The Glomerular Epithelial Cell Anti-Adhesin Podocalyxin Associates with the Actin Cytoskeleton through Interactions with Ezrin

Materials

Frozen rat kidneys were purchased from PelFreeze (Rogers, AR). SuperSignal chemiluminescence system and protein G-agarose were from Pierce (Rockford, IL). Cytochalasin D and latrunculin B were purchased from CalBiochem (San Diego, CA), and Texas-Red conjugated phalloidin was obtained from Molecular Probes (Eugene, OR). Polyvinylidene difluoride membranes were obtained from Millipore (Bedford, MA). Detergents were from Fisher Biotech (Tustin, CA) or ICN (Costa Mesa, CA). Cell culture media was obtained from the UCSD Cell Culture Core Facility, and tissue culture plastics were from Corning (Corning, NY). Sialidase (Arthrobacter ureafaciens) was purchased from Boehringer Mannheim (Indianapolis, IN). Kodak Biomax MR film was obtained from Sigma (St. Louis, MO). The Rho-kinase inhibitor, Y-27692 (28), was generously provided by Welfide Corporation (Osaka, Japan).

Antibodies

Anti-podocalyxin monoclonal antibodies (mAb) 5A and 1A, raised against Triton X-114 detergent-extracted rat glomeruli (29), reacts with the ectodomain of rat podocalyxin. Anti-podocalyxin polyclonal antibody (pAb) (0601) was generated against a synthetic peptide corresponding to the C-terminal 13 amino acids of rat podocalyxin as described previously (1). Anti-ezrin pAb was raised against recombinant human ezrin as described previously (30,31) and demonstrates no cross reactivity with moesin and radixin, two closely related members of the ERM family. Anti-ezrin mAb (3C12) was purchased from NeoMarkers, Inc. (Fremont, CA). The epitope for mAb 3C12 is reported to be between amino acids 362 and 585. Anti-actin mAb AC-40 (32) was purchased from Sigma. Cross-absorbed anti-rabbit Alexa 488 IgG (H+L) and anti-mouse Alexa 594 F(ab′)₂ were purchased from Molecular Probes, anti-rabbit (10 nm) and anti-mouse (5 nm) gold conjugates were purchased from Amersham-Pharmacia (Piscataway, NJ), and horseradish peroxidase—conjugated anti-rabbit and anti-mouse IgG was purchased from BioRad (Hercules, CA).

Immunofluorescence Microscopy

Kidneys of Sprague Dawley rats were perfused with Dulbecco's modified Eagle's medium, followed by 4% paraformaldehyde in 100 mM phosphate buffer (pH 7.4; 15 min). Kidneys were excised and immersion fixed in 4% paraformaldehyde (45 min) followed by 8% paraformaldehyde (15 min). Samples were cryoprotected and frozen in liquid nitrogen (33). Semithin cryosections (0.5 to 1.0 μm) were cut on a Leica Ultracut UCT microtome (equipped with a cryoattachment at - 100°C), placed on gelatin-coated microscope slides, and incubated for 2 h at room temperature with affinity-purified anti-podocalyxin (0601) and anti-ezrin mAb 3C12, followed by incubation for 1 h with goat anti-rabbit Alexa 488 and anti-mouse Alexa 594. Sections were observed with a Zeiss Axiophot microscope equipped for epifluorescence. Images were captured with a CCD camera and SCION Image software and processed with Adobe Photoshop 5.0. Final images were printed on Kodak Ektatherm XLS paper using a Kodak ColorEase PS printer.

Immunoelectron Microscopy

Rat kidneys were fixed and prepared as described above for immunofluorescence. Ultrathin cryosections (approximately 80 nm) were cut, blocked for 30 min with 10% fetal calf serum in phosphate-buffered saline (33), and incubated sequentially with anti-ezrin pAb for 16 h at 4°C followed by anti-podocalyxin mAb 5A for 2 h at 23°C. Bound primary antibodies were detected by incubation with anti-rabbit and anti-mouse gold conjugates (5 and 10 nm, respectively; 1:50). Sections then were stained with 2% neutral uranyl acetate (20 min); adsorption-stained with 0.2% uranyl acetate, 0.2% methyl cellulose, and 3.2% polyvinyl alcohol; and examined with a JEOL 1200 EX-II electron microscope.

Glomerular Isolation

All procedures were performed at 4°C. Frozen rat kidneys were dissected to obtain cortex, and isolated cortex was finely minced and suspended in 20 mM HEPES (pH 7.4), 150 mM NaCl containing protease inhibitor mix (2 μg/ml chymostatin, 2 μg/ml leupeptin, 2 μg/ml antipain, 2 μg/ml pepstatin, 10 μg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride). Material was gently pressed through a 350-mesh nylon screen followed by a 150-mesh screen. Filtered material was poured onto a 70-mesh screen and rinsed with cold phosphate-buffered saline. Glomeruli remaining on top of the screen were collected, pelleted by centrifugation (500 × g, 5 min, 4°C), and rinsed with cold phosphate-buffered saline. The preparation consisted of >95% glomeruli when examined with an inverted phase-contrast microscope.

Cell Culture

MDCK-PC8 and CHO-PC13 cells stably expressing rat podocalyxin were generated as described previously (1). Cells were grown in modified Eagle's medium-Earle salts and Ham's F-12, respectively, supplemented with 10% fetal calf serum (Hyclone, Logan, UT), 100 μg/ml streptomycin sulfate, 100 U/ml penicillin G, and 0.5 mg/ml G418 (CalBiochem, San Diego, CA).

Actin Depolymerization and Texas Red—Phalloidin Staining of Actin

Cells were grown to approximately 75% confluence on glass coverslips, incubated for 30 min at 37°C with Dulbecco's modified Eagle's medium alone or either cytochalasin D (20 μg/ml) or latrunculin B (10 μg/ml) diluted into Dulbecco's modified Eagle's medium, fixed for 20 min with 2% paraformaldehyde in PBS, and quenched by incubation for 20 min with 20 mM Tris (pH 7.4), 150 mM NaCl (TBS), 0.1 M glycine. Cells were then permeabilized for 5 min with 0.05% Triton X-100 in TBS, 0.1 M glycine, and incubated for 1 h with Texas Red phalloidin (1:100) in TBS, 10% calf serum. After rinsing, cells were mounted in Gelvatol (Monsanto, St. Louis, MO) containing 1 mg/ml paraphenylenediamine and examined with a Zeiss Axiophot equipped for epifluorescence. Images were captured and processed as described above and printed by use of a Codonics NP-1660 black and white thermal printer.

Immunoprecipitations and Immunoblotting

In some cases, before immunoprecipitations, MDCK-PC8 and CHO-PC13 cells were treated with A. ureafaciens sialidase, as described previously (1), or incubated at 37°C for 4 h with Y-27632 at the indicated concentrations diluted into their respective growth media. For immunoprecipitation, proteins from isolated rat kidney glomeruli or MDCK-PC8 or CHO-PC13 were solubilized with 1% Triton X-100 in TBS and protease inhibitor mix. Detergent insoluble material was removed by centrifugation (12,000 × g, 5 min, 4°C), and soluble proteins were incubated with a mix of anti-podocalyxin mAb 5A and 1A (3 μg each) and protein G-agarose for 16 h at 4°C. Immunoprecipitates were washed with TBS containing 1% Triton X-100 (3×) followed by TBS (1×), solubilized in Laemmli sample buffer supplemented with 4% β-mercaptoethanol and heated at 95°C for 3 min. For immunoblotting, antibody-bound proteins were resolved by sodium dodecyl sulfate—polyacrylamide gel electrophoresis and electrotransferred to polyvinylidene difluoride membranes using a wet tank transfer method (BioRad). Membranes were blocked with TBS, 0.1% Tween-20, and 5% calf serum for 30 min at 23°C and incubated with either anti-podocalyxin or anti-ezrin (1:4000) antisera or anti-actin mAb AC40 (100 ng/ml) in blocking buffer for 1 h at 23°C. Membranes were washed 3 times (10 min) with TBS, 0.1% Tween-20, 1% calf serum, and bound antibodies were detected with species-specific horseradish peroxidase—conjugated secondary antibodies, followed by chemiluminescence detection according to the manufacturer's instructions (Pierce).

Selective Detergent Extraction and Co-Sedimentation Analysis

Cortex was removed from frozen rat kidneys, finely minced, re-suspended in 20 mM HEPES (pH 7.4) and 150 mM NaCl, and processed using a TissueMizer (Tekmar, Cincinnati, OH) equipped with 0.5-cm rotary blades (3 min). Cortical homogenate was incubated for 3 h at 4°C with an equal volume of 10 mM imidazole (pH 7.2), 75 mM KCl, 5 mM MgCl₂, 1 mM ethyleneglycoltetraacetic acid containing 2% (1% final concentration) of Triton X-100 (TX-100), 3-[(3-Cholamidopropyl)-dimethylammonio]-1-propane-sulfonate hydrate, Tween-20, sodium dodecyl sulfate, or N-lauroyl sarcosine and protease inhibitor mix. In some cases, 0.6 M KI was added to the solubilization step for actin depolymerization (34,35,36). Detergent-solubilized proteins were centrifuged at 100,000 × g at 4°C for 45 min. Pelleted material was resuspended to identical volume as the supernatant, and proteins were analyzed by immunoblotting as described above with anti-podocalyxin, anti-ezrin, and anti-actin antibodies.

Podocalyxin and Ezrin Co-Localize in Rat Kidney Glomeruli

We showed previously that both podocalyxin and ezrin are expressed in glomerular epithelial cells and that they have a similar distribution at the apical plasma membrane of podocyte foot processes (14). Here we performed double-labeling studies to determine whether podocalyxin and ezrin co-localize in these cells. By immunofluorescence, podocalyxin is found primarily in podocytes (Figure 1A) but also is present in endothelium of glomerular and peritubular capillaries (Figure 1D). In glomeruli, ezrin is found in podocytes (Figure 1B), where it strikingly overlaps with podocalyxin (Figure 1C). Ezrin also is found in the microvillar region of proximal tubule epithelia (Figure 1E).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Localization of podocalyxin and ezrin in rat kidney by immunofluorescence. Semithin cryosections from rat kidney were incubated with anti-podocalyxin pAb (0601) (A, D) and anti-ezrin mAb (3C12) (B, E), followed by anti-rabbit Alexa 488 and anti-mouse Alexa 594 secondary antibodies. In glomeruli, the distribution of podocalyxin (A) and ezrin (B) overlaps in podocytes (C). Podocalyxin also is highly expressed in peritubular capillaries (D), and ezrin (E) is concentrated in the microvillar region of proximal tubule cells. No overlap in the distribution of podocalyxin and ezrin expression is evident outside of glomeruli (F).

Because podocalyxin and ezrin are co-expressed in podocytes, we next determined whether they co-localize along the apical plasma membrane of podocyte foot processes by performing double immunogold labeling of ultrathin cryosections from rat kidney. Podocalyxin and ezrin were found to be concentrated along the cell body and the apical plasma membrane of the foot processes above the slit diaphragms (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Co-localization of podocalyxin and ezrin in podocytes. Ultrathin cryosections from rat kidney were incubated with anti-podocalyxin mAb 5A (recognizes the ectodomain of podocalyxin) and anti-ezrin pAb, followed by gold-conjugated anti-mouse (10 nm) and anti-rabbit (5 nm) IgG (A, B). Podocalyxin and ezrin can be seen along the cell body and apical plasma membrane of foot processes above the level of the slit diaphragms. (C, D) Higher magnification views of fields in (A) show podocalyxin (large gold) on the outer surface of the plasma membrane (PM) and ezrin (small gold) on the inner surface of the PM of the podocyte. Magnification bar, 0.1 μm.

Podocalyxin and Ezrin Are Associated in Glomerular Epithelial Cells

Co-localization of podocalyxin and ezrin at the apical surface of podocytes prompted us to examine whether these two proteins are associated in a molecular complex and can be co-immunoprecipitated. Toward this end, proteins were detergent-extracted from isolated rat kidney glomeruli and subjected to immunoprecipitation with anti-podocalyxin mAb, followed by immunoblotting with anti-ezrin pAb. As shown in Figure 3, ezrin could be co-immunoprecipitated with podocalyxin, indicating that it is complexed with podocalyxin in glomeruli. After immunoprecipitation with mAb 20B to megalin, another kidney antigen present in podocytes and the proximal tubule brush border, no ezrin could be detected in the precipitates, thus demonstrating the specificity of ezrin's interaction with podocalyxin.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Ezrin co-immunoprecipitates with podocalyxin from isolated rat glomeruli. Immunoblot analysis of detergent-extracted glomerular proteins shows that ezrin is abundantly expressed and migrates at its expected relative molecular weight of approximately 85 kD (lane 1). Glomerular proteins were solubilized in Triton X-100 and immunoprecipitated with anti-podocalyxin mAb (5A/1A), followed by immunoblotting of the precipitate with anti-ezrin polyclonal antibody (pAb). Ezrin is detected in the immunoprecipitates, which indicates that it is part of a detergent-resistant complex with podocalyxin (lane 2). When immunoprecipitation is done with an unrelated mAb (20B) to megalin, no ezrin is found in the co-precipitated material (lane 3), demonstrating that ezrin is specifically complexed with podocalyxin.

Podocalyxin Is Associated with Actin Filaments

Because ezrin is known to mediate interactions between cell surface proteins and the actin cytoskeleton, we next examined whether podocalyxin is associated with actin filaments. Actin filaments are stable structures that are resistant to mild detergent extraction. After detergent treatment, actin filaments as well as tightly associated proteins remain insoluble and are pelleted by centrifugation at 100,000 × g. Therefore, we extracted rat kidney membranes with both mild nondenaturing detergents (Triton X-100, 3-[(3-Cholamidopropyl)-dimethyl-ammonio]-1-propane-sulfonate hydrate, or Tween-20) or with harsh denaturing detergents (sodium dodecyl sulfate or N-lauroyl sarcosine), centrifuged the extracts at 100,000 × g, and immunoblotted the resulting detergent soluble (supernatant) and insoluble (pellet) fractions with anti-podocalyxin and antiactin antibodies. As expected, actin was found primarily in the insoluble fraction when membranes were extracted with mild detergents and in the soluble fraction when extracted with denaturing detergents (Figure 4). A significant fraction of the podocalyxin co-sedimented with actin in the insoluble fraction after mild detergent extraction, whereas all of the podocalyxin together with actin was found in the soluble fraction after extraction with denaturing detergents.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Podocalyxin co-sediments with actin after extraction in mild detergents. Kidney homogenates were incubated with the indicated detergents followed by centrifugation at 100,000 × g to separate proteins into soluble (S, supernatant) and insoluble (P, pellet) fractions. Proteins were fractionated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and immunoblotted with antipodocalyxin pAb (0601) or anti-actin mAb (AC40). In the absence of detergent, podocalyxin and actin are found in the pellet (P) as expected. With mild (nondenaturing) detergent extraction, i.e., Triton X-100 (TX-100), CHAPS, or Tween-20 (TW-20), the majority of the actin is insoluble, and a significant portion of podocalyxin is found in the pellet. However, after extraction with denaturing detergents, i.e., SDS and N-lauroyl sarcosine (sarkosyl), that are known to disrupt actin filaments, both actin and podocalyxin are found in the soluble fraction.

We next examined the distribution of podocalyxin and ezrin in detergent-soluble and -insoluble fractions before and after actin depolymerization, to determine whether the co-sedimentation of podocalyxin with actin was dependent on intact actin filaments. Rat kidney cortical homogenates were extracted with mild detergents in the absence or presence of 0.6 M KI, which was shown previously to depolymerize actin (34,35,36). In the absence of KI, both podocalyxin and ezrin co-sedimented with actin (Figure 5). However, in the presence of KI actin, podocalyxin and ezrin were found in the soluble fraction. These data indicate that sedimentation of podocalyxin requires intact actin microfilaments and that podocalyxin is associated with the actin cytoskeleton through complex formation with ezrin.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Podocalyxin is released into the supernatant after depolymerization of actin filaments. Kidney homogenates were incubated with nondenaturing detergent (TX-100) in the presence (+KI) or absence (-KI) of 0.6 M potassium iodide. Detergent-soluble and -insoluble fractions were prepared by centrifugation (100,000 × g), and proteins were immunoblotted with anti-podocalyxin pAb (0601), anti-ezrin pAb, or anti-actin mAb (AC40). In the absence of KI, a significant portion of podocalyxin co-sediments with actin filaments after extraction with TX-100. However, after depolymerization of actin filaments with KI, all of the podocalyxin is found in the detergent-soluble fraction.

Interaction between Podocalyxin and Ezrin Is Not Dependent on Intact Actin Filaments

We next addressed whether the in vivo interaction between podocalyxin and ezrin is dependent on polymerized actin filaments. To this end, MDCK-PC8 and CHO-PC13 cells stably expressing rat podocalyxin (1) were treated with the actin depolymerizing agents cytochalasin D or latrunculin B, followed by immunoprecipitation with anti-podocalyxin and immunoblotting with anti-ezrin. As expected, treating MDCK-PC8 or CHO-PC13 cells with either cytochalasin D or latrunculin B resulted in complete disruption of the actin cytoskeleton, as seen by Texas Red—phalloidin staining (Figure 6A). Without treatment, ezrin was found to co-precipitate with podocalyxin from both MDCK-PC8 or CHO-PC13 cells (Figure 6B), indicating that exogenous expression of podocalyxin in these cell types is sufficient for interactions with endogenous ezrin and that this interaction does not require co-expression of additional podocyte-specific proteins. Cells treated with either cytochalasin D or latrunculin B showed no significant reduction in the amounts of ezrin co-precipitating with podocalyxin. These results indicate that the association between podocalyxin and ezrin in vivo is maintained after actin depolymerization and therefore is not dependent on intact actin microfilaments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Actin depolymerization has no affect on interactions between podocalyxin and ezrin. (A) MDCK-PC8 cells (upper panels) or CHO-PC13 cells (lower panels) stably expressing exogenous podocalyxin were incubated with or without cytochalasin D (20 μg/ml) or latrunculin B (10 μg/ml) and processed for fluorescence staining of actin with Texas Red-phalloidin, as described in the Materials and Methods section. Untreated cells demonstrate well-organized actin filaments. In the case of MDCK-PC8 cells, actin is concentrated around the PM, whereas in CHO-PC13 cells it is concentrated in stress fibers. Cytochalasin D and latrunculin B treatments result in depolymerization and fragmentation of actin in both cell types. (B) Parallel cultures of cells were incubated with or without cytochalasin D (Cy) or latrunculin B (La), and proteins were solubilized with TX-100. After immunoprecipitation with anti-podocalyxin mAb 5A/1A and immunoblotting with anti-podocalyxin pAb (0601), it can be seen that treating cells with Cy or La has no affect on podocalyxin expression. Immunoblotting with anti-ezrin shows that these actin depolymerizing reagents have no significant affect on the amount of ezrin coprecipitating with podocalyxin.

Interaction between Podocalyxin and Ezrin Is Not Affected by Sialidase Treatment or Rho-Kinase Inactivation

Events that alter extracellular domains of cell surface proteins often affect their intracellular interactions with cytosolic proteins. Because the polyanionic charge of podocalyxin is responsible for podocalyxin's anti-adhesion properties (1), we examined whether removal of sialic acid would result in loss of its association with ezrin. Toward this end, MDCK-PC8 or CHO-PC13 cells were treated with sialidase and subjected to co-immunoprecipitation analysis. As shown in Figure 7A, sialidase treatment of podocalyxin had no quantitative effect on the amount of ezrin that co-precipitated with podocalyxin. Thus, interactions between ezrin and the cytoplasmic tail of podocalyxin in transfected MDCK or CHO cells are independent of the ionic charge on podocalyxin's ectodomain.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Podocalyxin/ezrin associations are insensitive to sialidase treatment and Rho-kinase inhibition. MDCK-PC8 and CHO-PC13 cells were treated with sialidase for 1 h at 4°C (A) or incubated with the indicated concentrations of Rho-kinase inhibitor Y-27632 for 3 h at 37°C (B), followed by solubilization in 1% TX-100, immunoprecipitation with anti-podocalyxin mAb (5A/1A) and immunoblotting with either anti-podocalyxin pAb (0601, top panels) or anti-ezrin pAb (lower panels). (A) After sialidase treatment, there is no significant difference in the amount of ezrin (arrowhead) that co-immunoprecipitates with podocalyxin (^*), compared with untreated cells. The slower mobility of podocalyxin after sialidase treatment is due to removal of highly anionic sialic acid moieties (1,38). (B) Incubation of cells with various concentrations of the Rho-kinase inhibitor Y-27632 also has no affect on the amount of ezrin (arrowhead) that co-immunoprecipitates with podocalyxin (^*), although this reagent is able to disrupt actin organization in these cells (data not shown).

Rho-kinase—mediated phosphorylation of ezrin is necessary to activate ezrin to serve as a link between membrane proteins and actin filaments. Thus, if the association between podocalyxin and ezrin is in dynamic equilibrium, inhibition of Rho-kinase might be expected to lead to accumulation of inactive ezrin and reverse its association with podocalyxin. To examine more closely the dynamic stability of the interaction, we incubated MDCK-PC8 and CHO-PC13 cells with various concentrations of the potent Rho-kinase inhibitor Y-27632 (28) and quantified the association of podocalyxin and ezrin by co-immunoprecipitation analysis. When cells were treated with 40 μM Y-27632, extensive disruption and fragmentation of the actin cytoskeleton were seen in both cell lines as visualized by Texas Red—phalloidin staining (data not shown); however, no quantitative change was seen in the amount of ezrin that co-precipitated with podocalyxin (Figure 7B). These results indicate that once formed, interactions between podocalyxin and ezrin are very stable and not reversed by inactivation of Rho-kinase. Although previous studies showed that Rho-kinase is necessary for ezrin activation, this is the first demonstration that inhibition of Rho-kinase is not sufficient for inactivation of ezrin once it is complexed with a membrane protein.