An Essential Role of the Universal Polarity Protein, aPKCλ, on the Maintenance of Podocyte Slit Diaphragms

Glomerular visceral epithelial cells (podocytes) contain interdigitated processes that form specialized intercellular junctions, termed slit diaphragms, which provide a selective filtration barrier in the renal glomerulus. Analyses of disease-causing mutations in familial nephrotic syndromes and targeted mutagenesis in mice have revealed critical roles of several proteins in the assembly of slit diaphragms. The nephrin–podocin complex is the main constituent of slit diaphragms. However, the molecular mechanisms regulating these proteins to maintain the slit diaphragms are still largely unknown. Here, we demonstrate that the PAR3–atypical protein kinase C (aPKC)–PAR6β cell polarity proteins co-localize to the slit diaphragms with nephrin. Furthermore, selective depletion of aPKCλ in mouse podocytes results in the disassembly of slit diaphragms, a disturbance in apico-basal cell polarity, and focal segmental glomerulosclerosis (FSGS). The aPKC–PAR3 complex associates with the nephrin–podocin complex in podocytes through direct interaction between PAR3 and nephrin, and the kinase activity of aPKC is required for the appropriate distribution of nephrin and podocin in podocytes. These observations not only establish a critical function of the polarity proteins in the maintenance of slit diaphragms, but also imply their potential involvement in renal failure in FSGS.


Introduction
Glomerular diseases remain the major cause of chronic and end-stage renal disease, and the number of patients with glomerular diseases is increasing [1,2]. Because most glomerular diseases involve dysfunction of podocytes along with disassembly of slit diaphragms [1][2][3][4][5], it is necessary to understand the molecular basis for the maintenance of slit diaphragms. Mutations affecting several proteins in slit diaphragms, including nephrin and podocin (encoded by NPHS1 and NPHS2, respectively), lead to glomerular disease owing to disruption of slit diaphragms [2,[5][6][7]. However, the molecular mechanisms to regulate these proteins are unclear.
The slit diaphragms share several features with the epithelial tight junctions. First, both structures are specialized cell-cell junctions critical for selective paracellular permeability [4,[8][9][10][11]. Second, both structures are formed at the border of apical and basolateral membrane domains and are reported to play a role in regulation of apico-basal cell polarity [3,4,10,12]. Third, both structures are supported by common cytoplasmic proteins, including ZO-1, MAGI-1, and actin [13][14][15]. An evolutionarily conserved protein complex consisting of the serine/threonine protein kinase, aPKC, and two PDZ domain-containing scaffold proteins, PAR3 and PAR6, has been found to localize at the tight junctions and regulate the formation, and apico-basal epithelial polarity [10,16,17]. It has been revealed that PAR3 serves as a scaffold to associate aPKC with junctional proteins, and PAR6 regulates the kinase activity of aPKC [16,17]. Currently, there is convincing evidence that many cell polarity events are commonly regulated by this complex [16,17], although the functions of the aPKC-PAR complex in podocytes have not been clarified.

Results
The aPKC-PAR polarity proteins localize to slit diaphragms Based on the analogy of the slit diaphragms with tight junctions, we hypothesized that the aPKC-PAR complex plays a critical role in podocyte slit diaphragms. Consistent with this hypothesis, staining for aPKC, activated phospho-aPKC, PAR6b, and PAR3 revealed that these proteins are strongly expressed in podocytes surrounding the glomerular capillary tufts ( Figure 1A) and that they co-localize with nephrin ( Figure 1B). Electron microscopy revealed that PAR3 localizes at the cytoplasmic face of the slit diaphragms ( Figure 1C). These observations suggest a role of aPKC-PAR complex in the maintenance of structure and function of slit diaphragms.
Podocyte-specific aPKCl resulted in focal segmental glomerulosclerosis with disassembly of slit diaphragms To determine whether the aPKC-PAR complex is critical for slit diaphragms, we deleted aPKCl selectively from podocytes in mice. We deleted exon 5 of the loxP-flanked aPKCl gene (Prkci -Mouse Genome Informatics) using the Cre transgene driven by the podocyte-specific nephrin promoter [18,19] (aPkcl DE5/floxE5 ; Nphs1-Cre Tg mutant mice) (Supplemental Figure S1). Selective loss of aPKCl protein in the podocytes at postnatal day 0 (P0) was confirmed by double immunofluorescence for aPKCl and WT1, a podocyte-specific transcription factor [20] (Figure 2A). Mice of all genotypes were born at the expected Mendelian frequency (data not shown). Although the gross appearance of mutant kidneys was not significantly different from that of control kidneys at P0 ( Figure 2B), the mutant mice exhibited nephrotic-level proteinuria on P0 ( Figure 2C) and throughout their lives (data not shown). By 3 weeks of age, mutant mice developed severe renal dysfunction, including a significantly increased serum creatinine level and blood urea nitrogen level compared with control mice ( Figure 2D and not shown). By 4 weeks of age, mutant mice showed growth retardation ( Figure 2E), resulting in a drastically reduced life span with a median age of death of 6 weeks ( Figure 2F). These observations indicate that aPKCl plays a critical role in podocytes in supporting appropriate renal function.
Histopathologic analyses of the mutant kidneys revealed the development of progressive glomerulosclerosis (Figure 3). At P0, the mutant kidneys showed normal nephrogenesis and glomerulogenesis with normal staining patterns of nephrin and synaptopodin [21], an actin-associating protein, in podocyte foot processes. At P10, however, the mutant glomeruli demonstrated partial detachment of podocytes, adhesion of glomeruli to Bowman's capsules, mesangial expansion, dilated capillaries, and an irregular pattern of nephrin staining. At P21, the mutant glomeruli demonstrated more advanced glomerular damage with characteristic segmental to global sclerotic lesions. These histopathologic observations are consistent with the development of severe renal dysfunction in mutant mice by the age of 3 weeks. Together, these are features characteristic of focal segmental glomerulosclerosis.
To further analyze the onset of the glomerular damages, we examined the mutant podocytes at an ultrastructural level ( Figure 4A). At P0, fine foot processes were embedded in the glomerular basement membrane, and in particular, normal slit diaphragms were formed between the foot processes. These observations, however, do not exclude the possibility that aPKCl is required for the development of slit diaphragms, because the activation of the nephrin promoter is concomitant with the development of slit diaphragms [22] and Nphs1-Cre-mediated recombination cannot be accomplished before the onset of slit diaphragm formation (data not shown). At P7-10, however, we observed apically dislocated slit diaphragms and foot process effacement. Moreover, podocalyxin, a marker protein of the apical domains of podocytes [23], localized not only in the apical but also in the basal domains, whereas ZO-1 localized at the irregular cellcell junctions ( Figure 4B), indicating disturbance in the apico-basal cell polarity. Altogether, these observations indicate that aPKCl is necessary for the maintenance of slit diaphragm integrity and apico-basal cell polarity in podocytes. This implies that the disturbance in apico-basal cell polarity in podocytes can cause disturbance of the slit diaphragms resulting in glomerulosclerosis.

aPKC-PAR3 complex associates with the nephrinpodocin complex
To further explore the molecular mechanisms, by which the aPKC-PAR complex organizes slit diaphragms, we examined the potential association between the aPKC-PAR complex and nephrin in podocytes. First, isolated rat renal tubules (Tub) and glomeruli (Gl) were extracted sequentially under different conditions in detergents (see Methods), and analyzed the distribution of each protein by immunoblotting ( Figure 5A). Consistent with previous reports [14,24,25], nephrin and podocin were concentrated in the Gl-S2 and Gl-P2 fractions, which contain proteins associated with lipid raft microdomains and actin cytoskeletons, respectively. Similarly, aPKCl from glomeruli, but not from renal tubules, was efficiently concentrated to the Gl-S2 and Gl-P2 fractions, suggesting that a significant amount of aPKCl in glomeruli codistributes with nephrin and podocin. Next, we used immunoprecipitation to examine whether endogenous PAR3 forms a protein complex with nephrin and podocin. As demonstrated previously [26], aPKCl was reproducibly coprecipitated with PAR3 from both of the soluble fractions extracted from glomeruli ( Figure 5B). Furthermore, nephrin coprecipitated with PAR3 from both fractions. Importantly, podocin consistently coprecipitated with PAR3 from the Gl-S2 fraction. These data indicate that PAR3 forms a complex not only with aPKCl but also with the nephrin-podocin complex in vivo. This suggests that PAR3 can serve as a scaffold to associate aPKC and the nephrin-podocin complex to organize slit diaphragms.
To examine the region of PAR3 that is responsible for the formation of the complex with nephrin, various PAR3 mutants ( Figure 6A) were incubated with immobilized glutathione S-transferase fused with the cytoplasmic tail of nephrin (GST-nephrinICD). T7-tagged PAR3 mutants overexpressed in 293T cells specifically interacted with GST-nephrinICD, except for the mutants lacking all PDZ domains or mutated in the third PDZ domain [27,28], indicating the third PDZ domain is required for the formation of the complex with nephrin ( Figure 6B). Endogenous aPKC in 293T cells also interacted with GST-nephrinICD in a PAR3-dependent manner, confirming the formation of a ternary complex of PAR3, aPKC, and nephrin.
We next addressed the possibility of direct interaction between PAR3 and nephrin in vitro. Purified GST-nephrinICD2 precipi-tated the three PDZ domains of PAR3 fused with maltose binding protein (MBP-PDZ123) ( Figure 6, A and C). This indicates that the cytoplasmic tail of nephrin can bind directly to the PDZ domains of PAR3. Although GST-nephrinICD2 can bind with the isolated first or third PDZ domains of PAR3, all three PDZ domains are required for efficient binding. To identify the PAR3binding region in nephrin, we examined the binding ability of various deletion fragments of nephrin ( Figure 6A) with immobilized MBP-PDZ123. We found that deletion of the carboxyterminal 11 amino acids of nephrin (ICD2DC) reduced their binding affinity for MBP-PDZ123 and that these 11 amino acids (CT11) can bind to MBP-PDZ123 ( Figure 6D). This indicates that CT11 partly mediates the interaction between nephrin with PAR3, and is consistent with the sequence similarity of CT11 to the type II PDZ-binding motif [29,30]. Furthermore, a region encompassing amino acids 1134-1164 of nephrin (ICD-A) is enough for binding with MBP-PDZ123 ( Figure 6D). Altogether, these data suggest that the aPKC-PAR complex interacts with the nephrin-podocin complex by direct binding between the PDZ domains of PAR3 and two cytoplasmic regions of nephrin, including the CT11 and ICD-A regions.
aPKC activity is required for the appropriate distribution of nephrin and podocin in podocytes To determine whether aPKC activity affects the distribution of nephrin and podocin, we treated the isolated glomeruli with or without an aPKC-specific inhibitor [31,32], and analyzed the distribution of nephrin and podocin in sequentially extracted fractions using 0.1% TritonX-100 (S1) and RIPA buffer (S2) ( Figure 7A). Inhibition of aPKC significantly decreased the amount of nephrin in the highly soluble fraction (S1) within 30 min. Contrastingly, the amount of nephrin was significantly increased in the RIPA-soluble (S2) and -insoluble (P2) fractions. The distribution of podocin in the S1 and S2 fractions was similar to that of nephrin. These data suggest that the inhibition of aPKC augments the association of nephrin and podocin with the lipid raft microdomains and actin cytoskeletons.  The distribution of PAR3, aPKCl, nephrin, and podocin was analyzed in differentially extracted fractions of rat renal tubules or glomeruli (Tub, Gl). Tub-S1, Gl-S1, highly soluble fractions extracted with 0.1% TritonX-100; Tub-S2, Gl-S2, relatively soluble fractions extracted with 1.0% TritonX-100 and 20 mM CHAPS; Tub-P2, Gl-P2, insoluble fractions. The blotted filter with anti-nephrin antibody was stripped and reprobed with anti-PAR3 antibody. (B) Immunoprecipitation for PAR3 was carried out on both glomerular fractions (Gl-S1, Gl-S2) using two independent anti-PAR3 antibodies (UBI, C2-2AP) and the immunoprecipitates were analyzed by immunoblotting. The blotted filter with anti-nephrin antibody was stripped and reprobed with anti-PAR3 antibody. doi:10.1371/journal.pone.0004194.g005 We further assessed whether the inhibition of aPKC affects the formation of the nephrin-podocin complex in rat glomeruli ( Figure 7B). Immunoprecipitation of podocin from the S1 and S2 fractions of the control glomeruli revealed that nephrin was more effectively coprecipitated from the S2 fraction as compared to the S1 fraction. This is consistent with previous reports showing that the functional nephrin-podocin complex is formed in the lipid raft microdomains [25,33]. Although aPKC inhibition significantly increased the amount of both nephrin and podocin in the S2 fraction, the amount of coprecipitated nephrin was hardly increased. These data suggest that aPKC activity is required in order to prevent the excessive accumulation of nephrin and podocin, which do not form a complex in the lipid raft microdomains. The excessive accumulation of nephrin in the RIPA-insoluble (P2) fraction might lead to the formation of abnormal aggregations of nephrin, which in turn might strongly associate with actin cytoskeletons, because the forced aggregation of nephrin is reported to induce localized actin polymerization through direct binding to Nck [34,35].

Discussion
Podocytes are highly polarized cells with apical, basal, and slit diaphragm domains [3,4]. Because each membrane domain plays a critical role in the selective filtration barrier, and organizing the glomerular basement membrane so that foot processes can anchor in, the regulation of cell polarity is considered to be important for podocytes [1,3,4]. We have provided direct evidence that the cell polarity protein aPKCl is critical for the maintenance of the slit diaphragms. Our data suggest that the aPKC-PAR complex establishes apico-basal cell polarity in podocytes and appropriately arranges the slit diaphragms to form the filtration barrier. Our data further suggest that the association of the aPKC-PAR3 complex with the nephrin-podocin complex regulates the state of equilibrium in which nephrin and podocin are appropriately distributed among raft and non-raft microdomains to avoid unnecessary aggregation of nephrin. Because it is recently reported that PAR3 form a protein complex with Neph1 [36], another nephrin-binding transmembrane protein in slit diaphragms [37], PAR3-Neph1 complex may influence this equilibrium. In the regulation of actin cytoskeletons, the direct binding of the aPKC-PAR3 complex to the cytoplasmic tail of nephrin may compete with Nck and prevent the rearrangement of actin cytoskeletons in podocytes to maintain the integrity of the slit diaphragms [34,35]. Our observations provide a new pathophysiologic mechanism whereby any condition disturbing aPKC activity can lead to defective organization of membrane domains, damage podocytes, and result in FSGS.

Animals
We bred aPkcl DE5/+ mice [18] with Nphs1-Cre Tg mice [19] to generate aPkcl DE5/+ ; Nphs1-Cre Tg mice, and bred these mice with homozygous floxed aPkcl (aPkcl floxE5/floxE5 ) mice [18] to obtain aPkcl DE5/floxE5 ; Nphs1-Cre Tg mutant mice. The Rosa26R strain was purchased from the Jackson Laboratory. Serum creatinine levels were measured using a Jaffé assay kit (Wako Pure Chemical) according to the manufacturer's instructions. Male Wistar rats (220-240 g, 7 weeks old) were purchased from Charles River Japan. All animal experimentations were conducted in accordance with the Guidelines for Proper Conduct of Animal Experiments (Science Council of Japan), and all protocols were approved by our institutional review boards.

Histology and immunostaining
Two-micrometer-thick paraffin sections of 4% PFA-fixed kidney were stained with periodic acid-Schiff (PAS) and hematoxylin (Muto Chemical). For Immunostaining, the sections were autoclaved in target retrieval solution (Dako S3308 or Dako S1700, for 30 min at 90uC and for 10 min at 121uC, respectively), and processed for immunostaining as described previously [38]. The sections were examined and photographed with a DMR microscope (Leica) equipped with a Pro600ES color CCD camera (Pixera) or with a BX50 epifluorescence microscope (Olympus) equipped with a SenSys CCD camera (Photometrics). All images were arranged and labeled using Photoshop 5.5 (Adobe Systems).

Transmission electron microscopy
The mice were perfused with 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer at pH 7.4 (PB). The kidneys were removed, cut into small pieces, and immersed in 2.5% glutaraldehyde containing 1% tannic acid in 0.1 M PB for 2 h at 4uC. They were then post-fixed with 1% OsO 4 , dehydrated and embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate and then examined under a JEM 1230 electron microscope (JEOL).

Immunogold labeling
Mouse kidneys were perfused with PLP fixative and immersed in the same fixative for 30 min at 4uC. The samples were rinsed with 5% sucrose for 30 min at 4uC. Tissue samples were then infiltrated with 40% polyvinylpyrrolidone/2.3 M sucrose buffered with 0.1 M PB, embedded on nails, and frozen quickly in liquid nitrogen. Ultrathin sections were cut with an Ultracut UCT equipped with an EM FCS cryoattachment (Leica) at 2110uC. Sections were transferred to Formvar-coated nickel grids (150 mesh). Subsequent incubation steps were performed by floating the grids on droplets of the filtered solution. Free aldehyde groups were quenched with PBS-0.01 M glycine, and the sections were incubated overnight with PBS containing 20% fetal bovine serum (FBS). Next, the grids were incubated with affinity-purified rabbit anti-podocalyxin antibody (1:200 dilution with PBS containing 20% FBS) and mouse anti-ZO-1 antibody (Zymed, 1:100 dilution) for overnight at 4uC. The grids were then incubated with antirabbit IgG coupled with 5 nm-gold (diluted 1:100) and anti-mouse IgG coupled with 10 nm-gold (diluted 1:100) for 1 h. After immunostaining, the samples were fixed with 2.5% glutaraldehyde buffered with 0.1 M PB (pH 7.4). The sections were then contrasted with 2% uranyl acetate solution for 20 min, and absorption-stained with 3% polyvinyl alcohol containing 0.2% uranyl acetate for 20 min. All sections were observed using a JEM 1230 electron microscope.

Statistical analysis
The two-tailed Mann-Whitney U-test (VassarStats, http:// faculty.vassar.edu/lowry/VassarStats.html) or the two-tailed Student's t-test (Microsoft Excel 2007) were used to analyze the differences between the pairs of groups. Values were regarded significant at p,0.05.