE3 F1D. A representative experiment is shown which was repeated twice.

View larger version (53K): [in this window] [in a new window]   Figure 10. Half-life of plasma membrane NHE3 F1 double mutant but not the total half-life is prolonged compared with wild-type NHE3 in PS120 cells. (A) Half-lives of plasma membrane NHE3 and NHE3 F1 double mutant in PS120 cells were estimated using surface biotinylation. Plasma membrane NHE3 and NHE3 F1 double mutant proteins were biotinylated at 4°C using sulfo-NHS-LC-biotin, and, after incubation at 37°C for varying times, were harvested. After solubilization, biotinylated protein was recovered with streptavidin-agarose from total cellular protein. The agarose bound biotinylated proteins were separated by SDS-PAGE and biotinylated NHE3 and NHE3 F1 double mutant were detected by anti-VSV-G antibody (P5D4) (A, a). The Western blots were quantified by densitometric analysis using ImageQuant software. Graphs (below) show quantitation of Western blots for NHE3 (antibody) and NHE3 F1 double mutant (A, c). A single representative experiment of three similar experiments is shown. (B) Half-life of total NHE3 and NHE3 F1 double mutant determined with pulse chase. PS120 cells expressing NHE3 and NHE3 F1 double mutant were pulse labeled with 0.2 mCi/ml [35S]Met/Cys cell labeling mix for 4 h and chased in medium containing Met/Cys at varying time points. After solubilization, NHE3 and NHE3 F1 double mutant were immunoprecipitated (P5D4) and subjected to SDS-PAGE. Proteins were transferred to nitrocellulose, and autoradiography was performed. Densitometric analysis was performed using ImageQuant software. Graph shows quantitation of autoradiograph for NHE3 (B, b) and for NHE3 F1 double mutant. (B, c). Total half-lives were estimated as 16.7 h for NHE3 and 15.9 h for NHE3 F1 double mutant. A representative experiment of three similar experiments is shown.

Half-Life of Plasma Membrane NHE3VF1 Double Mutant Is Prolonged
Half-life of total NHE3 and that of the NHE3 population initially on the plasma membrane were determined, using pulse-chase and surface biotinylation, respectively, as described previously (Cavet et al., 2001). Comparison was made of NHE3 with the F1 double mutant to examine a construct with significantly decreased transport activity and ezrin binding. The amount of total surface biotinylated NHE3 and NHE3 F1 double mutant proteins remaining with the cells 0-30 h after initial biotinylation of surface proteins was determined, with results normalized to surface NHE3 at time 0 (cells always at 4°C). These results (Figure 10A) demonstrated that plasma membrane NHE3 and NHE3 F1 double mutant have different degradation rates. The wild-type NHE3 in PS120/NHERF2 cells had a plasma membrane half-life of 10.4 h, which is similar to our previous data in PS120 cells (Cavet et al., 2001). In contrast, the plasma membrane half-life of NHE3 F1 double mutant was much longer (Figure 10A). Of interest, NHE3 F1 double mutant had an initial rapid decrease in amount at 2.5 h, similar to wild-type NHE3, but then had a minimal decrease over the remaining 30 h. This result indicates that the mutation of the direct ezrin binding site of NHE3 increases its plasma membrane half-life. The similar initial decrease in plasma membrane NHE3 in the F1 double mutant is consistent with the normal endocytosis rate measured over 30 min as shown in Figure 9A.

The half-lives of the total NHE3 and NHE3 F1 double mutant were measured using ³⁵S-pulse-chase labeling. Even though the surface half-life was much longer in NHE3VF1 double mutant than wild type, the half-lives of total NHE3 and NHE3 F1 double mutant were not significantly different (Figure 10B). This means that ezrin binding to NHE3 affects half-life by an effect at the plasma membrane and not intracellularly.

Lateral Mobility of NHE3 F1D at the Apical Membrane of OK Cells Is Decreased as Assessed by FRAP
Further studies were designed to examine the apical membrane mobility of the NHE3 F1D mutant in polarized epithelial cells. This was done to address the long half-life of plasma membrane FID mutant NHE3 with normal endocytosis. Moreover, we felt study of an epithelial cell BB would allow greater examination of plasma membrane effects. We have shown that the lateral mobility of NHE3 at the apical surface of OK cells is dependent on the actin cytoskeleton. For example, the lateral mobility of NHE3-EGFP was significantly decreased by 0.05 µM latrunculin B 30-min treatment (Cha et al., 2004). In this study, we compared the lateral mobility of BB NHE3 F1D-EGFP with wild-type NHE3 under conditions we had shown that trafficking did not contribute to fluorescence recovery in the BB (Cha et al., 2004). As shown in Figure 11, BB mobility of NHE3 F1D mutant was decreased compared with wild-type NHE3. This is consistent with our previous results in which NHE3 truncated not to bind NHERF proteins exhibited actin cytoskeleton-dependent BB mobility (Cha et al., 2004).

View larger version (33K): [in this window] [in a new window]   Figure 11. Lateral mobility of NHE3 F1D-EGFP in OK cell BB was significantly decreased compared with wild-type NHE3. (A) OK cells depleted in NHE3 expression by acid suicide technique were transiently transfected with NHE3-EGFP and NHE3 F1D-EGFP and studied by FRAP using a confocal microscope concentrating on the apical membrane. FRAP was measured at the apical nonjuxtanuclear region of NHE3-EGFP or NHE3 F1D-EGFP in polarized OK cells. The results shown are a representative FRAP experiment with initial fluorescence intensity before photobleaching set at 100%. Data were collected as 50 images every 9 s. Similar results were obtained in three similar experiments. (B) The mobile fractions of NHE3-EGFP and NHE3 F1D-EGFP were determined by FRAP. Data shown are means ± SEM. *P values are compared with control NHE3-EGFP (unpaired t test).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The same ezrin domain (FERM) is used in both types of binding to NHE3, which indicates the involvement of two separate ezrin molecules with a single NHE3 molecule. This concept seems to apply to ezrin interaction with multiple ligands. How ezrin interacts with its partners has recently been partially clarified. There are two patterns, direct binding and indirect binding via NHERF interactions. Substrates that only directly bind ezrin include ICAM-1, ICAM-2, ICAM-3, CD44, syndecan2, and NHE1 (Tsukita et al., 1994; Helander et al., 1996; Yonemura et al., 1998; Denker and Barber, 2002; Lozupone et al., 2004). Indirect associations have been described with the 2-adrenergic receptor, CFTR and NHE3 (Tsukita et al., 1994; Hall et al., 1998; Wang et al., 1998; Sun et al., 2000). The renal podocyte plasma membrane protein podocalyxin binds ezrin both directly and indirectly via NHERF1 or -2 (Schmieder et al., 2004). Of note, the function of the direct ezrin binding to podocalyxin has not been identified. Although podocalyxin is the protein that seems to interact with ezrin in a manner most similar to NHE3, there are other multiple ezrin-interacting proteins. Ezrin binds the phosphoinositide-3 kinase (PI-3 kinase) p85 subunit using two sites in ezrin, the N-terminal domain and pY353, which binds to the c-Src homology 2 domain of p85 (Sun et al., 2000). Also, the Drosophila ERM-related protein coracle, which is present in and helps organize the epithelial cell septate junction, requires binding to its ligands with both its N-terminal (FERM domain) and its nonactin binding C terminus for normal septate junction formation (Ward et al., 2001). The increasing recognition of more than one ezrin molecule or multiple sites on one ezrin molecule binding to a single partner is consistent with ezrin existing as a dimer (Bretscher et al., 2002; Chambers and Bretscher, 2005) and suggests this pattern should be sought to determine how frequently it is involved in ERM function.

The involvement of the ezrin N-terminal FERM domain in both NHERF binding and NHE3 direct binding is not surprising given the recognized importance of this N-terminal domain in ezrin binding to its partners, including ICAM-1-3, CD43, CD44, PIP2, calmodulin, CD95, and NHE1, among others. Multiple parts of the FERM domain are involved in binding; for example, ICAM-2 binds the third part, whereas CD95 (Fas) binds the second part of the FERM domain (Lozupone et al., 2004). Whether NHE3 binding to FERM domain 3 is specifically to these aa, however, is not known.

These studies extend understanding of some aspects of the ezrin binding sites in its ligands, which is known to involve multiple positively charged aa (Ward et al., 2001; Bretscher et al., 2002). As with NHE3 shown here, secondary structure analysis of ezrin ligand binding sites indicates that the clustered positive charges for ezrin binding can be in an -helical as well as in a linear configuration. The direct ezrin binding domain of podocalyxin involves its last 15 aa between aa 405 and 422, which by our analysis is predicted to be -helical (Schmieder et al., 2004). Thus, both secondary structures must be considered when searching for ezrin binding sites. Moreover, clustered positively charged aa are not sufficient for ezrin binding (the NHE3 F2 domain-positive aa cluster does not bind ezrin), whereas a high isoelectric point, as is present in the direct NHE3 binding area, is often necessary. Importantly, whether positively charged aa are involved in direct ezrin contact is not established. Indeed, in crystal structure to 2.4 Å of the radixin FERM domain with the cytoplasmic domain of ICAM-2, binding involves a -sheet peptide of ICAM-2. Whereas positively charged aa are involved in this binding, they flank but are not in the binding domain, and their suggested role is in interacting with membrane phosphatidylinositol bisphosphate (Hamada et al., 2003).

The dependence on direct ezrin binding (and thus implications of binding to actin) of trafficking of NHE3 to the plasma membrane (both newly synthesized and from the recycling compartment) is consistent with the previous demonstration of the actin dependence of these processes in general. The involvement of actin in exocytosis involves multiple steps, including depolymerization of actin to remove the terminal web barrier, which prevents vesicles from approaching the membrane, and a more active role of actin in which a myosin motor is involved with moving the vesicle to the membrane for docking (Valentijn et al., 1999; Bader et al., 2002, 2004; Wenk and Camilli, 2004). Moreover, two pools of vesicles, at least for the transporter GLUT4, are involved, a pool for immediate release and a reserve pool, which accounts for the majority of the transporter, with actin apparently involved in regulation of trafficking between these two pools (Rudich and Klip, 2003). Recently, multiple distinct intracellular and plasma membrane pools of NHE3 have been described (Alexander et al., 2005). Which of these roles of actin and which pool of NHE3 are involved in NHE3 exocytosis is not known. That direct ezrin binding to NHE3 seems necessary for endocytic recycling to the plasma membrane is also consistent with previous results indicating a role of ezrin in movement of PI3-kinase to the apical membrane of the renal proximal tubule cell line, LLC-PK1, because basal as well as stimulated NHE3 activity in multiple cell models (PS120 and OK cells) are PI 3-kinase dependent (Gautreau et al., 1999). These studies, however, do not negate that the indirect NHE3/ezrin/actin association via NHERF1/NHERF2 binding may also be necessary for some aspect(s) of trafficking, particularly those related to NHE3/NHERF family complex formation and regulation of NHE3 activity (Donowitz et al., 2005). The lack of involvement of the direct ezrin binding domain of NHE3 in endocytosis (Figure 9A) is consistent with the previous demonstration that the domain of the NHE3 C terminus necessary for cAMP inhibition, part of which occurs via increased rates of endocytosis, was C-terminal of the direct ezrin binding domain (Cabado et al., 1996). Relevant to the aggregate studies in Figures 9 and 10, the long plasma membrane half-life of the NHE3 F1 double mutant may represent unmasking of the recently recognized nontrafficking apical membrane pool of NHE3 (Alexander et al., 2005).

Relating to other functional consequences of NHE3 direct ezrin binding, we suggest that the reduced mobility of the BB pool of the NHE3 F1 double ezrin binding mutant and its prolonged plasma membrane half-life may be related. Both these changes are due to changes in the plasma membrane pool of NHE3. There are at least two parts of the half-life curve of surface NHE3 F1 double mutant (Figure 10A), with an initial normal rate of decrease (consistent with a normal endocytosis rate) and a much reduced later phase. We hypothesize these results are best explained by a normal basal endocytic pool of NHE3, representing that initially in the intervillus cleft area. In addition, however, there may be a defect in the ability to deliver NHE3 from the plasma membrane to that pool (Figure 12) This would predict that there is likely to be a defect in stimulated endocytosis in the NHE3 direct ezrin binding mutant.

View larger version (23K): [in this window] [in a new window]   Figure 12. Model of effects of direct ezrin binding to NHE3 on trafficking in an epithelial dell. Delivery of NHE3 to the plasma membrane is reduced in the absence of direct ezrin binding both by the effects on direct delivery from the synthetic pathway and from endocytic recycling. Although the endocytic process is normal, delivery of NHE3 from the microvillus to the endocytic clefts (indicated by reduced lateral mobility of microvillar NHE3 shown in Figure 11) may also be dependent on direct ezrin binding.

We have shown that direct binding of NHE3 to ezrin is necessary for NHE3 trafficking and mobility in the BB (Figure 12). However, direct binding of NHE3/ezrin did not seem to affect organization of the membrane cytoskeleton, either in PS120 cells or OK cells as judged by the presence of microvilli. Although NHE1/direct ezrin binding has been shown to be important for organizing the cytoskeleton (Denker et al., 2000; Denker and Barber, 2002; Baumgartner et al., 2004), there is no evidence yet for a role of NHE3/direct ezrin binding in similar organization either of the lamellipodia of fibroblasts (Figure 8D) or of the epithelial cell apical cytoskeleton, although this has not been systematically examined.

	ACKNOWLEDGMENTS

 
We acknowledge the expert editorial assistance of H. McCann. This study was supported in part by National Institutes of Health National Institute of Diabetes and Digestive and Kidney Diseases Grants R01-DK26523, R01-DK61765, and P01DK44484; R24-DK64388 (The Hopkins Basic Research Digestive Diseases Development Core Center); and The Hopkins Center for Epithelial Disorders.

	Footnotes

 
This was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E05-09-0843) on March 15, 2006.