NEPH2 Is Located at the Glomerular Slit Diaphragm, Interacts with Nephrin and Is Cleaved from Podocytes by Metalloproteinases

Plasmids and Antibodies

Full-length nephrin, NEPH1, and NEPH2 cDNAs were recently described (4,7). The Fc fusion proteins contain the NEPH1, NEPH2, or nephrin sequences as indicated, followed by the CH2 and CH3 domain of human IgG. All truncations were generated by PCR and standard cloning procedures, and verified by automated sequencing. The domain structures of nephrin, NEPH1, and NEPH2 were predicted using SMART (http://smart.embl-heidelberg.de) and motif scanner (http://scansite.mit.edu). MT1-matrix metalloproteinase (MT1-MMP) cDNA was a kind gift from J. Weiss, University of Michigan (Ann Arbor, MI). The plasmids pMstGV (pGal4/VP16), pMstGV-APP695 (pAPP-Gal4/VP16), and pMstGV-LDL receptor (pLDLR-Gal4/VP16), and the pG5E1B-Luc plasmid, which contains five Gal4 binding sites and the E1B minimal promoter preceding the luciferase firefly gene, were generously provided by T. Südhof, University of Texas Southwestern Medical Center, (Dallas, TX). For construction of neph2-Gal4/VP16 and neph2-cyt-Gal4/VP16, the human full-length neph2 cDNA without stop codon or the cytoplasmic tail containing Kozak and first ATG 5′ of the sequence were PCR-amplified and cloned into the BglII and NheI sites of pMstGV as described previously (11). The resulting constructs encode fusion proteins containing the neph2 full-length protein or the cytoplasmic tail with a VP16 transactivation domain, a nuclear localization signal, and a Gal4 binding region. The human fetal libraries were purchased from Edge BioSystems (Gaithersburg, MD). For detection of human NEPH2, the 5′-oligonucleotide 5′-CCAACATCTACAGCACCCTGAGCGGCC and 3′-oligonucleotide 5′-TTAGACGTGAGTCTGCATCCG were used to yield a 311-bp PCR fragment. Polyclonal NEPH2 and nephrin antisera were generated by immunization of rabbits (Eurogentec, Herstal, Belgium). The following peptides were used: NEPH2.ec, MAKDKFRRMNEGQVY (human, mouse: aa34-48); and NEPH2.ic, SDPSRPLQRRMQTHV (human: aa764-778, mouse: aa752-766). The peptide used for nephrin was KDSRTVTESRLPQES (human: aa480-494). All antisera were affinity-purified against the peptides. Monoclonal anti-NEPH2 antibody was raised against the intracellular fragment of human NEPH2 (aa 593-778) and used for immunohistochemistry at 1:20 and for Western blot at 1:1000. M2 anti-FLAG antiserum was obtained from Sigma-Aldrich (Taufkirchen, Germany), and protein G-sepharose was obtained from Amersham Pharmacia Biotech (Freiburg, Germany). Anti-MT1-MMP was obtained from Chemicon International (Temecula, CA), and the secondary antisera were obtained from Jackson Immuno Research Laboratories (West Grove, PA).

Reporter Gene Assays

Reporter assays were performed as described previously (11). Luciferase gene expression was analyzed using the Luciferase Assay System (Promega, Mannheim, Germany). A 10-s integral measurement of light emission from each reaction mixture was recorded in a plate luminometer and normalized for β-galactosidase activity. All transfections for reporter gene assays were performed in triplicate and repeated in at least two independent experiments.

Cell Culture

Conditionally immortalized mouse podocytes were generated as described previously (12) and grown at permissive temperature in the presence of gamma IFN at 10 U/ml. To induce differentiation, the cells were maintained on type I collagen at 37°C without gamma IFN for at least 14 d. For shedding experiments, differentiated podocytes were incubated in medium containing the agents as depicted for 60 min at 37°C. GM6001 was obtained from Chemicon International; pervanadate was freshly prepared from sodium vanadate and H₂O₂ immediately before use.

Co-Immunoprecipitation

Co-immunoprecipitations were performed as described (13). Briefly, HEK 293T cells were transiently transfected using the calcium phosphate method. After incubation for 24 h, cells were washed twice and lysed in a 1% Triton X-100 lysis buffer (20 mM Tris-HCl, pH 7.5; 1% Triton X-100; 50 mM NaCl; 50 mM NaF; 15 mM Na₄P₂O₇; 0.1 mM EDTA; 2 mM Na₃VO₄; and protease inhibitors). After centrifugation at 15,000 × g (15 min, 4°C) and ultracentrifugation at 100,000 × g (30 min, 4°C), cell lysates containing equal amounts of total protein were precleared with protein G-sepharose, and then incubated for 1 h at 4°C with the appropriate antibody, followed by incubation with 30 μl of protein G-sepharose beads for 3 h. The beads were washed extensively with lysis buffer, and bound proteins were resolved by 10% SDS-PAGE. For endogenous interaction, eight mouse kidneys were homogenized using a glass potter, cleared by centrifugation, and solubilized in lysis buffer supplemented with 20 mM CHAPS and 3 mM ATP. Before immunoprecipitation, cellular lysates were further precleared by ultracentrifugation and absorption to protein G beads. All tissues were freshly prepared and perfused in situ with ice-cold PBS before lysis. Control samples were incubated with equal amounts of rabbit IgG followed by protein A beads. Human kidney lysates were prepared from the healthy pole of tumor nephrectomies.

Immunohistochemistry

Mouse kidneys were subject to retrograde perfusion fixation using 3% glutaraldehyde and embedded in paraffin; 4-μm sections were cut and deparaffinized. After standard immunohistological procedures, including antigen retrieval with proteinase K, the sections were incubated with anti-NEPH2 antiserum or mouse serum (1:1000), followed by goat anti-mouse rhodamine red. For injection studies, sera were heat-inactivated at 56°C for 30 min. Adult male C57Bl/6 mice were injected with 200 μl intravenously of either anti-NEPH2 antiserum or pre-immune serum and euthanized 24 h later. Kidneys were fixed and embedded as described, and sections were incubated with goat anti-rabbit rhodamine red. An Axiophot 2 microscope (Zeiss, Jena, Germany) equipped with a CCD camera was used for image documentation.

Electron Microscopy

Adult mice were anesthetized with sodium pentobarbital and transcardially perfused using 4% paraformaldehyde and 0.1% glutaraldehyde in PBS. Kidneys were removed and postfixed in the same fixative (overnight, 4°C). Tissues were washed in PBS, and then horizontal sections (50 μm) were cut on a vibratome and cryoprotected in a solution containing 25% sucrose and 10% glycerol in 50 mM PBS. The sections were freeze-thawed and incubated in blocking solution containing 20% normal goat serum in 50 mM Tris-buffered saline for 1 h, followed by incubation with an anti-Neph2.ec antiserum (overnight, 4°C). After washing, the sections were with incubated with 1.4-nm gold-coupled goat anti-rabbit secondary antibody (1:100, Nanogold; Nanoprobes, Stony Brook, NY) for immunogold reaction. Immunogold labeling was then enhanced with HQ silver kit (Nanoprobes). After treatment with OsO₄, the sections were stained with uranyl acetate, dehydrated, and flat-embedded in epoxy resin (Durcupan ACM, Fluka; Sigma-Aldrich, Gillingham, UK). Ultra-thin sections were cut and analyzed in a Philipps CM 100 electron microscope.

Urine Preparation

Eight ml urine per sample were cleared from cell debris and crystals by centrifugation and diluted in acetone to achieve an 80% acetone solution. After incubation o.n. at −20°C, the solutions were centrifuged at 10,000 × g (10 min, 4°C). Supernatants were removed and pallets washed with 100% acetone (−20°C). Dry pallets were weighted and resuspended in Laemmli buffer at 1 μg/μl.

Software

Alignments and phylogenetic trees were created using Lasergene 5.51 (DNASTAR Inc, Madison, WI).

Sequence Comparison of the NEPH Gene Family

Human NEPH proteins are type 1 transmembrane proteins characterized by five extracellular Ig domains. Searching GenBank, we found orthologs in a variety of species (Figure 1A). These included well-characterized Drosophila proteins as well as predicted unnamed proteins in zebrafish or Xenopus. Phylogenetic analysis reveals a close relation between the mammalian NEPHs, with >97% identity of human and mouse NEPH2 (Figure 1B). We therefore decided to use peptide sequences unique to NEPH2 but common to the human and the mouse protein as epitopes for the generation of polyclonal antisera. The same approach was used for anti-nephrin antiserum. Fusion proteins used for NEPH2 protein interaction studies are shown in Figure 1C.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

(A) Phylogenetic tree of NEPH proteins. GenBank was searched for NEPH2-related proteins. We use the term NEPH for unnamed proteins with high homology to human NEPHs. GenBank accession numbers are: human NEPH1, CAH73881, human NEPH2, NP_115920; human NEPH3, NP_954649; mouse NEPH1, AAN73043; mouse NEPH2, BAC32333; mouse NEPH3, NP_766486; rat NEPH1, AAP78673; rat NEPH2, XP_235986; rat NEPH3, XP_218486; Drosophila melanogaster kirre, AAF86308; Drosophila melanogaster irreC-rst, Q08180; anopheles gambiae kirre, XP_310916; Xenopus laevis NEPH, AAH57728; SYG-1, AAC47047; danio rerio NEPH, AAH66766. (B) Homology with human NEPH2 of selected Ig family proteins. Data were obtained using Clustal V (Megalign, DNASTAR). A high degree of similarity is seen among mammalian NEPH2 orthologs and between NEPH2 and NEPH1. (C) NEPH2 fusion proteins used for interaction studies. The Ig domains were predicted using the SMART prediction program (http://smart.EMBL-Heidelberg.de/). All constructs include the authentic signal peptide (SP). NEPH2.ec.Fc contains the entire extracellular (ec) domain of NEPH2 fused to the carboxy-terminus of the CH2 and CH3 domain of human IgG (Fc).

Characterization of Anti-NEPH2 Antisera

Antisera directed against the peptides NEPH2.ec and NEPH2.ic were tested in HEK 293T cells. The localization of the peptide sequences within the NEPH2 molecule is depicted in Figure 2A. We used lysates from HEK 293T cells transiently overexpressing human NEPH2 and nephrin, respectively, to test the antisera by Western blot analysis. Cells overexpressing NEPH1 served as negative control. As demonstrated in Figure 2B, both anti-NEPH2 antisera recognized a single band at approximately 95 kD, which was not detected in the control lysates. Similarly, anti-nephrin detected a band at 175 kD, the expected size of nephrin without apparent cross-reactivity. Next, we tested if the antisera were suitable for immunoprecipitation. Flag-tagged NEPH2 and nephrin, respectively, were transiently transfected. Five μg of each antibody coupled to protein A beads precipitated the overexpressed proteins as detected by anti-flag on Western blot (Figure 2, C and D). In addition, we raised a monoclonal antibody against the intracellular domain of NEPH2. On Western blot, the antibody specifically detected the flag-tagged NEPH2 C-terminus in lysates from overexpressing HEK 293T cells (Figure 2E).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Generation of antisera against NEPH2 and nephrin. (A) Amino acid sequences and location within the NEPH2 protein of peptides used for rabbit immunization. NEPH2.ec is located near the N-terminal signaling peptide, and NEPH2.ic is close to the C-terminal PDZ binding domain. (B) NEPH2, NEPH1, and nephrin were transiently overexpressed in HEK 293T cells. Lysates were separated on SDS-PAGE and Western blot analysis performed using affinity purified rabbit antisera directed against NEPH2.ec, NEPH2.ic, and the N-terminus of nephrin (anti-nephrin.ec). Bands in the expected region are visible in the NEPH2 and nephrin lanes, but not in the NEPH1 lanes (negative control). (C) Anti-NEPH2.ic and anti-NEPH2.ec both precipitate NEPH2.F from transiently transfected HEK 293T cells. Lysates were subject to immunoprecipitation with 5 μg anti-NEPH2.ec and anti-NEPH2.ic, respectively, Western blot analysis with anti-M2. Note that NEPH2.F shows up as a double band on Western blot, possibly a result of differential posttranslational modifications. (D) Anti-nephrin.ec immobilizes nephrin from transiently transfected HEK 293T cells. 5 μg anti-nephrin.ec antibody bound to protein A beads were used followed by Western blot analysis performed with anti-M2. (E) Characterization of anti-NEPH2 mAb. The flag-tagged C-terminal domain of NEPH2 (F.NEPH2 cyt) was transiently overexpressed in HEK 293T cells. Lysated were separated on SDS-PAGE. Anti-NEPH2 mAb recognizes a band in the correct region (right lane).

NEPH2 Is Expressed in Various Mouse Tissues

Western blot technique was used to evaluate the tissue expression of mouse NEPH2. Both anti-NEPH2.ic and anti-NEPH2.ec antiserum detected a band at approximately 125 kD in mouse podocyte lysates (Figure 3A). This band was not seen in Western blots reprobed with pre-immune serum. Because no significant immunoreactivity was detectable in mouse tissue lysates, we immunoprecipitated NEPH2 using anti-NEPH2.ic for immobilization and anti-NEPH2.ec for immunoblotting. Whereas the predicted size of NEPH2 is 85.4 kD, tissues from heart, lung, liver, kidney, testis, brain, and muscle showed positive bands at approximately 100 kD (Figure 3B). Some tissues expressed an additional band at approximately 125 kD (heart, testis, kidney, lung). The strongest expression was seen in muscle and brain. These data suggest that NEPH2 is alternatively spliced, partially degraded, or differentially glycosylated in some tissues. To test for earlier expression, we performed NEPH2 PCR from human fetal libraries. As demonstrated in Figure 3C, there is moderate expression of NEPH2 mRNA in fetal human kidneys and high expression in fetal human brains.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Expression of NEPH2. (A) Mouse podocytes express NEPH2. Anti-NEPH2.ic antiserum or anti-NEPH2.ec antiserum were used to detect endogenous NEPH2. NEPH2 immunoreactivity is detected at approximately 125 kD by both antisera, but not by pre-immune serum. (B) NEPH2 is widely expressed in mouse tissues. Lysates were prepared from freshly explanted mouse organs and separated on SDS-PAGE. Immunoprecipitations were performed using the NEPH2.ic antibody coupled with protein A beads. NEPH2 expression was determined using the NEPH2.ec polyclonal antibody. Strong immunoreactivity is detected at approximately 100 kD in all tissues tested with strongest expression in muscle and brain. Note the additional bands at approximately 125 kD in some tissues. Specificity was confirmed by re-incubation of the blot with pre-immune serum (lower panel). (C) NEPH2 is expressed in fetal tissues. cDNA library PCR with intron-spanning primers detects a specific transcript for NEPH2 (311 bp) in human fetal kidney and human fetal brain.

NEPH2 Forms Homodimers and Heterodimers with Nephrin

Candidate extracellular ligands for NEPH2 were detected by immunoprecipitation in overexpressing HEK 293T cells. As depicted in Figure 4A, nephrin.ec.Fc, a construct containing the entire extracellular domain of nephrin, precipitated NEPH2. The NEPH1 extracellular domain, or a control construct containing parts of the PKD1 extracellular domain (PKD.Fc), however, did not precipitate NEPH2. As reported for the nephrin-NEPH1 interaction, the interaction between NEPH2 and nephrin was not confined to single Ig domains (Figure 4B). Several nephrin truncations precipitated NEPH2, whereas the control construct did not. Figure 4C demonstrates that NEPH2 forms homodimers; NEPH2.F is precipitated by the NEPH2 extracellular domain, but not by the control construct.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Homodimeric and heterodimeric interactions of NEPH2. (A) NEPH2 precipitates nephrin but not NEPH1. Fc-tagged extracellular domains of nephrin, NEPH1, or Ig domains of PKD1 (PKD.Fc) were co-expressed with NEPH2 in transiently transfected HEK 293T cells. Cellular lysates were incubated with protein G to precipitate the Fc fusion proteins. NEPH2, bound to nephrin.Fc, was detected by Western blot analysis, using the NEPH2.ec polyclonal antibody. In contrast, NEPH1.Fc and control protein PKD.Fc failed to immobilize NEPH2. (B) Nephrin fusion proteins precipitate NEPH2. (B) Nephrin fusion proteins containing either the entire extracellular domain of nephrin (nephrin.Fc), the nephrin Ig domains 1 to 3 (nephrin.Ig1-3.Fc), or Ig domains 5 and 6 (sFc.nephrin.Ig5 + 6) fused with Fc precipitated NEPH2, whereas PKD.Fc did not. Western blot analysis was performed with anti-NEPH2.ec. (C) NEPH2 fusion protein precipitates NEPH2. Co-expression of NEPH2.Fc, containing the entire extracellular domain of NEPH2, but not PKD.Fc immunoprecipitated flag-tagged NEPH2 (NEPH2.F). NEPH2.F, bound to NEPH2.ec.Fc, was detected by Western blot analysis, using the M2 anti-flag monoclonal antibody. (D) Interaction between NEPH2 and nephrin in vivo. Lysates prepared from human kidneys were incubated with anti-nephrin antiserum or rabbit IgG and subsequently precipitated with protein A. Immobilized NEPH2 was detected by Western blot analysis using anti-NEPH2.ec antiserum. Nephrin but not rabbit IgG immobilizes NEPH2, suggesting that nephrin and NEPH2 interact in vivo. Immunoprecipitation with anti-NEPH2.ec antiserum served as size control (right lane).

The in vivo interaction between nephrin and NEPH2 is shown in Figure 4D. Nephrin co-precipitated NEPH2 in human kidney lysates. Rabbit IgG, used as negative control, did not precipitate either protein.

NEPH2 Is Located at the Glomerular Slit Diaphragm

The interaction between NEPH2 and nephrin suggests that NEPH2 is expressed on podocytes. Accordingly, our monoclonal anti-NEPH2 antibody detected NEPH2 immunoreactivity in a glomerular staining pattern (Figure 5, A and B). Injection of anti-NEPH2.ec but not pre-immune serum showed staining in the glomerular basement membrane region (Figure 5C, D). Finally, immunogold electron microscopy revealed that NEPH2 is located at the glomerular slit diaphragm (Figure 5, E and F). Only minor additional immunoreactivity was detected at other locations of the podocyte foot processes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

(A, B) Localization of NEPH2 to the renal glomerulus. Anti-NEPH2 mAb specifically labels glomerular cells in mouse kidney sections (A), whereas mouse control serum does not (B, original magnification 200×). (C, D) Localization of NEPH2 near the glomerular basement membrane. Anti-rabbit IgG labels the glomerular basement membrane and some glomerular cells in mice injected with anti-NEPH2.ec antiserum (C), but not in mice injected with pre-immune serum from the same rabbit (D, original magnification 400×). (E, F) Immunogold localization of Neph2 to the slit diaphragm in adult mouse glomeruli. Pre-embedding silver-enhanced immunogold labeling using rabbit anti-Neph2.ec antibody. Neph2 is mostly localized to the slit diaphragm (arrows). FP indicates foot process; GBM, glomerular basement membrane; E, endothelial cell.

Shedding of the NEPH2 Extracellular Domain

Because shedding of nephrin and NEPH2 has been reported, we incubated supernatants of murine podocytes with anti-NEPH2.ec. Western blot analysis revealed a band at approximately 95 kD (Figure 6A). Incubation of the cells with the tyrosine phosphatase inhibitor pervanadate leads to an increase in soluble NEPH2, whereas the band vanished after incubation with the MMP inhibitor GM6001. These findings were confirmed in several independent experiments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Shedding of NEPH2. Shedding of NEPH2 in vitro. (A) Immortalized murine podocytes were incubated with agents as indicated for 60 min. Supernatants were then precleared and incubated with anti-NEPH2.ec coupled to protein A beads. Anti-NEPH2.ec immunoreactivity is detectable in supernatants of pervanadate exposed cells, and to a lesser extent of untreated cells. No immunoreactivity can be found in supernatants of cells exposed to GM6001. (B) Lysates obtained from murine podocytes were tested for MT1-MMP. Using specific polyclonal antiserum, intact MT1-MMP is detected at 63 kD, its inactivated form is detected at 42 kD. (C) NEPH2 was overexpressed with or without MT1-MMP in transiently transfected HEK 293T cells. Supernatants were incubated with anti-NEPH2.ec antibody bound to protein A beads. Precipitates were analyzed on Western blot using anti-NEPH2.ec antiserum. In the presence of MT1.MMP, a band of approximately 95 kD appears in supernatants of cells expressing both constructs. This band vanishes if either construct is missing. This band is not seen when using anti-NEPH2.ic antibody for Western blot analysis. Bcl2 is used as replacement for NEPH2 as a negative control. (D) Detection of the NEPH2 cytoplasmic fragment in a gene reporter assay. All constructs contained the transcription factor Gal4-VP16 near the C-terminus and were co-transfected with the pG5E1B-Luc plasmid. Full-length NEPH2 leads to a luciferase activity of 6000-fold, suggestive of shedding and release of the cytoplasmic NEPH2 fragment. Negative controls are the LDL receptor (LDLR, lane 4), or empty vector (lane 1). Positive controls are a cytoplasmic fragment of NEPH2 (lane 2), β-amyloid precursor protein (APP, lane 5) and a cytoplasmic construct (pMstGV, lane 6).

Next, we tested whether the podocyte cell line expressed the slit diaphragm protein MT1-MMP as a possible cleaving enzyme for the NEPH2 extracellular domain. Figure 6B demonstrates significant expression of the proteinase (63 kD) and its major degradation product (42 kD) in the cell lysates of cultured podocytes.

To further evaluate the possible role of MT1-MMP, we transiently overexpressed MT1-MMP and NEPH2 in HEK 293T cells. Figure 6C demonstrates the presence of a positive band in the supernatants of HEK 293 T cells when both NEPH2 and MT1-MMP were co-transfected. In the absence of MT1-MMP or expression of a control construct (Bcl2), no significant immunoreactivity was detectable. Because the immunoreactive band is not detectable after reprobing with anti-NEPH2.ic, our results suggest that MT1-MMP (and potentially other proteases) cleave and release a NEPH2 fragment that lacks the intracellular C-terminal domain of NEPH2.

To monitor the release of the intracellular C-terminal domain, we performed a reporter gene assay that requires episomal translocation of the cleaved C-terminal NEPH2 domain. The chimeric transcription factor Gal4-VP16 was inserted into NEPH2 near the C-terminus. As shown in Figure 6D, co-expression with the pG5E1B-Luc plasmid led to a strong increase in luciferase activity as compared with negative control. Another transmembrane Gal4-VP16 construct (LDLR) did not have these effects, suggesting cleavage of NEPH2 but not LDLR.

To evaluate the significance of NEPH2 shedding in vivo, we tested human urine samples for anti-NEPH2.ec immunoreactivity. Urinary protein was precipitated and subject to SDS-PAGE followed by Western blotting and incubation with anti-NEPH2.ec antiserum. A 95-kD band was detectable in the urine of healthy controls, suggesting that NEPH2 is secreted under physiologic conditions (Figure 7A). Additional bands were detected at 65 kD and 55 kD, and more variably at 23 kD, whereas the reprobe of the Western blots for NEPH2.ic was negative. In patients with membranous glomerulonephritis, there seems to be an increase in NEPH2 shedding as compared with a healthy control (Figure 7B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Shedding of NEPH2 in vivo. (A) Urine from healthy individuals was subject to acetone precipitation. 25 μg were separated on SDS-PAGE. Western blot analysis with anti-NEPH2.ec antibody revealed NEPH2-shedding products. (B) Shedding of NEPH2 in vivo under disease conditions. Urine samples of a healthy individual and three patients with membranous glomerulonephritis were subjected to acetone precipitation and 15 μg and 25 μg were separated on SDS-PAGE. Western blot analysis with anti-NEPH2.ec antibody suggested that patients with membranous glomerulonephritis (urinary protein levels 1 to 2 g/d) excreted significantly more NEPH2 shedding products than a healthy control (left two lanes).

NEPH2 Expression and Interactions

NEPH2 belongs to the NEPH family of proteins, which are highly conserved within mammalian and submammalian species. Searching the GenBank database, we found NEPH proteins in mice and rats, but also in Xenopus laevis, Danio rerio, and insects. NEPH2 has been shown to interact with cytoplasmatic proteins of the slit diaphragm. Extracellular ligands, however, have not been identified so far. We now demonstrate homodimerization of extracellular NEPH2 domains as well as extracellular interactions with nephrin. Accordingly, NEPH2 was expressed at the glomerular slit diaphragm.

Similar findings have been reported for NEPH1 (4). It is expressed at the slit diaphragm and knockout mice show a renal phenotype with podocyte foot process effacement and congenital nephrotic syndrome (5,6). NEPH1 forms homodimers and interacts with nephrin, and as in NEPH2, these interactions cannot be mapped to single Ig domains (4). Similar to NEPH1, NEPH2 was expressed in a wide variety of tissues, including fetal and adult kidneys and a podocyte cell line. Both NEPH1 and NEPH2 interact with the slit diaphragm components podocin and ZO-1 and, as shown here, both are expressed at the slit diaphragm and are extracellular ligands of nephrin. These observations suggest a similar role for both NEPH proteins in renal filtration. There are differences, however. NEPH1 expression is much reduced in adult mice as compared with embryos, whereas we found high protein levels of NEPH2 in adult animals. Although both proteins form homodimers and interact with nephrin, they did not interact with one another.

In Drosophila, the nephrin-like proteins hibris (Hbs) and sticks and stones (Sns), as well as the NEPH proteins kirre/dumbfounded (Duf) and irregular chiasm roughest (IrreC-rst), have been implicated in myoblast fusion and myotube guidance (14). S2 cell aggregation assays revealed heterotypic interactions of Duf with Hbs and Sns (15). Muscle precursor cells in Drosophila carry either NEPH proteins (founder cells) or nephrin-like proteins (fusion-competent myoblasts). These cells will only fuse heterogeneously, i.e., NEPH protein expressing cells will never fuse with one another (16). As a first step in muscle fusion, fusion-competent myoblasts extend filopodia directed toward founder cells. These processes will be randomly orientated in Duf mutant embryos and they may be misguided by ectopic expression of either Duf or IrreC-rst (15,17). These observations indicate that NEPH proteins may guide cell processes and control their separation or fusion. Similar findings have recently been reported for Caenorhabditis elegans (18). Synaptic guidepost cells required to drive synapses from the motor neuron onto adjacent target cells transiently express the nephrin-like protein SYG-2 during synapse formation. SYG-2 interacts with the NEPH protein SYG-1 expressed on presynaptic axons and directs SYG-1 accumulation and synapse formation to adjacent regions of the axon. Mutants of either SYG-1 or SYG-2 cause defects in synaptic specificity with ectopic synapse formation on inappropriate targets. Furthermore, SYG-2 misexpression causes aberrant accumulation of SYG-1 and synaptic markers adjacent to the SYG-2-expressing cells. NEPH2 is highly expressed in the mammalian muscle and brain. Further studies will need to examine whether NEPH2 controls synapse and myocyte formation during mammalian development.

NEPH2 Shedding

A number of membrane proteins have soluble forms that are released into the extracellular space. Although these soluble forms may result from alternative splicing, they most often derive from proteolysis of the extracellular domain. The cleavage occurs close to the transmembrane domain, often releasing physiologically active protein. In many cases, shedding can be blocked by metalloproteinase inhibitors, suggesting the involvement of MMP.

We found NEPH2 immunoreactivity in the supernatants of cultured mouse podocytes using Western blot analysis. There was a strong increase in soluble NEPH2 after incubating cells with pervanadate, an inhibitor of tyrosine phosphatases. In contrast, soluble NEPH2 disappeared from the supernatants in the presence of GM6001, an inhibitor of MMP. These findings prompted us to suspect that a cleaving enzyme of the MMP family is responsible for NEPH2 shedding. A prominent candidate was MT1-MMP, also known as MMP14, a membrane-bound calcium-dependent MMP. MT1-MMP cleaves several other membrane proteins, e.g., betaglycan, in a tyrosine-phosphorylation-dependent manner (19). MT1-MMP can be blocked by GM6001 in MCF7 cells and is activated by furin (20,21). Importantly, both MT1-MMP and furin have recently been shown by electron microscopy to co-localize at the glomerular slit diaphragm (22). We found that supernatants of HEK 293T cells, expressing both NEPH2 and MT1-MMP, contained NEPH2. HEK 239T cells express endogenous NEPH2, yet no immunoreactivity was detectable in the absence of MT1-MMP. A faint band appeared in cells transfected with MT1-MMP and a control vector (Bcl2), probably as a result of shed endogenous NEPH2. Using a gene reporter assay, we were able to demonstrate the release of the cytoplasmic tail of NEPH2 from the plasma membrane. Further studies will be required to determine potential signaling capacities of this fragment.

Shedding appears to occur under physiologic conditions. Examining urine of healthy subjects with antiserum directed against the extracellular NEPH2 domain, we detected several NEPH2 fragments, ranging from approximately 95 kD to 23 kD.

COS7 cells transiently overexpressing mouse NEPH2 shed a protein of approximately 70 kD (10), suggesting that the size of fully glycosylated NEPH2 in podocytes might be larger than the reported 97 kD. The calculated mass of the NEPH2 extracellular domain is approximately 60 kD, but it contains five predicted N-glycosylation sites. Transient overexpression of human NEPH2 cDNA in HEK 293T cells results in a band of approximately 100 kD. However, mouse podocytes express a 125-kD protein, suggesting cell-specific differences in posttranslational modification and/or splicing. The presence of multiple tissue-specific forms is supported by the presence of several splice variants of the closely related NEPH3 (9). Soluble NEPH2 could also represent a splice variant lacking the transmembrane domain; this has been reported for nephrin (23). However, antiserum directed against the intracellular NEPH2 domain failed to react with urine proteins, arguing against this possibility.

Doublier et al. found rapid loss of nephrin immunoreactivity on the surface of cultured glomerular epithelial cells in response to aggregated human IgG4 and puromycin, suggesting an involvement of the cytoskeleton in nephrin shedding (24). Interestingly, nephrin immunoreactivity was rapidly lost when incubating the cells with TNF-α, a strong inducer of MT1-MMP expression (25). Shedding of nephrin has also been reported in rats with streptozotocin-induced diabetes mellitus (26). More recently, nephrinuria was found in 19% to 35% of type 1 diabetic patients using Western blotting techniques (27). Although nephrinuria was only detectable in diabetic patients, NEPH2 was clearly detectable in urine from healthy volunteers. In contrast to nephrin, shedding of NEPH2 may therefore represent an integral part of its physiologic function. In addition, we observed increased NEPH2 shedding in patients with membranous glomerulonephritis and proteinuria. Thus, urinary NEPH2 might also serve as a novel biomarker of glomerular disease activity and/or efficacy of anti-proteinuric treatment.

We propose that the network created by podocyte foot processes requires dynamic remodeling in response to changes in intraglomerular pressure and fluid composition. NEPH proteins and nephrin may play an important role in the guidance of podocyte foot processes to establish the complex three-dimensional structure that characterizes these epithelial cells. Shedding of the extracellular domains may regulate turnover and regeneration, and can be controlled by environmental changes that stimulate metalloproteinases. Further work will be required to test this hypothesis.