Polarity and Developmental Regulation of Two PDZ Proteins in the Retinal Pigment Epithelium

Abstract

purpose. Identification of binding partners for ezrin, an actin-binding protein crucial for morphogenesis of apical microvilli and basolateral infoldings in RPE cells.

methods. Rat eyes, rat primary RPE, the rat RPE-J cell line, and a clonal line of RPE-J cells transfected with human ezrin cDNA were analyzed by immunofluorescence microscopy and immunoblot. Immunofluorescence localization of two ezrin-binding proteins was performed in cryosections of rat eyes of various ages and in monolayers extracted with the detergent Triton X-100 and fixed in paraformaldehyde. The interaction of both proteins with ezrin and gluthathione-S-transferase (GST)-ezrin fusion proteins was analyzed by SDS-PAGE and immunoblot.

results. Immunofluorescence microscopy of adult rat eyes detected a polarized distribution of ERM (ezrin, radixin, and moesin)-binding phosphoprotein of 50 kDa (EBP50) at the apical microvilli and synaptic–associated protein of 97 kDa (SAP97) at the basolateral surface of RPE cells, which overlapped with ezrin. These two PDZ (postsynaptic density protein [PSD-95]/disc large [DLG]-A/ZO-1) domain proteins had a similar polarized distribution and high resistance to detergent extractability, indicative of cytoskeletal association, both in primary cultures of rat RPE and in a clonal RPE-J cell line expressing high levels of transfected ezrin. RPE cell lysates from rat retinas of various postnatal ages revealed increasing levels of EBP50 and SAP97 compared with αv integrin, a protein expressed at constant adult levels from birth. GST pull-down and immunoprecipitation experiments demonstrated a direct interaction between EBP50 and SAP97 and ezrin.

conclusions. The data indicate that EBP50 localizes at the apical microvilli, whereas SAP97 localizes at the basolateral surface of RPE cells, probably through a direct interaction with ezrin.

The RPE performs a variety of vectorial transport and secretion functions that are essential for photoreceptor survival. The RPE transporters and structural proteins responsible for these polarized functions are located asymmetrically into either the very long apical microvilli that surround the photoreceptors or the basolateral membrane, that displays intricate basal infoldings abutting on Bruch’s membrane. We have recently shown that the normal development of both microvilli and basal infoldings during postnatal maturation of the rat RPE is dependent on the actin-binding protein ezrin, ¹ a member of the closely interrelated ezrin, radixin, and moesin (ERM) protein family. Ezrin is the first and most studied protein of the ERM family, which is included within the band 4.1 superfamily. The NH₂-terminal domain of ERM proteins, designated as FERM, ^{2 3 4} is highly conserved (more than 85% identity) and is followed by a long central α-helical region and a charged COOH-terminal domain. The FERM domain interacts directly with plasma membrane proteins, such as the hyaluronate receptor CD44, ⁵ CD43, ^{6 7} intercellular adhesion molecule (ICAM)-1, ⁸ ICAM-2, ^{8 9} and ICAM-3, ¹⁰ as well as with PDZ-domain–containing proteins (see later description). The COOH-terminal and the NH₂-terminal domains of ERM proteins bind F-actin, ^{11 12 13 14 15} thus allowing these proteins to link the plasma membrane with the actin cytoskeleton. ^{4 16 17}  

Recently, it has become apparent that ERM proteins may interact with multispanning plasma membrane proteins through adaptor proteins containing PDZ-domains, protein-recognition motifs named after the proteins where these domains were first observed (postsynaptic density protein [PSD]-95, the Drosophila tumor suppressor DLG[ disc large]-A, and the tight junction protein ZO1). PDZ domains are generally used to cluster several copies of a transmembrane protein into membrane subdomains or to assemble groups of signaling proteins into a functional signaling cascade. ^{18 19} Two polarized PDZ domain proteins that interact with ezrin have been identified and characterized in epithelia other than RPE. Ezrin binding protein of 50 kDa (EBP50) has two PDZ domains and localizes to the apical microvilli of some native epithelia and cultured cells. ^{20 21 22} EBP50 links apical transporters such as the cystic fibrosis transmembrane conductance regulator (CFTR), ²¹ the kidney proximal tubule Na⁺/H⁺ exchanger (NHE3), ^{23 24} and the β₂-adrenergic receptor ^{25 26} to ezrin and the actin cytoskeleton. ²⁷ Synaptic–associated protein of 97 kDa (SAP97), the rat homologue of the Drosophila tumor suppressor disc large (Dlg), has three PDZ domains, and was identified as an interacting partner of ezrin at the basolateral surface of gastric parietal cells. ²⁸ The human homologue of SAP97, hDlg, was shown to bind ezrin and band 4.1. ^{29 30 31 32} SAP97 interacts with Shaker-type K⁺ channels, ^{33 34} the N-methyl-d-aspartate (NMDA) receptor subunits, ³⁵ and the GluR1 subunit of theα -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor ³⁶ and localizes to postsynaptic complexes of neurons and to the lateral membrane of different epithelial cells. ^{31 37 38 39}  

Having identified ezrin as a major player in RPE morphogenesis, it is important to identify ezrin-binding partners that cooperate with ezrin in such a process. Herein, we report for the first time the presence of EBP50 and SAP97 in RPE, localized respectively at the microvilli and basal infoldings. Immunoblot analysis of these proteins during the first two postnatal weeks in the rat detected a 10- and 4-fold increase in their expression, that correlated with the development of microvilli and basal infoldings. Our results are consistent with an important role of these two PDZ-domain–containing proteins in RPE morphogenesis. 

Methods

Antibodies

Affinity-purified polyclonal antibody against EBP50 was the generous gift of Anthony Bretscher (Cornell University, Ithaca, NY). Polyclonal antibody directed to SAP97/hDlg and monoclonal antibody directed to ezrin were purchased from Affinity Bioreagents, Inc. (Golden, CO) and NeoMarkers Inc. (Fremont, CA), respectively. Polyclonal antibodies directed to the integrin αv and to ezrin (C-15) were purchased from Chemicon International, Inc. (Temecula, CA) and Santa Cruz Biotechnology Inc. (Santa Cruz, CA), respectively. Secondary antibodies to mouse IgG and rabbit IgG, conjugated to either FITC or CY3, were purchased from Cappel Laboratories (Cochranville, PA) and Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). 

Cultured Cells

RPE-J cells were cultured at the permissive temperature of 32°C as previously described. ⁴⁰ Briefly, cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 4% heat-inactivated (30 minutes at 56°C), fetal calf serum (CELLect Gold; ICN Pharmaceuticals, Inc., Costa Mesa, CA), glutamine, nonessential amino acids, and penicillin-streptomycin (Gibco, Grand Island, NY). To obtain a differentiated epithelial phenotype, cells were plated on polycarbonate Transwell filters (Costar, Cambridge, MA) coated with a thin layer of synthetic basement membrane (Matrigel; Collaborative Research, Bedford, MA), cultured in growth medium supplemented with 10⁻⁸ M retinoic acid for 6 to 7 days, and then switched to the nonpermissive temperature of 40°C for 36 to 48 hours. ⁴⁰  

Primary RPE cultures were obtained from 2- to 3-week-old Long-Evans rats (Harlan Sprague-Dawley, Indianapolis, IN), as previously described. ¹ Briefly, animals were killed by CO₂ asphyxiation, and the eyes were enucleated and stored in 10 mM HEPES buffered Hanks’ balanced salt solution (HBSS). A circumferential incision was made above the ora serrata and the cornea, lens, iris, and vitreous body were removed. The eyecups with the neural retina exposed were incubated in 320 U/ml hyaluronidase in HBSS for 1 hour at 37°C, the neural retina was peeled off from the RPE, and the eyecups were incubated in 2 mg/ml trypsin in HBSS for 60 minutes at 37°C. RPE sheets were teased from the underlying choroid with needles, collected, and incubated with trypsin-EDTA for 1 minute. Cells were plated on synthetic membrane (Matrigel; Collaborative Research)–coated filters (Transwell) and cultured without further passaging in DMEM supplemented with 10% FCS, l-glutamine, nonessential amino acids, and antibiotics for 2 to 4 weeks. All animal procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Cryosections

Long-Evans rats anesthetized with cold-induced hypothermia (pups) or CO₂ (adults) were perfused intracardiacally with 4% paraformaldehyde in PBS. The eyes were enucleated and postfixed by immersion in the same fixative for 3 hours at 4°C. A circumferential incision was made bellow the ora serrata, and the cornea, lens, iris, and vitreous body were removed. Eyecups were quenched with 50 mM NH₄Cl made in 0.1 M modified PHEM buffer (pH 6.9; 60 mM PIPES, 20 mM HEPES, 10 mM EGTA, 5 mM MgCl₂, 70 mM KCl) for 1 hour at 4°C, infused successively with 15% and 30% sucrose made in the same buffer and with optimal cutting temperature compound (Tissue-Tek 4583; Miles Inc., Elkhart, IN). Alternatively, eyecups were treated with hyaluronidase to detach the neural retina from the RPE as described earlier. Cryosections (10 μm) were cut on a cryostat (Bright Instrument Co., Huntingdon, UK), and collected on slides (SuperFrost; Fisher, Springfield, NJ). Sections were allowed to dry for at least 1 hour at room temperature, washed three times in PBS, 0.3 mM CaCl₂, 1 mM MgCl₂, and 0.2% BSA (PBS-CM-BSA) to remove the mounting medium. Retina-free eyecups were treated with 0.2% Triton X-100 made in PBS-CM-BSA for 15 minutes before blocking and labeling the sample. 

Immunofluorescence

To detect the association of proteins with the cytoskeleton, monolayers were extracted with a detergent-containing buffer and double-stained with antibodies against EBP50 or SAP97 and ezrin as follows. Monolayers were washed in PBS-CM and dipped four times for 5 seconds at room temperature in four different beakers containing extraction buffer MES (50 mM MES, 3 mM EGTA, 5 mM MgCl₂, 0.5% Triton X-100 [pH 6.4]), as previously described, ^{1 41} fixed in 4% paraformaldehyde, and processed for immunofluorescence. Filters and tissues were incubated for more than 1 hour in the appropriate primary antibody (or antibodies) diluted in PBS-CM-BSA. After several washes, filters and tissues were reacted with the appropriate secondary antibodies conjugated to FITC, Texas red, or CY3. In some cases, cell nuclei were stained with 1 μg/ml 4′,6-diamidino-2-phenyllindole (DAPI) in PBS-CM for 5 minutes. A series of 0.5 μm xy (en face) or single xz (transverse) sections were collected in a laser scanning confocal microscope (LSCM; LSM510; Carl Zeiss, Oberkochen, Germany). Confocal images were further analyzed with imaging software (MetaMorph; Universal Imaging, West Chester, PA). Each individual xy image of the RPE monolayers stained with EBP50 and ezrin antibodies represents a three-dimensional projection of the entire monolayer (sum of all images in the stack), but each individual xy image of the RPE monolayers stained with SAP97 and ezrin antibodies had the top 2 μm of the stack eliminated in both channels to remove the strong apical ezrin staining, and then the rest of the planes were summed up into one image. Alternatively, cryosections were analyzed using an epifluorescence microscope (model E600; Nikon, Melville, NY) and digital images were collected with a cooled charge-coupled device (CCD) camera and the software. Figure panels were composed with image analysis software (Photoshop ver. 5.0; Adobe, San Jose, CA). 

Immunoelectron Microscopy

Adult rat eyecups freshly isolated were fixed in 4% paraformaldehyde, 0.1% glutaraldehyde, 0.2% picric acid in 0.1 M modified PHEM buffer (pH 6.9; 60 mM PIPES, 20 mM HEPES, 10 mM EGTA, 5 mM MgCl₂, 70 mM KCl) for 1 hour at 4°C. Samples were postfixed in 0.25% tannic acid for 1 hour at 4°C, dehydrated in ethanol, embedded in resin (Unicryl Kit; Ted Pella Inc., Redding, CA) and polymerized under UV light for 96-hour at −20°C. Ultrathin sections on grids were hydrated at room temperature for 1 hour with PBS, 1% BSA, 0.01% Tween-20, 10% goat serum, incubated in 50 mM NH₄Cl in PBS-CM for 30 minutes, followed by sequential incubation with a rabbit anti-SAP97 antibody (1:50) and 10 nm colloidal gold-conjugated antibody (1:50) (BBInternational; Ted Pella Inc.) for 1 hour each. Samples were examined under a microscope (model 100CX-II; JEOL, Peabody, MA) using an accelerating voltage of 80 kV. 

Quantification of EBP50 and SAP97 in Cell Lysates

Whole cell lysates from RPE-J monolayers, RPE collected from rat eyecups and primary cultures of RPE were solubilized in RIPA buffer (0.1% SDS, 1% Triton X-100, 1% deoxycholate, 0.15 M NaCl, 2 mM EDTA, 25 mM Tris [pH 7.4]) supplemented with a cocktail of protease and phosphatase inhibitors (Sigma, St. Louis, MO). Protein per sample (20μ g) was resolved in a 7.5% SDS-PAGE and electrotransferred to nitrocellulose membranes (Schleicher & Schuell, Keene, NH). Membranes were incubated with antibodies to EBP50 and SAP97 in buffer (Blotto A; 20 mM Tris/HCl, 0.9% NaCl, 0.05% Tween-20 [TBST], 5% skimmed milk) for 1 hour. Alternatively, lysates were run under nonreducing conditions electrotransferred and probed for the integrin αv, as previously described. ⁴² Protein detection was performed with secondary antibodies conjugated to peroxidase and visualized using chemiluminescence reagent (Reagent Plus; NEN Life Science Products, Inc., Boston, MA) detection system. Nitrocellulose membranes were exposed to film, and the intensity of scanned protein bands was analyzed with NIH Image software (ver. 1.62; provided in the public domain by the National Institutes of Health, Bethesda, MD, and available at http://www.nb.nih.ncbi.gov). 

Glutathione-S-Transferase Pull-Down Experiments

The plasmids pGEX-EzN and pGEX-Ez, encoding the NH₂-terminal domain and the full-length human ezrin fused to glutathione-S-transferase (GST), the generous gift of Monique Arpin (Curie Institute, Paris, France), have been described. ⁴³ GST-ezrin and GST/NH₂-terminal ezrin fusion proteins were expressed in Escherichia coli strains TG1 and DH5α, respectively, and affinity purified using glutathione-Sepharose beads (GSH-Sepharose; Amersham Pharmacia Biotech, Uppsala, Sweden). Bacteria were grown overnight at 37°C in Luria-Bertani (LB) medium containing 100 μg/ml ampicillin. The overnight cultures were diluted 1:20 in LB medium with ampicillin and induced with 0.5 mM IPTG for 60 minutes (30 minutes for pGEX-EzN). Bacteria were harvested by centrifugation at 4000g for 15 minutes; the pellet was resuspended in 1 ml ice-cold PBS supplemented with 1.5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF) and protease and phosphatase inhibitors and sonicated for 2 minutes. Lysates were centrifuged at 14,000 rpm for 20 minutes at 4°C. The supernatant was rotated overnight at 4°C with 30 μl of a 50% slurry of the GSH-Sepharose beads, prepared according to the manufacturer’s instructions. Beads were thoroughly washed with ice cold PBS. GST, GST-ezrin, and GST-NH₂ ezrin beads were preincubated with lysis buffer supplemented with 4% BSA for 1 hour at 4°C. An RPE-J clone overexpressing exogenous ezrin (clone 16, described in Ref. ¹ was lysed by sonication in hypotonic lysis buffer (10 mM Tris-HCl [pH 7.5], 0.5 mM MgCl₂, 1 mM EGTA, and 1 mM PMSF) supplemented with protease and phosphatase inhibitors as previously described. ⁴³ For SAP97 binding assays, RPE monolayers were treated with 5 mM of the amine-reactive cross-linking reagent dithiobis[succinimidyl propionate] (DSP; Pierce, Rockford, IL) for 30 minutes at 4°C before lysis. Debris and nuclei were pelleted by centrifugation at 700g for 5 minutes at 4°C. Lysates were supplemented with 0.5% Triton X-100, 140 mM NaCl, and 0.8% BSA and incubated with GST beads overnight at 4°C. Beads were washed with the same ice-cold lysis buffer modified with 0.1% BSA and 0.01% Triton X-100. Bound proteins were eluted from the beads by boiling in sample buffer for 5 minutes at 100°C. Proteins were resolved in 7.5% SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted with EBP50 and SAP97 antibodies, as described earlier. 

Immunoprecipitation of EBP50 and SAP97 from Adult RPE Sheets

Long-Evans rats, 2 to 3 weeks old, were anesthetized by exposure to CO₂, the eyes were enucleated, and the cornea, lens, iris, and vitreous body were removed through a circumferential incision made below the ora serrata. Eyecups were treated with hyaluronidase to detach the neural retina from the RPE, and with trypsin to allow the teasing of RPE sheets from the eyecups, as described earlier. RPE sheets were washed with ice-cold TBS and lysed for 30 minutes at 4°C in 500 μl of either RIPA buffer (150 mM NaCl, 25 mM Tris [pH 7.4], 2 mM EDTA, 1% Triton X-100, 1% deoxycholate, 0.1% SDS, and 1 mM PMSF, with protein and phosphatase inhibitor cocktails), for EBP50, or 500 μl lysis buffer (TBS with 1% NP-40, 10% glycerol, and 2 mM EDTA, with protein and phosphatase-inhibitor cocktails), for SAP97 immunoprecipitation. Lysates were clarified by spinning at 16,000g for 15 minutes to obtain a postnuclear supernatant (PNS) that was used for immunoprecipitation. Supernatants were transferred to new tubes and incubated with 25 μl pansorbin (Calbiochem, La Jolla, CA) prewashed three times with RIPA or lysis buffer and then rotated for 30 minutes at 4°C. Precleared supernatants were centrifuged, transferred to new tubes, and incubated with protein-Sepharose beads (A/G PLUS; Santa Cruz Biotechnology) coupled to anti-EBP50 and SAP97 antibodies. Samples were rotated in the cold for at least 60 minutes, centrifuged, and washed five times with RIPA or lysis buffer. Bound proteins were eluted by incubation with two times sample buffer and boiling of samples for 5 minutes. Immunoprecipitated proteins were resolved in a 7.5% SDS gel and transferred to nitrocellulose membranes. Membranes were then reacted with a goat antibody specific to ezrin. 

Results

Expression of EBP50 and SAP97 in RPE Cells In Situ

Our laboratory has previously shown that inhibition of ezrin synthesis by antisense oligonucleotides results in the disappearance of apical microvilli and basal infoldings of primary RPE cells. ¹ To gain further understanding of the ezrin-dependent mechanisms responsible for the morphogenesis of surface differentiations, we investigated the expression of the PDZ-containing proteins EBP50 and SAP97 in immature as well as in fully developed rat RPE. Immunofluorescence of rat eyecup cryosections detected EBP50 at the apical surface of both immature (P0; Fig. 1A ) and adult (Fig. 1D) RPE, overlapping with the distribution of ezrin, (Figs. 1B 1E) , as demonstrated by the yellow color in digitally merged images (Figs. 1C 1F) . On RPE maturation, apical EBP50 and ezrin immunofluorescence became more intense and extended into the outer segment (OS) layer, consistent with the ensheathing of the growing OS by the elongating RPE microvilli. In contrast, indirect immunofluorescence with SAP97 antibody of retinal cryosections at P0 (Fig. 1G) and adulthood (Fig. 1J) detected the protein in association both with the lateral (Fig. 1G , arrows) and basal surface of RPE. Specific SAP immunofluorescence was observed in the choroid of immature eyes (Fig. 1G) , but the significance of this observation remains obscure. Adult neural retina-free eyecups labeled with SAP97 displayed a clear lateral and basal localization (Fig. 1J) that colocalized with the ezrin staining (Fig. 1K) , as demonstrated by the yellow areas in digitally merged images (Fig. 1L) . The comparison of ezrin with EBP50 and SAP97 labeling suggests that, both in immature and adult RPE, ezrin and EBP50 primarily colocalize in apical microvilli, whereas ezrin and SAP97 colocalize at the basolateral plasma membrane of both immature and mature RPE cells. 

Localization of SAP97 to Basal Infoldings of Mature RPE In Situ

To confirm the presence of SAP97 in the basal infoldings of RPE, we performed immunogold electron microscopy localization experiments on ultrathin sections of adult eyecups embedded in resin (Unicryl; Ted Pella, Inc.). We have previously demonstrated the presence of ezrin at both apical microvilli and basal infoldings by using the same procedure. ¹ Unicryl sections were sequentially labeled with a polyclonal antibody to SAP97 followed by a secondary donkey anti-rabbit IgG antibody conjugated to 10 nm colloidal gold particles. Control samples labeled with secondary antibody alone showed no specific labeling of the samples (Fig. 2A) . In contrast samples labeled with SAP97 antibody showed colloidal gold particles specifically labeling the basal infoldings of RPE (Fig. 2B , arrowheads). 

EBP50 and SAP97 Expression in Developing RPE

If EBP50 and SAP97 play a structural role in the morphogenesis of apical and basal differentiations of RPE, it would be expected that, as previously shown for ezrin, ¹ their expression would dramatically increase during postnatal maturation of the rat retina. Immunoblots of lysates of RPE cells obtained from rats of different postnatal ages demonstrated a 10- and 4-fold increase in the expression levels of EBP50 and SAP97 between P3 and P21 (Figs. 3A 3B , respectively), when normalized to the expression of the integrinα v, a protein previously shown to remain unalterted during this time (Fig. 3C) . ⁴²  

Colocalization of EBP50 with Ezrin in the Apical Surface and of SAP97 with Ezrin at the Lateral Membrane of RPE Cells

To further analyze the interaction of EBP50 and SAP97 with ezrin in RPE cells we studied the distribution of these proteins in primary rat RPE cultures, in the rat RPE-J cell line, and in a clonal RPE-J cell line overexpressing human ezrin (clone 16). All these RPE cultures preserve native RPE characteristics, ^{40 44} including the ability to perform phagocytosis. ^{42 45} Whereas primary cultures show extensive apical microvilli and basal infoldings, RPE-J cells possess sparse and short microvilli and no basal infoldings. ¹ However, the RPE-J clone 16 significantly increased the number and extension of surface differentiations after ezrin overexpression. To characterize the association of EBP50 (Fig. 4) and SAP97 (Fig. 5) with the cytoskeleton we examined by confocal microscopy RPE monolayers extracted with Triton X-100 before fixation with paraformaldehyde. En face examination of primary RPE monolayers grown on polycarbonate filters revealed a punctate staining pattern for both ezrin (Fig. 4A) and EBP50 (Fig. 4B) consistent with plasma membrane association. xz sections confirmed the association of both ezrin (Fig. 4C) and EBP50 (Fig. 4D) with apical microvilli-like structures. In contrast, wild-type RPE-J had low levels of ezrin remaining after detergent extraction (Figs. 4E 4G) , and low levels of EBP50 associated with the plasma membrane (Figs. 4F 4H) . Of interest, ezrin overexpression in RPE-J clone 16 largely increased the levels of detergent-resistant ezrin and EBP50. The two proteins displayed a punctate pattern en face (Figs. 4I 4J) and an apical membrane distribution in xz optical sections (Figs. 4K 4L) . 

A similar approach was used to compare the distribution and cytoskeletal association of ezrin and SAP97 in cultured RPE cells. Monolayers of cells plated on filters were extracted with Triton X-100 followed by fixation, staining with specific antibodies, and observation under confocal microscope (Fig. 5) . To facilitate the visualization of these proteins in the lateral membrane, the strong apical fluorescence of ezrin in primary RPE cultures and RPE-J clone 16 was eliminated by deleting the top four confocal sections (corresponding to 2 μm) and summing up the remaining sections of each channel for en face examination. Under these conditions, lateral ezrin and SAP97 signals were easily detected in long-term (8 weeks) primary RPE cultures (Figs. 5A 5B , respectively), and in RPE-J clone 16 overexpressing ezrin (Figs. 5I 5J , respectively), but were absent or very weak in wild-type RPE-J cells (Figs. 5E 5F , respectively). Vertical optical sections of the whole monolayer displayed a strong apical ezrin signal and comparatively low levels of lateral ezrin, and a preferentially lateral SAP97 localization in primary RPE cultures (Figs. 5C 5D , respectively) as well as in RPE-J clone 16 cells (Figs. 5K 5L) . Furthermore, SAP97 could be detected in the basal surface of some of the cells of the monolayers of rat primary RPE, RPE-J, and clone 16. The significance of this observation is not clear. As shown above for ezrin and EBP50, a similar treatment of RPE-J cells extracted most of the cellular ezrin (Fig. 5E) and resulted in very a weak signal of SAP97 on the basolateral membrane (Figs. 5F 5H) . 

Confirmation of EBP50 and SAP97 Interaction with Ezrin

To further characterize the interaction of ezrin with EBP50 and SAP97, GST-Sepharose beads linked to either full-length ezrin or to its NH₂-terminal domain were mixed with lysates of RPE-J clone 16 cells. Bound proteins were eluted from the beads, resolved in a 7.5% SDS-PAGE gels, transferred to nitrocellulose membranes, and immunoblotted with the antibodies to EBP50 (Fig. 6A) or SAP97 (Fig. 6B) . Under these conditions, EBP50 bound to the NH₂-terminal domain of ezrin (Fig. 6A , lane 2) but not to full-length ezrin (Fig. 6A , lane 3) or to control GST beads (Fig. 6A , lane 1). These results support previous data that suggest that inactive ezrin exists in a closed configuration that requires activation to allow the unfolding and interaction of its NH₂-terminal domain with EBP50. ⁴⁶ Conversely, SAP97 bound more strongly to full-length ezrin (Fig. 6B , lane 3) than to ezrin’s NH₂-terminal domain (Fig. 6B , lane 2) and did not bind to GST beads (Fig. 6B , lane 1). It should be noted that the detection of SAP97-ezrin interactions required stabilization of lysates by a cross-linker (DSP). 

To further characterize the interaction of ezrin with EBP50 and SAP97, immunoprecipitation experiments were performed on lysates of RPE sheets collected from adult rat eyes. Samples were lysed, precleared, and incubated with protein-Sepharose beads (A/G Plus; Santa Cruz Biotechnology) coupled to anti-EBP50 and SAP97 antibodies. Immunoprecipitated proteins were resolved in an SDS-PAGE gel, and transferred to nitrocellulose membranes. The membranes were then incubated with an antibody specific to ezrin (Fig. 6C) . Control samples incubated with rabbit IgG failed to bring down the ezrin antibody (Fig. 6C , lanes 1, 3). Of note, both EBP50 (Fig. 6C , lane 2) and SAP97 (Fig. 6C , lane 4) immunoprecipitates included ezrin, suggesting a stronger interaction between SAP97 and ezrin in RPE cells in the eye than in RPE cultures. 

Discussion

The RPE–photoreceptor boundary is the site of many vital biochemical and biophysical processes required for the normal function of the retina. Polarity studies have shown that the neural retina influences the organization of the apical surface of RPE by promoting the apical localization of certain proteins, such as N-CAM, Na,K-ATPase and Emmprin, but the mechanisms involved in these phenomena are still unknown. ^{47 48 49 50} Understanding the molecular organization of the apical and basal surfaces of the RPE should provide important clues to the mechanisms involved in the interaction of RPE with the neighboring photoreceptors and choroid and in the control of RPE cell polarity. We recently identified a key role for the actin-binding protein ezrin in the establishment of the apical microvilli and basal infoldings of RPE. ¹ In the current study, we report the presence of two PDZ domain–containing proteins, EBP50 and SAP97, in RPE; their polarized distribution; and their interaction with ezrin and the actin cytoskeleton. 

This is the first report of the presence of EBP50 in the apical microvilli of RPE cells. A previous study detected very low levels of EBP50 in eye lysates solubilized with SDS but did not identify the RPE as the source. ²⁰ In addition, in our study EBP50 also localized to the apical surface of RPE cells in primary culture, and its apical localization and association with the cytoskeleton was strikingly increased by ezrin overexpression in a rat RPE cell line (RPE-J) that expresses low levels of endogenous ezrin. Previous work has demonstrated enrichment of EBP50 in the apical microvilli of certain nonocular epithelia. ²⁰  

This is also the first report of the presence of SAP97 at the basolateral surface of RPE cells. Colloidal gold immunoelectron microscopy for SAP97 labeled the elaborated basal infoldings of RPE cells obtained from adult rat eyes. A previous study reported the immunocytochemical detection of various synapse-associated proteins during the postnatal development of the rat retina, but the RPE and the cell layers underneath were not included in those observations. ⁵¹ Similarly, SAP97 had been detected at the plasma membrane of various cell types including neurons and epithelia. ^{31 37 38} The localization of SAP97 to basolateral infoldings in RPE in situ is identical with that of ezrin. ¹ It is worth noting that there are very few reports of a basolateral localization of ezrin. This has been shown in immature enterocytes in intestinal crypts, in the intricate podocyte extensions to the glomerular basement membrane in renal corpuscles, ⁵² and in the basolateral membrane infoldings of both resting and stimulated parietal cells. ⁵³ It is therefore of interest that SAP97 localized primarily to the lateral membrane of RPE cells in primary culture, maintaining a good colocalization with ezrin at that level. The lateralization of SAP97 may represent the relative loss of basal infoldings in culture. A lateral distribution of SAP97 and its human homologue hDlg has been previously reported in other epithelial cells in culture. ^{29 37 38 39} An important observations was that overexpression of ezrin in RPE-J cells led to an increase in resistance to detergent extraction of both SAP97 and ezrin, suggesting an interaction of these two proteins that may be functionally relevant. However, the low levels of expression of SAP97 in the basal surface of clone 16 may be correlated to the fact that ezrin overexpression regenerates both microvilli and basal infoldings but not to the same extent and organization observed in primary RPE cultures and to the RPE in vivo. The basal localization of SAP97 in rat primary RPE cultures, observed more frequently in long-term (8-week) cultures, suggests that basal localization of SAP97 is stabilized by unknown factors in cells in culture. 

In rats, both RPE and photoreceptors mature during the first 2 weeks after birth, which involves the development of long apical microvilli and elaborated basal infoldings in RPE and the formation of outer segments in photoreceptors. Immunoblots of rat RPE cell lysates at different postnatal ages demonstrated that both EBP50 and SAP97 levels increased several fold as the RPE matured when compared with the expression of the integrin αv, a protein expressed at constant adult levels from birth. The increased expression of EBP50 and SAP97 reflect the maturation of the actin cytoskeleton as the RPE consolidates its interaction with the surrounding tissues. Previous work has shown that the actin content of mature RPE in 21-day-old chicken embryos (in chicken, as in humans, the maturation of the retina occurs before birth) is four times that of the immature RPE of 11-day-old chicken embryos ⁵⁴ and that ezrin levels increase fourfold during postnatal maturation of rat RPE. ¹  

The pull-down experiments rendered with the GST-ezrin fusion proteins further characterized the interaction between EBP50 and SAP97 with ezrin in RPE. Our data confirmed previous data in other cells showing that EBP50 binds to the NH₂-terminal domain but not to the full-length ezrin, because of intramolecular binding of the NH₂- and COOH-terminal domains of ERM proteins. ^{20 46} SAP97 bound with more intensity to full-length ezrin than to the NH₂-terminal domain, suggesting the existence of additional binding sites localized away from the NH₂-terminal domain of ezrin. A similar behavior was shown for the binding of ezrin toα -actin. ⁵⁵ It has been established that hDlg is capable of binding to members of the band 4.1 superfamily through the first two PDZ domains and the I3 domains. ^{31 32} Nonetheless, in those experiments a high-affinity interaction was evident between the two proteins, whereas our data suggest a low-affinity interaction between SAP97 and ezrin in RPE cells, in that its detection required the use of the cross-linker DSP. This could reflect the expression of different isoforms of SAP97 with different ezrin-binding properties in RPE cells or a different regulation of ezrin in RPE cells. Indeed, the stimulation of microvilli assembly in RPE solely by overexpression of full-length ezrin in the absence of external stimulation is a relatively unique observation—a phenomenon that has not been reported in other cell types. The interaction of both EBP50 and SAP97 with ezrin was further analyzed by immunoprecipitation performed in RPE sheets collected from adult eyes. In vivo, a stronger interaction between SAP97 and ezrin was observed, because the interaction of both proteins could be detected without the use of a cross-linking agent before the immunoprecipitation. This reflects additional stabilizing mechanisms in native RPE cells. 

The health and integrity of the neural retina photoreceptors depend on a well-regulated extracellular environment. The RPE performs nursing functions that regulate and determine the health of the photoreceptors. These functions include structural support of the photoreceptor OS, ⁵⁶ daily phagocytosis of shed fragments of photoreceptor OS, ^{57 58} transport and metabolism of retinal lipids involved in the visual cycle and visual pigment regeneration, ^{59 60} regulation of the transport of metabolites and ions between the neural retina and the choroid, absorption of scattered light, and the control of the volume and composition of the fluid in the subretinal space through the transport of ions, fluid, and metabolites. ⁶¹ All these functions rely on the presence of diverse plasma membrane transporters and receptors present either in the apical or basolateral membrane domains of RPE. Because most ion transporters and channels are polytopic (extend through the membrane several times) and as an increasing number of these transporters has been shown to interact with PDZ domain proteins such as EBP50 and SAP97 in other systems, our results open the way to identify transporting functions of RPE cells that may be coordinated or regulated through the interaction with polyvalent PDZ-containing adaptor proteins. The ability of PDZ proteins to cluster transmembrane receptors and channels has great functional significance for the activation of these proteins, as shown by the example of the inactivation no after-potential D (INAD) signaling conglomerate in Drosophila’s photoreceptors. ⁶² It is likely that the uncovering of partners of EBP50 and SAP97, a current objective in our laboratory, will help in understanding not only details of the morphogenesis of RPE but also various aspects of the maintenance of a healthy neural retina and of derailment into a disease state. 

 Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2000.

 Supported by National Institutes of Health Grant EY08538, by a Jules and Doris Stein Professorship from the Research to Prevent Blindness Foundation (ER-B), by a Norman and Rosita Winston Foundation fellowship (VLB), and by the Dyson Foundation.

 Submitted for publication July 6, 2001; accepted August 21, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Polarized expression of EBP50 and SAP97 at the RPE microvilli in vivo. Rat eyes of different ages (P0 and adult) were fixed with 4% paraformaldehyde immediately after enucleation. Cryosections (10 μm) were stained with specific antibodies to ezrin (B, E, H, K, Cy3, red), EBP50 (A, D, FITC green), or SAP97 (G, J, FITC, green). Nuclei were stained with DAPI (blue). The labeled cryosections were observed under an epifluorescence microscope, and images were collected with a cooled CCD camera. Digitally acquired images were translated using image-management software. Immunofluorescence images of all three channels were merged digitally by image processing; overlapping green and red fluorescence appears as yellow (C, F, I, L). At all ages, both ezrin and EBP50 were detected at the apical RPE surface, suggesting that, in the eye, ezrin and EBP50 colocalize at RPE microvilli. The apical extension of EBP50 and ezrin immunofluorescence in mature RPE (D, E) reflects the growth of long and thin microvilli that surround the mature photoreceptor outer segments. SAP97 distributed both on the lateral (arrows) and basal RPE surfaces at all ages (G, J). However, the basolateral localization of SAP97 in adults was best observed in samples without the neural retina atop the RPE layer (J). A minor fraction of ezrin was detected at the basal surface of RPE cells, both in immature (H) and mature (K) RPE cells and was partially codistributed with SAP97. RN, retinal nuclei; Ch, choroid; ONL, outer nuclear layer; IS, photoreceptor inner segments. Bar, 10μ m.

Figure 2.

 

Immunogold localization of SAP97 to basal infoldings in adult rat RPE. Adult rat eyecups were fixed in 4% paraformaldehyde, 0.1% glutaraldehyde, and 0.2% picric acid prepared in PHEM buffer. Tissue was sequentially dehydrated in methanol, embedded in resin, and polymerized at −20°C under UV light. Ultrathin sections were sequentially reacted with a rabbit polyclonal antibody to SAP97 followed by a gold-conjugated (10 nm) donkey antibody to rabbit IgG. (A) Control samples reacted with a donkey anti-rabbit IgG antibody showed no labeling. (B) Immunogold labeling was specifically associated with basal infoldings (BI, arrowheads). P, pigment granule; BM, Bruch’s membrane. Bar, 1 μm.

Figure 3.

 

Increased EBP50 and SAP97 expression during postnatal maturation of rat RPE. RPE of different ages (P2–P21) was harvested and lysed in RIPA buffer supplemented with protease and phosphatase inhibitors. Protein (20 μg) from each lysate was separated in a 7.5% SDS gel, transferred to nitrocellulose membranes, and probed with antibodies specific to both EBP50 (A) and SAP97 (B) followed by enhanced chemiluminescence (ECL) detection of immunoreactivity. In parallel, lysates were run under nonreducing conditions, transferred to membranes, and probed for the integrin αv, a protein previously characterized to have their levels of expression unchanged throughout the postnatal period (C). Membranes were exposed to film and signal intensities were analyzed on computer. Intensity signals of both EBP50 and SAP97 were normalized to the αv integrin and plotted as a percentage. During RPE maturation, both EBP50 (gray bars) and SAP97 (filled bars) expression was upregulated approximately 10 and 4-fold. Data correspond to one representative experiment. Values reported are the mean of three independent experiments ± SEM.

Figure 4.

 

EBP50 predominantly localized to the apical surface of primary RPE cells and RPE-J cells overexpressing ezrin. Primary rat RPE (A–D), RPE-J (E–H), and RPE-J clone 16, overexpressing exogenous ezrin (I–L) were plated on synthetic membrane–coated semipermeable filters to promote differentiation. The monolayers were permeabilized with a buffer containing Triton X-100 for 40 seconds, fixed with paraformaldehyde, and processed for indirect immunofluorescence with ezrin (A, C, E, G, I, K) or EBP50 (B, D, F, H, J, L) antibodies. Samples were observed in an LSCM. For each staining, horizontal sections were summed up into one image. Horizontal (xy) confocal sections of primary RPE cultures and of clone 16 displayed punctate staining typical of apical microvilli localization of ezrin (A, I) and EBP50 (B, J). Vertical (xz) confocal sections through the epithelial monolayers confirmed that ezrin (C, K) and EBB50 (D, L) colocalize mostly at the apical surface of primary RPE and clone 16 monolayers. In contrast, both ezrin and EBP50 are extracted in larger amounts by Triton X-100 in wild-type RPE-J cells (E, F). A small amount of EBP50 was found associated with the apical plasma membrane of RPE-J cells (F, H). Bar, 5 μm.

Figure 5.

 

Ezrin and SAP97 predominantly colocalize to the basolateral surface of cultured cells. Long-term (8 weeks) primary rat RPE (A–D), RPE-J (E–H), and RPE-J clone 16, overexpressing exogenous ezrin (I–L) were plated on synthetic membrane–coated semipermeable filters to promote differentiation. Differentiated monolayers were permeabilized with a buffer containing Triton X-100 for 40 seconds, paraformaldehyde fixed, and immunolabeled for ezrin (A, C, E, G, I, K) and SAP97 (B, D, F, H, J, L). Samples were observed in an LSCM. Each individual xy image of the RPE monolayers stained with SAP97 and ezrin antibodies had their top 2 μm eliminated in both channels to remove most of the apical ezrin staining, and the remainder of the planes were summed up into one image. Horizontal cross sections (xy scans) of the long-term primary RPE monolayers (A–D) and the clone 16 (I–L) displayed a lateral ezrin localization (A, I) that overlapped with the SAP97 basolateral distribution (B, J). Vertical sections (xz scans) through the monolayers revealed that ezrin localized mostly to the apical surface of long-term primary RPE and RPE-J clone 16 (C, K), whereas SAP97 localized mostly to the basolateral membrane (D, L). An overlapping localization of both proteins was observed at the top lateral membrane. As previously shown, in polarized RPE-J cells, ezrin was mostly extracted by Triton X-100 treatment (E, G), but weak labeling of SAP97 could still be detected mostly at the lateral membrane (F, H). Bar, 5μ m.

Figure 6.

 

EBP50 and SAP97 directly interact with ezrin in RPE cells in vitro and in vivo. To further analyze the interaction of ezrin with both EBP50 (A) and SAP97 (B), whole cell lysates of RPE-J clone 16 monolayers were incubated with agarose beads complexed to GST (lane 1), GST-NH₂-terminal (lane 2), and GST-full-length ezrin (lane 3). Precipitates were resolved in a 10% SDS-PAGE gel, transferred to nitrocellulose followed by immunoblot to EBP50 and SAP97. EBP50 was shown to bind to the GST-NH₂-terminal domain ezrin. Lysates were cross-linked with DSP before the incubation with the GST-ezrin fusion proteins, to allow detection of the interaction with SAP97. SAP97 was shown to bind to both the NH₂-terminal domain of ezrin and full-length ezrin. (C) Furthermore, the interactions of EBP50 (lanes 1, 2) and SAP97 (lanes 3, 4) with ezrin were analyzed in immunoprecipitates from RPE sheets from adult eyes. Control samples (lanes 1, 3) incubated with rabbit IgGγ -globulin failed to bring down ezrin. However, both EBP50 (lane 2) and SAP97 (lane 4) were able to bring down ezrin. Each lane contains RPE sheets collected from an adult eye.

The authors thank Dena Almeida for expert help with the preparation of cryosections and Leona Cohen-Gould for excellent assistance with the electron and confocal microscopy.