Renal Physiology

Myosin 1e is a component of the glomerular slit diaphragm complex that regulates actin reorganization during cell-cell contact formation in podocytes

J. Bi, S. E. Chase, C. D. Pellenz, H. Kurihara, A. S. Fanning, M. Krendel


Glomerular visceral epithelial cells, also known as podocytes, are critical to both normal kidney function and the development of kidney disease. Podocyte actin cytoskeleton and their highly specialized cell-cell junctions (also called slit diaphragm complexes) play key roles in controlling glomerular filtration. Myosin 1e (myo1e) is an actin-based molecular motor that is expressed in renal glomeruli. Disruption of the Myo1e gene in mice and humans promotes podocyte injury and results in the loss of the integrity of the glomerular filtration barrier. Here, we have used biochemical and microscopic approaches to determine whether myo1e is associated with the slit diaphragm complexes in glomerular podocytes. Myo1e was consistently enriched in the slit diaphragm fraction during subcellular fractionation of renal glomeruli and colocalized with the slit diaphragm markers in mouse kidney. Live cell imaging studies showed that myo1e was recruited to the newly formed cell-cell junctions in cultured podocytes, where it colocalized with the actin filament cables aligned with the nascent contacts. Myo1e-null podocytes expressing FSGS-associated myo1e mutant (A159P) did not efficiently assemble actin cables along new cell-cell junctions. We have mapped domains in myo1e that were critical for its localization to cell-cell junctions and determined that the SH3 domain of myo1e tail interacts with ZO-1, a component of the slit diaphragm complex and tight junctions. These findings suggest that myo1e represents a component of the slit diaphragm complex and may contribute to regulating junctional integrity in kidney podocytes.

  • Myosin
  • actin
  • podocyte
  • slit diaphragm

myosin 1e (myo1e, previously known as myo1c) is a member of the myosin superfamily of actin-dependent motor proteins that is expressed in many cell types in humans and mice (3, 26, 49). Like other members of the myosin superfamily, myo1e contains an NH2-terminal motor domain responsible for actin binding and ATPase activity, a neck domain that includes a light chain-binding IQ motif, and a tail domain that may determine cargo-binding properties and intracellular localization. Myo1e tail domain consists of three tail homology (TH) regions: TH1 domain that binds acidic phospholipids (14), proline-rich TH2 domain, and SH3 domain, also known as TH3, which binds to proline-rich proteins (3, 28).

In the kidney, myo1e expression is particularly high in glomerular visceral epithelial cells known as podocytes, which play a key role in the selective filtration of proteins in the renal glomerulus (27). Myo1e-knockout (KO) mice exhibit defects in renal filtration that can be attributed to podocyte dysfunction (7, 27), and mutations in the myo1e gene are associated with glomerular disease in humans (1, 39, 46). The precise functional role of myo1e in glomerular podocytes is unknown.

Podocytes are highly specialized epithelial cells that form arborized, actin-rich projections known as foot processes. The intertwined foot processes cover the entire surface of glomerular capillaries and are connected with each other via modified cell-cell junctions called slit diaphragms. Slit diaphragms are unusual junctional complexes that contain tight junction proteins (ZO-1, occludin, and JAM-A), adherens junction proteins (P-cadherin), and junctional proteins that are specific to podocytes such as nephrin, Neph-1, and podocin (15, 16). During renal development, slit diaphragms arise from the tight junctions, which translocate from the apical position, characteristic of cuboidal podocyte precursors, to the basolateral position, typical of mature slit diaphragms (16, 29, 30, 42). The presence of the intact foot processes and slit diaphragms is necessary for normal filtration since mutations in the components of the podocyte actin cytoskeleton or the slit diaphragm proteins are associated with proteinuria (4, 2325). Myo1e localization pattern in mouse glomeruli is similar to that of the foot process and slit diaphragm proteins (7, 27). In earlier studies examining myo1e expression in cultured cells, endogenous myo1e (or its rat homolog, Myr3) was observed in cell-cell junctions (49, 51, 52). Taken together, the filtration defects of myo1e-null mice and the localization pattern of myo1e in the kidney and cultured cells suggest that myo1e may contribute to cell-cell adhesion and that some of the defects in myo1e-null podocytes may arise from disruption of cell contacts.

To test this hypothesis in the current study, we used a subcellular fractionation protocol that allows isolation of the detergent-resistant slit diaphragm fraction from mouse glomeruli (43). In this earlier study (43), myo1e was detected in the slit diaphragm fraction by mass spectrometry. In our experiments, the detergent-insoluble slit diaphragm fraction was enriched in myo1e, confirming that myo1e is tightly associated with the slit diaphragm proteins. We examined the dynamics of myo1e recruitment to nascent cell-cell contacts in cultured podocytes and the effects of disease-associated myo1e mutation on its localization to the junctions and on junctional actin assembly, mapped the regions of myo1e required for junctional localization, and identified tight junction protein ZO-1 as one of the binding partners of myo1e.



Rabbit anti-myo1e antibody was described previously (27, 49). The following commercial antibodies were used: mouse monoclonal anti-myc (clone 9E10; BioLegend), rabbit anti-GFP and mouse monoclonal anti-ZO-1 (Invitrogen), and rabbit anti-podocin (Sigma). Secondary antibodies labeled with Alexa 488 and 568 were from Invitrogen.

Expression constructs and cloning.

pEGFP-C1 myo1e and pEGFP-C1 myo1e tail constructs were described previously (28), as were myc-tagged ZO-1 constructs (12). Truncated myo1e constructs in pEGFP-C1 vector (Clontech) were produced using In-Fusion HD PCR cloning (Clontech). Primers used to subclone TH1, TH2, and TH3 domain were AA717-920, AA913-1060, and AA1052-1108, respectively. To verify expression level and molecular weight of the full-length and truncated myo1e constructs, GFP-tagged constructs were expressed in human embryonic kidney (HEK)-293 cells and analyzed by immunoblotting against GFP (data not shown). EGFP-myo1e constructs in pShuttle vector for adenoviral expression (Agilent Technologies) were described previously (39). mCherry-Lifeact and mCherry-ZO-1 (human) adenoviral constructs in the pShuttle vector were prepared according to the manufacturer's instructions. Glutathione-S-transferase(GST)-tagged proteins were prepared as described (28).

Animal studies.

All animal experiments were conducted in accordance with the protocols approved either by the State University of New York Upstate Medical University Committee for the Humane Use of Animals (mice) or by the Animal Care and Use Committee of the Juntendo University School of Medicine (rats for immunoelectron microscopy). Male Wistar rats (4–6 wk old) and neonatal rats (1–2 days old) were obtained from Charles River Japan (Kanagawa, Japan).

Cell culture, transfection, and infection.

Wild-type podocytes (40) were a generous gift from Dr. Peter Mundel, Massachusetts General, Boston, MA. Conditionally immortalized mouse kidney podocytes lacking myo1e were obtained from Myo1e knockout mice (27) expressing a transgene encoding a thermosensitive, interferon-inducible variant of the SV-40-T-antigen [Immortomouse (21)]. Podocyte isolation was performed, following the procedure of Mundel et al. (40). Individual podocyte clones were obtained by dilution cloning and tested by immunoblotting to verify the lack of myo1e expression as well as the presence of podocyte markers WT-1, synaptopodin, and podocin.

Madin-Darby canine kidney (MDCK) II Tet-Off cells (clone T23; Clontech) and AD-293 (ATCC) were grown in DMEM containing 10% FBS. Immortalized podocytes were cultured as described (40). Podocyte differentiation was induced by a switch from 33 to 37°C and removal of interferon-γ from culture medium; cells were differentiated for 2 wk prior to experiments. MDCK and HEK-293 cells were transfected using JetPEI DNA transfection reagent (Polyplus). Mouse podocytes were infected with adenovirus.

Glomerular purification and isolation of slit diaphragm-containing protein fraction.

Glomeruli were purified from mice using differential sieving. Isolation of slit diaphragm proteins was performed as described (43). Fractionation of glomeruli to isolate slit diaphragm proteins was performed in eight independent experiments with similar results.

GST pulldown assay.

For pulldown from 293 cells, transfected cells were lysed in lysis buffer containing 50 mM Tris·HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 10% glycerol, and complete protease inhibitor (Roche) and centrifuged for 10 min at maximum speed in a tabletop microcentrifuge at 4°C. The supernatant was incubated with GST-tagged proteins and glutathione agarose for 3 h at 4°C. The bead pellets and unbound proteins were separated by centrifugation. Beads were washed one time with lysis buffer and then separated again as described above. Finally, both bead pellets and unbound proteins were processed for SDS-PAGE.

Immunoelectron microscopy and immunocytochemistry.

Electron microscopy and immunogold labeling were performed as described previously (31). For immunofluorescence staining, mouse kidney cryosections were prepared and processed as described (7).

Fluorescence microscopy and live-cell imaging.

For live-cell imaging, MDCK cells or podocytes were plated onto glass-bottom Mattek dishes (Mattek); for podocyte plating, dishes were precoated with collagen IV. Twenty-four hours posttransfection/infection, cells were observed using a Perkin-Elmer UltraView VoX Spinning Disk Confocal system mounted on a Nikon Eclipse Ti microscope and equipped with an environmental chamber to maintain cells at 37°C. To observe contact reassembly in a calcium switch assay, podocytes were preincubated for 20 min in Hanks' balanced salt solution without calcium and magnesium containing 2 mM EGTA to remove calcium and then transferred into RPMI with 10% FBS and normal calcium content for 5–10 min at 37°C before being placed on the microscope stage.

Image analysis.

All fluorescence intensity measurements were done using Image J software (National Institutes of Health). Quantitative analysis was performed using KaleidaGraph (Synergy Software). The statistical significance was calculated using unpaired Student t-test with unequal variance.


Myo1e is a component of the slit diaphragm in the renal glomeruli.

To investigate the possibility that myo1e is enriched in the glomerular slit diaphragm, we used a slit diaphragm purification protocol developed by Pierchala et al. (43) to isolate the slit diaphragm complex from glomerular preparations. To determine whether myo1e was present in the slit diaphragm complex-containing fraction, we performed immunoblotting against endogenous myo1e. Myo1e was indeed enriched in the glomerular slit diaphragm-containing fraction along with the slit diaphragm markers podocin and ZO-1 (Fig. 1A). Surprisingly, myo1e was retained in the detergent-resistant slit diaphragm fraction, whereas in cultured cell lysates myo1e is usually soluble in nonionic detergents (see, for example, Ref. 28). Thus, myo1e appeared to interact with the components of the detergent-resistant slit diaphragm complex. Unlike myo1e, actin-binding protein synaptopodin was present primarily in the cytoplasmic fraction (Fig. 1A), indicating that the slit diaphragm isolation procedure was specific for the slit diaphragm components. To examine the localization of myo1e at the ultrastructural level, immunogold labeling of rat kidney sections was performed using antibodies against myo1e. In the adult rat kidney, myo1e localized to the intracellular side of the plasma membrane of podocyte foot processes, where it was concentrated at the base of the slit diaphragm (Fig. 1B, open arrowheads; slit diaphragm is indicated by the arrows). In the neonatal kidney, where slit diaphragm precursors were undergoing a gradual translocation from the apical to the basal position, myo1e localization was more variable. Although myo1e was still enriched in the plasma membrane in the regions corresponding to the slit diaphragms, some immunogold particles labeling myo1e were positioned more apically than in the adult kidney, whereas others were present close to the glomerular basement membrane (Fig. 1C).

Fig. 1.

Myosin 1e (myo1e) is a component of slit diaphragm. A: Western blot analysis of glomerular subfractions. Samples were blotted against endogenous ZO-1, myo1e, podocin, and synaptopodin. Images on both left and right (from 2 representative fractionation experiments) indicate that myo1e, podocin, and ZO-1 are enriched in the detergent-resistant (DR) slit diaphragm fraction, whereas another podocyte marker, synaptopodin, is present in the cytoplasmic fraction. The white line at right indicates that 2 separate parts of the blot have been placed next to each other. B: immunoelectron micrograph of the adult rat glomerulus shows that myo1e is concentrated at the base of the slit diaphragm (left). The same micrograph with the outlines of podocyte foot processes is shown on the right. Arrows indicate slit diaphragm, and an open arrow points to a foot process, whereas arrowheads point to gold particles that mark the anti-myo1e antibody. C: immunoelectron micrograph of the neonatal rat glomerulus displays variable location of myo1e along the lateral surface of podocyte foot processes. Most immunogold particles marking myo1e position are located adjacent to the slit diaphragms; however, the distance between myo1e signal and the glomerular basement membrane is highly variable. Bars, 100 nm.

Myo1e colocalizes with the slit diaphragm marker ZO-1 in mature and immature glomeruli.

Immunostaining of cryosections of mouse kidneys showed that myo1e colocalized with the slit diaphragm component ZO-1 in mature glomeruli (Fig. 2A). Myo1e staining did not completely overlap with ZO-1 labeling (arrows in Fig. 2A point to the regions stained for myo1e only), indicating that myo1e is present in podocyte cell bodies in addition to being enriched in the slit diaphragm region. Immunostaining of myo1e in immature glomeruli in cryosections of 1-wk-old mouse kidneys showed that myo1e was concentrated at the basal aspect of developing podocytes, where it colocalized with ZO-1 but not with the apical marker podocalyxin (Fig. 2B). Thus, during podocyte maturation, myo1e is enriched at the basal surface of podocytes in the region where the formation of new slit diaphragms was taking place.

Fig. 2.

Myo1e colocalizes with the slit diaphragm marker ZO-1 in mouse glomeruli and in cultured podocytes. A: immunostaining of cryosections of 2-wk-old mouse kidney with antibodies against ZO-1 and myo1e shows colocalization of myo1e and ZO-1 (yellow in merged image) in mature glomeruli as well as areas stained for myo1e only (green in merged, indicated by arrows). B: immunostaining of cryosections of 1-wk-old mouse kidney against podocalyxin, ZO-1, and myo1e shows that ZO-1 and myo1e localize to the basal aspect of podocytes in the immature glomeruli, whereas podocalyxin marks the apical side. C: immunostaining of wild-type mouse podocytes reveals localization of myo1e and ZO-1 to cell-cell junctions. Scale bar, 10 μm.

Myo1e is recruited to cell-cell contacts in cultured podocytes.

To elucidate the role of myo1e during remodeling of cell-cell junctions, we used conditionally immortalized podocytes to examine the localization of myo1e. Myo1e localized to the cell-cell junctions in wild-type podocytes (Fig. 2C). To determine whether the loss of myo1e activity resulted in changes in assembly of cell-cell junctions, we expressed either wild-type or mutant myo1e in the myo1e-null podocytes (obtained from the Myo1e-KO mice in our laboratory). Myo1e-KO podocytes were infected with adenoviruses encoding GFP-tagged myo1e and mCherry-tagged Lifeact, a marker of actin filaments (44). To induce contact remodeling, podocytes were subjected to a calcium switch treatment that induced disassembly of cell-cell junctions by decreasing Ca2+ concentration and reassembly of the junctions by Ca2+ replenishment (18). Cell-cell contacts observed using this model likely represent cell-cell adhesions reminiscent of slit diaphragms, since they could be immunostained using anti-ZO-1 antibody (data not shown). Wild-type myo1e was recruited to the newly formed cell-cell adhesions during recovery from the low Ca2+ condition (Fig. 3), where it colocalized with the actin filaments labeled by Lifeact. Kymographs in Fig. 3B illustrate that myo1e and actin were recruited to the nascent adhesions at the same time during junction formation.

Fig. 3.

Myo1e is recruited to the nascent contacts in cultured podocytes. A: myo1e-knockout (KO) podocytes were infected with adenoviruses encoding green fluorescent protein (GFP)-myo1e and mCherry-Lifeact for 24 h and subjected to calcium removal and replenishment to induce disassembly and reassembly of cell contacts. Time lapse imaging was started 5 min after the addition of normal calcium-containing medium. Frames selected from a time lapse movie show myo1e and Lifeact localization at 10-, 20-, and 30-min time points during imaging. Myo1e and actin are recruited to the newly formed contacts. Enlarged images on the right correspond to the boxed regions. Bars, 10 μm. B: kymographs show that both myo1e and actin are recruited to the nascent contacts at the same time.

We also examined localization of a mutant form of myo1e, myo1eA159P, during formation of new cell-cell contacts in podocytes (Fig. 4). A159P is a missense mutation of a highly conserved amino acid residue in the myo1e motor domain that is associated with the recessive form of focal segmental glomerulosclerosis (FSGS) in humans (39). When expressed in myo1e-null podocytes, GFP-myo1eA159P was slightly enriched in the nascent cell-cell junctions but showed a complete lack of colocalization with the actin filaments (Fig. 4B, kymographs).

Fig. 4.

Myo1eA159P mutant does not colocalize with actin in the newly formed contacts. A: myo1e-KO podocytes were infected with adenovirus encoding GFP-myo1eA159P and mCherry-lifeact 24 h before calcium switch treatment and live-cell imaging. Selected frames from a time lapse movie show cells at 10-, 20-, and 30-min time points. Enlarged images on the right correspond to the boxed region. Bars, 10 μm. B: kymographs show that myo1eA159P does not colocalize with nascent actin bundles.

Cells lacking functional myo1e exhibit defects in actin assembly along cell-cell contacts and in contact spreading.

Formation of cell-cell contacts in myo1e-KO podocytes expressing GFP-tagged wild-type myo1e was accompanied by the assembly of actin punctae along the cell-cell junction (Fig. 5, AC, arrowheads, and Fig. 3A). These punctae rapidly fused and elongated to form actin cables along the cell-cell boundary. In addition to the actin filament bundles located along the contact, radially positioned short actin spokes (Fig. 5B, white arrow) were also observed. The tips of some of these radial spokes colocalized with the areas of enrichment of myo1e (Fig. 5B, black arrow), and in some cases the spokes elongated as the contacts matured and the cells pulled away from each other while retaining punctate connections. In contrast to the cells expressing wild-type myo1e, KO podocytes expressing myo1e mutant predominantly formed radial actin spokes rather than longitudinal actin cables (Fig. 5, DF, white arrows, and Fig. 4A). For each myo1e construct (wild type and A159P mutant), eight pairs of cells forming contacts were observed. Seven out of eight pairs of cells expressing GFP-myo1e were observed to assemble actin cables along the cell-cell junctions, whereas five out of eight pairs of cells expressing GFP-myo1eA159P failed to assemble actin along the junction and formed only radial actin spokes. Many of the contacts formed by cells expressing myo1e mutant were relatively small and failed to spread (Fig. 5G), suggesting that the lack of assembly of new actin filament bundles along the junction may correlate with inefficient contact formation. Immunoblotting of podocytes infected with adenoviral vectors at 24 h postinfection (same expression time as used for imaging) showed that the amount of full length GFP-myo1eA159P expressed in these cells ranged from 71 to 86% of the full-length GFP-myo1e level in two independent experiments (data not shown). Although we cannot exclude the possibility that the defects in cell contact formation could be linked to the differences in the expression level of the two constructs, it appears unlikely considering that only relatively small differences in the expression level of wild-type vs. mutant GFP-myo1e were detected by immunoblotting.

Fig. 5.

Podocytes expressing mutant myo1e fail to assemble actin filament bundles along cell-cell boundaries. Contact reassembly following calcium switch treatment was observed in myo1e-KO podocytes expressing GFP-myo1e and mCherry-lifeact (AC) or GFP-myo1eA159P and mCherry-lifeact (DF). AC: cells expressing wild-type myo1e assembled actin filaments along the cell-cell junction (white arrowheads) and perpendicular to the junctions (white arrows). Bundles running perpendicular to the junctions were associated with myo1e-containing punctae at the cell edge (black arrow). A and B include enlarged images of the areas indicated by white rectangular outlines. DF: cells expressing myo1e mutant assembled primarily radial bundles of actin filaments (white arrows). Bars, 10 μm. G: box-and-whiskers plots of the rate of contact spreading measured using 9- to 10-min long time lapse movies of myo1e KO podocytes expressing either GFP-myo1e (n = 5 cells) or GFP-myo1eA159P (n = 6 cells). Horizontal line within each box indicates the median for each group; boxes correspond to the 25th to 75th percentile range. ○, Outliers.

Myo1e localization to cell-cell junctions requires multiple binding motifs.

To further map the domains of myo1e that are necessary for its localization to the junctions, we used GFP-tagged myo1e constructs that lack specific tail domains. Since transfection of podocytes is challenging and production of an adenoviral vector for expression of each truncated construct is very resource intensive, we used MDCK cells for these domain-mapping studies (Fig. 6). MDCK cells were used previously as a complement to cultured podocytes for the studies of slit diaphragm proteins and podocyte signaling pathways (32, 56). MDCK cells are also of renal epithelial origin [although unlike podocytes, MDCK cells represent epithelium of distal tubules (17)]. MDCK cells have well-developed cell-cell contacts (adherens and tight junctions) enriched in ZO-1; therefore, we chose this cell line as a model system to study myo1e localization to cell-cell junctions. As a quantitative measurement of junctional localization, we used the ratio of mean fluorescence intensity of GFP-myo1e along the cell-cell junction to the mean cytosolic intensity of GFP-myo1e as an indicator of myo1e enrichment in the junctions (Fig. 6B). Both values were background corrected using mean fluorescence intensity of an adjacent untransfected cell as an indicator of background fluorescence. The junctional fold enrichment score for those constructs that are completely cytosolic would be expected to be ∼0.5 (the intensity along the junction, which includes both a transfected and an untransfected cell, would be approximately one-half of the cytosolic intensity in the transfected cell). Indeed, constructs that had diffuse cytoplasmic localization had fold enrichment scores close to 0.5 (Fig. 6C).

Fig. 6.

Myo1e localization to the junctions requires multiple domain interactions. A: confocal sections through the apical portion of Madin-Darby canine kidney cells expressing GFP-tagged myo1e constructs and mCherry-Lifeact. Full-length myo1e, myo1e tail, myo1e lacking Src homology 3 (SH3) domain, and myo1e construct consisting of tail homology (TH)1 and TH2 domains localize to the junctions, whereas myo1e lacking TH2 domain shows primarily cytosolic localization. Scale bar, 10 μm. Line scans in the images at right show fluorescence intensity of the GFP and mCherry signals along a line drawn across 2 of the cell-cell junctions (indicated by the white lines in merged images). Peaks of GFP and mCherry fluorescence at the junctions coincide, except for the myo1e construct lacking TH2 domain, which is primarily cytosolic. B: method used to calculate fold enrichment of myo1e in the junctions. A line drawn along the junction of a transfected cell (green) with an untransfected cell (white) was used to measure mean fluorescence intensity (MI) in the junction (J), whereas a line drawn in the cytosol of a transfected cell was used to measure mean fluorescence intensity in the cytosol (C). Mean fluorescence intensity of the background (B) was measured by drawing a line through an untransfected cell. C: analysis of the fold junctional enrichment of myo1e. The bar graphs represent mean fold enrichment for 10 cells/construct (16 cells for the full-length myo1e), whereas the error bars correspond to standard deviation. IQ indicates light chain binding IQ motif. *Constructs that show statistically significant difference in fold enrichment relative to the full-length myo1e, with P values <0.01. **P < 0.0001. P values for the t-test comparison to the full-length myo1e: ΔSH3 = 0.2957, TH1TH2 = 0.01135.

Full-length myo1e and myo1e tail alone were localized to the intercellular junctions in MDCK cells (Fig. 6A). Myo1e tail showed a stronger enrichment in cell-cell junctions than the full-length myo1e (Fig. 6C), suggesting that the motor domain may target a subset of myo1e molecules to other intracellular structures, reducing its junctional enrichment. Myo1e tail expressed in podocytes also exhibited strong localization to cell-cell junctions (data not shown), indicating that the behavior of truncated constructs in MDCK cells and podocytes was similar. Deletion of the SH3 domain of myo1e did not significantly impair myo1e localization to the junctions (Fig. 6, A and C). However, myo1e construct lacking the TH2 domain had mostly cytosolic localization (Fig. 6). Localization of the TH2 domain alone was also mostly diffuse (Fig. 6C), whereas the construct consisting of the TH1 domain along with the TH2 domain exhibited junctional localization (Fig. 6A), and its enrichment in the junctions was similar to that of the full-length myo1e (Fig. 6C). All myo1e constructs that were enriched in the junctions also colocalized with actin, as indicated by the line scans of GFP-myo1e and mCherry-Lifeact intensity (Fig. 6A, right). As shown in Fig. 6C, myo1e-FL (full-length), myo1e-tail, myo1e-ΔSH3, and myo1e-TH1TH2 all exhibited significant enrichment at the junctions. However, if we deleted the TH2 domain from the full-length myo1e or TH1 domain from myo1e tail, the fold enrichment score dropped significantly (Fig. 6C). Therefore, we have concluded that the TH1 and TH2 domains together are necessary and sufficient to determine junctional localization of myo1e.

Myo1e SH3 domain binds to the COOH-terminal portion of ZO-1.

Since the myo1e SH3 domain was not necessary for myo1e localization to the junctions, we hypothesized that although the main role for the TH1 and TH2 domains may be to target myo1e to the junctions, the SH3 domain may serve as an effector domain, binding or activating additional proteins following myo1e recruitment to cell-cell contacts. To identify effector proteins that may bind to myo1e, we analyzed known protein components of the slit diaphragm complex and their protein interaction motifs. Based on the presence of proline-rich motifs, we hypothesized that ZO-1 could be a potential binding partner for the SH3 domain of myo1e. To test this potential interaction, we performed an in vitro pulldown assay using GST-tagged SH3 domain of myo1e to precipitate myc-tagged ZO-1 (Fig. 7A) from HEK-293 cell lysates. As shown in Fig. 7B, purified GST-tagged myo1e-SH3 precipitated full-length ZO-1. We used GST-tagged SH3WK mutant as a negative control since this mutation of a tryptophan to a lysine residue has been shown to disrupt SH3 domain binding to proline-rich targets (28). Neither GST alone nor GST-tagged SH3 mutant showed specific binding to ZO-1, which indicates that this interaction was specific to SH3 domain. Sequence analysis of ZO-1 identified several proline-rich regions located at the COOH-terminal portion of ZO-1, which could represent potential binding motifs for the SH3 domain of myo1e (Fig. 7A). Indeed, only the proline-rich domain containing, COOH-terminal portion of ZO-1, but not the NH2-terminal portion of ZO-1, bound to myo1e (Fig. 7B). When fluorescently tagged myo1e and ZO-1 were expressed in cultured podocytes, they colocalized in cell-cell contacts, further confirming the interaction (Fig. 7C).

Fig. 7.

Myo1e SH3 domain interacts with the COOH-terminal portion of ZO-1. A: diagram showing domain organization of myo1e and ZO-1 constructs used in the pulldown experiments. All ZO-1 constructs were tagged with the myc tag. Protein domains are indicated by TH, SH3, PSD-95/Dlg1/ZO-1 (PDZ), guanylate kinase (GUK), unique 6 (U6), ZO-1 and UNC-5 proline-rich (ZU5). B: Western blot analysis of glutathione-S-transferase (GST) pulldown reactions using anti-myc antibody. GST-SH3 domain of myo1e but not GST alone or GST-SH3 mutant precipitated myc-tagged ZO-1 full-length and ZO-1 COOH terminus. C: GFP-labeled myo1e and mCherry-tagged ZO-1 (full-length) colocalize when expressed in mouse podocytes. Scale bar, 10 μm.


Previous studies have identified a role for myo1e in the maintenance of glomerular filtration. Mice lacking myo1e in podocytes exhibit proteinuria, whereas patients with homozygous mutations in myo1e are characterized by FSGS and proteinuria (7, 27, 39). The importance of myo1e for normal glomerular filtration, the localization pattern of myo1e in glomeruli, and a prior proteomic study that detected myo1e in the slit diaphragm fraction (43) led us to investigate a potential association between myo1e and the slit diaphragm complexes.

Using subcellular fractionation of glomerular lysates, we determined that myo1e was associated with the slit diaphragm complexes. Immunoelectron microscopy of renal glomeruli indicated that myo1e was located along the plasma membrane in the foot processes of podocytes in regions adjacent to the slit diaphragms. These findings suggest that myo1e may interact with the protein components of the slit diaphragm complex.

Myo1e was also found to colocalize with the slit diaphragm marker ZO-1 in cryosections of mouse kidneys. Whereas myo1e was enriched in the regions corresponding to the slit diaphragms, a significant amount of myo1e staining was detected in the cell bodies of podocytes as well as in other regions distinct from the slit diaphragms or cell-cell junctions (Figs. 2 and 3). Thus, in addition to being highly concentrated in the slit diaphragm region, myo1e may be associated with other intracellular structures in podocytes, such as endocytic vesicles and cell-substrate adhesions (27, 50). Intriguingly, during glomerular maturation, myo1e was highly enriched in the basal region of developing podocytes, overlapping with ZO-1, which marks developing slit diaphragm complexes. This observation suggested to us that myo1e may have an important function during formation or remodeling of cell-cell junctions between podocytes.

Live-cell imaging and immunostaining of podocytes in culture revealed the presence of myo1e in cell-cell contacts; this is consistent with the previous observations of actin-binding proteins localizing to cell-cell junctions in podocytes (2, 16, 47). During contact reassembly, GFP-tagged myo1e was recruited to the newly formed cell-cell junctions, where it colocalized with the actin filaments assembling along the cell-cell boundary. The precise molecular composition of these newly formed cell adhesions in podocytes is unknown, but immunostaining of cells during recovery from calcium removal indicated that ZO-1 was present at these nascent junctions (data not shown). The contacts formed during the initial recovery from calcium removal may be similar to the nascent adherens junctions (also called punctum adherens), which contain ZO-1 (55). Since cell-cell junctions formed by podocytes in culture may lack some of the slit diaphragm proteins that are expressed in intact glomeruli but downregulated in cultured podocytes (9), the precise relationship between these in vitro junctions and slit diaphragms formed in vivo is unclear, although they do share some of the protein components, such as ZO-1.

Myo1e colocalization with the newly assembled actin structures in cell-cell junctions was disrupted in the presence of the disease-associated motor domain mutation (A159P). Since myosin motor domain is responsible for myosin-actin interactions, the lack of colocalization with the actin punctae during contact assembly may reflect the loss of actin binding in the A159P mutant. Alternatively, the abnormal localization of the A159P mutant of myo1e may be caused by the misfolding of the mutant protein.

Myo1e-knockout podocytes infected with wild-type myo1e assembled two types of actin structures at the nascent junctions: radial actin spokes and circumferential actin filament cables along the cell-cell boundary. We have not been able to directly observe the transitions between the two types of actin filament structures, and therefore, we cannot postulate whether the presence of these structures corresponds to distinct stages of contact assembly. However, a similar transition from radial to circumferential actin organization (along with the transition from vinculin-enriched, radially organized tight and adherens junctions to linear junctions depleted of vinculin) has been observed in a calcium switch assay in MDCK cells (53). Myo1e was preferentially localized along the cell-cell junctions so that the sites of myo1e enrichment appeared to provide the framework for localizing and/or stabilizing circumferential actin cables. Cells expressing the mutant form of myo1e did not assemble circumferential actin cables along the cell-cell boundary as efficiently as the cells expressing wild-type myo1e. On the other hand, no overt disruption of actin filament assembly was observed using phalloidin staining of myo1e-knockout podocytes at steady state (data not shown). Thus, live cell observations suggest that myo1e may contribute to actin reorganization during highly dynamic processes such as formation or remodeling of cell-cell contacts, including transitions from actin punctae or spokes to circumferential actin cables.

Further mapping of the domains determining myo1e localization to the junctions was performed using MDCK cells. The mapping studies provided additional insight into the role of the various functional domains in myo1e localization to cell-cell junctions; however, these in vitro studies using MDCK cells may not fully reflect the localization of myo1e in vivo. The SH3 domain of myo1e was dispensable for myo1e localization to the cell-cell junctions. On the other hand, TH1 and TH2 domains of myo1e were necessary for junctional localization. The motor domain of myo1e was not required for localization of myo1e to cell-cell junctions since myo1e tail was able to localize to cell-cell contacts in both MDCK cells and podocytes. Moreover, we have observed that the tail domain was more highly enriched in the cell-cell contacts than the full-length myo1e (Fig. 7C). This observation is consistent with a previous study showing that the tail of Dictyostelium myosin IB (a long-tailed class I myosin, similar to myo1e in terms of tail domain organization) is highly enriched at the plasma membrane (3.6-fold enrichment) vs. the full-length myosin IB being moderately enriched at the plasma membrane (1.7-fold enrichment) (5). The difference in localization between the full-length and tail-only constructs of myosin IB in Dictyostelium appears to be due to the interaction of the full-length myosin IB with actin filaments in the cytoplasm, which increases the cytoplasmic (nonmembrane) fraction of myosin (5).

This observation is in contrast to the observed defects in the localization of the myo1eA159P mutant during live-cell imaging of podocytes. Thus, the A159P mutation may cause a more severe defect in myo1e localization compared with the deletion of the motor domain, possibly because of the misfolding of the mutant protein. Therefore, the A159P mutant of myo1e may be unable to interact with either the junctional complexes or the actin filaments. On the other hand, in the absence of the motor domain, myo1e tail retains its ability to interact with the junctional proteins and the plasma membrane.

How could the various functional domains of myo1e contribute to its localization to cell-cell junctions? Since the deletion of the proline-rich TH2 domain was sufficient to abolish myosin localization to cell junctions, this region may interact with myo1e binding partners within the junctional protein complexes; these partners may include SH3 domain-containing proteins that remain to be identified. Alternatively, TH2 domain may include an ATP-insensitive actin-binding site similar to those identified in the TH2 domains of amoeboid class I myosins (22, 45), which may promote interactions with the actin filaments in junctional complexes. TH1 domain of myo1e interacts with anionic phospholipids with high affinity, with some preference for PtdIns(4,5)P2 and PtdIns(3,4,5)P3 (14). Both PtdIns(4,5)P2 and PtdIns(3,4,5)P3 are enriched at the plasma membrane of MDCK cells, where they play important roles in the establishment of epithelial cell polarity and epithelial morphogenesis (36). Hydrolysis of PtdIns(4,5)P2 by bacterial phosphatase SigD from Salmonella results in the loss of junctional integrity, redistribution of ZO-1, and reorganization of junctional actin filaments in intestinal epithelial cells (37), indicating that phosphoinositides play key roles in regulation of epithelial junctional stability. Thus, the TH1 domain binding to specific plasma membrane phospholipids together with TH2 domain interactions with proline-rich motif binding proteins or actin filaments may lead to the enrichment of myo1e in cell-cell junctions in the presence of both lipid- and protein-based signals for the slit diaphragm assembly.

Finally, we sought to identify junctional proteins that interact with myo1e SH3 domain. ZO-1, a known component of the slit diaphragm complex, interacted with myo1e SH3 domain in a pulldown assay. This interaction was mapped to the proline-rich COOH-terminal portion of ZO-1. We hypothesize that myo1e may help recruit ZO-1 to cell-cell junctions via SH3 domain-proline-rich motif interactions. Previous studies have implicated ZO-1 in regulation of actin organization in cell-cell junctions in epithelial cells (13). Thus, ZO-1 and myo1e may act together to promote coordinated assembly of junctional complexes and the actin-based structural elements that support them.

We hypothesize that the TH1 and TH2 domains together provide signals for targeting myo1e to the junctions, whereas myo1e motor and SH3 domains may serve as effector domains, recruiting additional proteins such as actin and ZO-1, to nascent contacts (Fig. 8).

Fig. 8.

Model describing how myo1e may function in cell-cell junctions. A: myo1e localization to the junctions requires multiple domain interactions. Interactions of the TH1 domain with the plasma membrane lipids and binding of the proline-rich TH2 domain to junction-associated proteins target myo1e to the junctions. B: motor domain of myo1e interacting with actin filaments and SH3 domain interacting with proline-rich proteins, such as ZO-1, promote coordinated assembly of cell-cell junctions and junctional actin bundles.

This study identified myo1e, a class I myosin, as a component of cell-cell junctions. Other members of the myosin superfamily have been implicated in the regulation of cell-cell contact organization (33). Nonmuscle myosin II, which is associated with the contractile actin filament bundles that support adherens junctions, regulates junction assembly, organization, and stability (10, 20, 48). Other myosins that localize to cell-cell contacts in epithelial cells and regulate junction assembly include members of myosin classes VI, VII, IX, and X (6, 11, 19, 34, 35, 41, 54). Unlike other myosins associated with the cell-cell junctions, class I myosins are unlikely to be involved in the long-range transport of vesicles or proteins along actin filaments due to their lack of processivity. Instead, class I myosins such as myo1e may function as dynamic linkers between the plasma membrane lipids, protein components of cell-cell junctions, and actin filaments (38). Intriguingly, another class I myosin, myo1c, has been implicated in the targeting of the podocyte junctional protein Neph1 to the plasma membrane (2), highlighting the importance of class I myosins in podocyte functions. The ability of class I myosins to interact with both lipid and protein binding partners places these motor proteins in the key position to be able to regulate renal filtration of proteins via modulation of specialized cell-cell contacts between podocytes.


This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Awards 1-R01-DK-083345 (to M. Krendel) and RO1-DK-061397 (to A. S. Fanning).


No conflicts of interest, financial or otherwise, are declared by the authors.


J.B., S.E.C., C.D.P., H.K., and M.K. performed the experiments; J.B. and M.K. analyzed the data; J.B., S.E.C., C.D.P., H.K., A.S.F., and M.K. interpreted the results of the experiments; J.B., H.K., and M.K. prepared the figures; J.B. and M.K. drafted the manuscript; J.B., S.E.C., C.D.P., H.K., A.S.F., and M.K. edited and revised the manuscript; J.B., S.E.C., C.D.P., H.K., A.S.F., and M.K. approved the final version of the manuscript; H.K., A.S.F., and M.K. contributed to the conception and design of the research.


Myo1e construct lacking TH2 was provided by J. Ouderkirk. We are grateful to the members of Krendel laboratory and to V. Sirotkin, D. Pruyne, and J. Amack for the critical discussion of the manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 9.
  9. 10.
  10. 11.
  11. 12.
  12. 13.
  13. 14.
  14. 15.
  15. 16.
  16. 17.
  17. 18.
  18. 19.
  19. 20.
  20. 21.
  21. 22.
  22. 23.
  23. 24.
  24. 25.
  25. 26.
  26. 27.
  27. 28.
  28. 29.
  29. 30.
  30. 31.
  31. 32.
  32. 33.
  33. 34.
  34. 35.
  35. 36.
  36. 37.
  37. 38.
  38. 39.
  39. 40.
  40. 41.
  41. 42.
  42. 43.
  43. 44.
  44. 45.
  45. 46.
  46. 47.
  47. 48.
  48. 49.
  49. 50.
  50. 51.
  51. 52.
  52. 53.
  53. 54.
  54. 55.
  55. 56.
View Abstract