ARTICLES J Am Soc Nephrol 11: 991–1001, 2000

Primary Structure of Mouse and Rat Nephrin cDNA and Structure and Expression of the Mouse

HELI PUTAALA,* KIRSI SAINIO,† HANNU SARIOLA,† and KARL TRYGGVASON* *Division of Matrix Biology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden; and †Developmental Biology Research Program, Institute of Biotechnology, University of Helsinki, Finland.

Abstract. Nephrin is a central component of the glomerular the interpretation of sequence variants and mutations in the slit diaphragm and is essential for the normal renal nephrin gene in patients with congenital nephrotic syndrome. filtration process. This study describes the complete structure In situ hybridization analyses of whole mouse embryos and of the mouse nephrin gene, which was shown to be homolo- tissues revealed high expression of nephrin in glomeruli gous to the human gene, the major difference being 30 exons and, surprisingly, an intense and highly restricted expression in in the mouse gene as opposed to 29 in human. The complete a set of cells in hindbrain and spinal cord. No expression was primary structure of mouse and rat nephrins was also deter- observed elsewhere. This expression pattern may explain oc- mined. The sequence identity between the mouse and rat casionally occurring neural symptoms caused by inactivating was shown to be 93%, while both rodent proteins have mutations in the nephrin gene in patients with congenital only about 83% sequence identity with human nephrin. The nephrotic syndrome. availability of the three mammalian sequences is significant for

The renal filtration barrier, in which ultrafiltration of plasma meets the lateral cell membrane of the foot processes of podo- occurs, consists of the glomerular basement membrane (GBM) cytes (7,8). located between the fenestrated endothelial cell and epithelial We have recently identified a novel , termed nephrin podocyte layers, as well as the slit diaphragm located between (9), and shown its specific localization to the slit diaphragm the podocyte foot processes (1). During embryonic develop- (10). The nephrin gene is mutated in congenital nephrotic ment, epithelial cells becoming glomerular differ- syndrome (NPHS1) (9,11), a disease characterized by massive entiate from the metanephric mesenchyme at the S-shaped proteinuria and edema (12). Based on Northern and in situ body stage, when the podocyte layer is first seen at day 13 hybridization, nephrin has been shown to be expressed by the (E13) of mouse embryonic development (2). Although a glomerular podocytes (9). Using immunoelectron microscopy, wealth of knowledge has accumulated on the structure and we recently localized nephrin to the slit diaphragm where properties of GBM components (3,4), the filtration role of the nephrin is surmised to form a zipper-like isoporous filter struc- ture (10) similar to that proposed by Rodewald and Karnovsky podocyte cell layer, particularly that of the slit diaphragm, has based on transmission electron microscopic observations (13). long been unclear. Only a couple of proteins have been local- Because nephrin resides in the slit diaphragm, and mutations in ized to the slit diaphragm region, one of which was a 51-kD the gene cause severe nephrotic syndrome, this protein must antigen recognized by monoclonal antibody 5-1-6 (5,6). The have a crucial role in the filtration mechanism of the glomeruli antigen epitopes recognized by the antibody are unknown, and may participate more generally in the pathogenesis of however. Another protein that has been localized to the region proteinuria and renal failure. ␣Ϫ is the -isoform of the intracellular tight junction protein To further characterize nephrin and facilitate studies on its zonula occludens-1, located at sites where the slit diaphragm involvement in renal physiology and disease, we have now determined the primary structure of both mouse and rat nephrin based on cDNA sequences, determined the structure of the mouse gene, and studied its expression pattern both by North- Received September 10, 1999. Accepted November 22, 1999. The sequences reported here have been submitted to the GenBank database ern and in situ hybridization analyses during mouse embryo- with accession nos. AF172247 to AF172256 genesis. Correspondence to Dr. Karl Tryggvason, Division of Matrix Biology, Depart- ment of Medical Biochemistry and Biophysics, Karolinska Institute, S-17177 Stockholm, Sweden. Phone: ϩ46 8 728 7720; Fax: ϩ46 8 316 165; E-mail: Materials and Methods [email protected] Isolation and Characterization of Genomic Mouse 1046-6673/1106-0991 Bacterial Artificial Clones Journal of the American Society of Nephrology Genomic mouse bacterial artificial chromosome (BAC) clones Copyright © 2000 by the American Society of Nephrology containing the nephrin gene were isolated using 5Ј-ACG TGA GGG 992 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 991–1001, 2000

CCG AGC GGA CAT G-3Ј and 5Ј-CCA CAT AGT CCA GCC ACT instructions. Three positive primary plaques were obtained and, after TGA-3Ј primers (GenomeSystems) that are specific for amyloid ␤ secondary screening, three positive-phage DNA were isolated using a precursor-like protein 1 (APLP1), whose gene is located immediately lambda mini kit (Qiagen), digested with EcoRI and subcloned into adjacent to the nephrin gene (9,14). Approximately 50 ␮gofBAC pBluescript SK(ϩ). Three clones, 3.32, 5.92, and 6.21, were se- clone DNA was digested with EcoRI and BamHI restriction enzymes, quenced from both strands and the sequences were analyzed with the and the resulting fragments were run in a 1% agarose gel and South- FASTA program. A missing segment between clones 3.32 and 5.92 ern-blotted to a Biodyne B nylon filter (Pall Corp., Lund, Sweden) was amplified using Vent DNA polymerase from the rat kidney ␭gt10 (15). The filter was hybridized at 42°C with human nephrin cDNA, cDNA library using two primer pairs: r(forw)1 5Ј-CAT CCT GGC first with a fragment comprising exons 2 to 3, and then with a CAA CTC GTC CG-3Ј together with r(rev)1 5Ј-GGA GTA GGC fragment containing exons 26 to 29. These probes were radioactively TGA TCC ACC TG-3Ј; and r(forw)2 5Ј-CAG GCG ACG CCT TGA labeled with [␣-32P]dCTP (3000 Ci/mmol) (Amersham Pharmacia ACT TG-3Ј together with r(rev)2 5Ј-GCA AAT CGG ACG ACA Biotech), using random priming (Life Technologies). Positive frag- AGA CG-3Ј. The fragments were agarose-gel purified and subcloned ments, 2.4-kb EcoRI (positive for exons 2 to 3) and 7.2-kb BamHI into the pCR-Script SK(ϩ) vector. Two clones from each reaction (positive for exons 26 to 29), were purified with the Gel Extraction kit were sequenced from both strands. To obtain the missing 5Ј end, two (Qiagen, Chatsworth, CA) and subcloned into the pBluescript SK(ϩ) sets of amplification reactions were performed from the rat kidney vector. The plasmids were isolated using the QIAprep spin miniprep- Marathon-Ready cDNA (Clontech) using Vent DNA polymerase (5Ј kit (Qiagen) and sequenced on an ABI 310 Genetic Analyzer (Perkin RACE). First, a round using primer pair AP1 together with r(rev)3 Elmer). For identification of exons, the sequences were compared 5Ј-AGC ACG ATC TCC TTC CAG GC-3Ј was carried out, followed with those of the human nephrin cDNA sequence using the FASTA by a second nested round using primer pair AP2, together with r(rev)4 program. This comparison yielded exons 1 to 4 from an EcoRI 5Ј-GAA GCC TGG CAT CTT CGG G-3Ј. The resulting PCR frag- fragment and exons 24 to 30 from a BamHI fragment of the genomic ments were agarose gel-purified and sequenced directly from both BAC clone. strands.

Cloning of Mouse Nephrin cDNA Mouse nephrin exon-specific primers were designed from genomic Analysis of cDNA and Amino Acid Sequences sequence data obtained from the subcloned BAC clone fragments. The final cDNA and amino acid sequences were compared with the Two primer pairs were used for PCR to amplify the cDNA from human ones using the FASTA and BLAST programs. Alignment of mouse kidney QUICK-Clone cDNA (Clontech, Palo Alto, CA) with amino acid sequences was performed with the PileUp program (Ge- Advantage cDNA polymerase mix (Clontech): m(forw)1 from the 5Ј netics Computer Group, Madison, WI), and signal peptide prediction untranslated region (UTR) 5Ј-GAC AGC AAC AAA CAA GCT GCT with the SPScan program (Wisconsin Package version 10.0, Genetics GG-3Ј together with m(rev)1 primer from exon 29 5Ј-TCA CAC CAG Computer Group). Protein patterns were searched using the PROSITE ATG TCC CCT CAG C-3Ј; and the same m(forw)1 primer together protein pattern search tool (16). with m(rev)2 primer from the 3ЈUTR region 5Ј-GGA AAC AGG TGT CGT GAA GAG TC-3Ј. The PCR products were agarose-gel purified and subcloned into the pCR-Script SK(ϩ) vector (Stratagene, La Determination of Exon-Intron Boundaries of the Mouse Jolla, CA). Two clones from each amplification were sequenced from Nephrin Gene both strands using either ABI 310 or 377 Genetic Analyzers. The To determine the exon-intron boundaries, exon-specific primers sequences were compared to each other using the FASTA program, were designed and intervening intron sequences were PCR amplified and nonmatching sequence segments were reamplified using Vent from BAC clones using HotStarTaq DNA polymerase (Qiagen). The DNA polymerase (New England Biolabs, Beverly, MA) from mouse fragments were size-determined and agarose-gel purified. The exon- kidney QUICK-Clone cDNA in two rounds. The PCR fragments were intron boundaries were sequenced from both strands either from the agarose-gel purified and sequenced directly from both strands. The amplified fragment or from the subcloned BAC-fragments described final mouse cDNA sequence was determined by comparing the se- above. The smallest introns were sequenced entirely. For intron 23, quences to that of the gene. Furthermore, the 5Ј region of the cDNA the BAC Southern nylon filter was hybridized at 65°C with a mouse was amplified (rapid amplification of cDNA ends [5Ј RACE]) in two nephrin cDNA fragment comprising exons 23 to 24. A positive 8.7-kb separate reaction sets in two rounds from a mouse kidney Marathon- EcoRI fragment was agarose gel-purified, subcloned into pBluescript Ready cDNA (Clontech) with the Advantage cDNA polymerase mix. SK(ϩ), and sequenced with exons 23 and 24-specific primers to find In the first round, AP1 primer was used together with exon 2-specific the exon-intron junction. A region with high GC content in the m(rev)3 primer 5Ј-CAG AAG CAG CCC ATC CTT AGC-3Ј and boundary of exon 23 and intron 23 was sequenced until a suitable site m(rev)4 primer 5Ј-CAG GTA ACT GTG CTT CCT GCC TC-3Ј. The for PCR primer annealing was found. The rest of intron 23 was second, nested amplification round was performed with AP2 primer PCR-amplified with intron 23-specific primers. together with exon 2-specific m(rev)3b primer 5Ј-CAG ATA GAG CCC AGA AGC CTC G-3Ј and m(rev)3 primer that is upstream to the m(rev)4 primer, respectively. The nested PCR round products were ϩ Northern Analysis of Mouse Nephrin Expression agarose gel-purified and subcloned into pCR-Script SK( ) vector. ϩ Three subclones were sequenced from both strands. By using multiple-tissue Northern analysis, poly(A) RNA from eight adult mouse tissues and 7-, 11-, 15-, and 17-d-old mouse embryos were studied (Clontech). Hybridizations were carried out at Cloning of Rat Nephrin cDNA 65°C using a PCR-amplified fragment comprising exons 23 to 29 as To obtain the rat nephrin cDNA, a rat kidney ␭gt10 cDNA library probe, as described above. The Northern blots were also probed with (Clontech) was screened using a human nephrin cDNA fragment similarly labeled human ␤ actin cDNA to compare the loading in each comprising exons 3 to 29 as a probe, according to the manufacturer’s lane. J Am Soc Nephrol 11: 991–1001, 2000 Mouse and Rat Nephrin 993

In Situ Hybridization of the three clones obtained. Two restriction fragments of 2.4 For in situ hybridization, a probe from the 3Ј end of the gene was and 7.2 kb hybridizing with the human cDNA were sequenced amplified from a mouse kidney QUICK-Clone cDNA using exon 29 and shown to contain sequences corresponding to exons 1 to 4 specific m(forw)2 primer 5Ј-GAC CCC TAT GAC CTT CGC TG-3Ј, and 24 to 29 in the human gene, respectively. Primers based on together with a 3ЈUTR-specific m(rev)5 primer 5Ј-CAG ATG TCA these exon sequences were then used for amplification of the Ј GCT GGA GTC TTC-3 . The amplification was performed using mouse nephrin cDNA from pooled kidney mRNA, and the Advantage cDNA polymerase mix. The resulting fragment was sub- full-length cDNA could be compiled and used to determine the cloned into the pCR-Script SK(ϩ) vector and sequenced from both exon structure. strands. The 219-bp sense BamHI and antisense NotI 35S-labeled cRNA probes were transcribed, and in situ hybridization was per- The mouse nephrin gene spans about 27 kb (Figure 1). To formed essentially as described by Wilkinson and Green (17) with determine the exon-intron boundaries of the gene, exon-spe- some modifications. Briefly, E11 to newborn whole mouse embryos cific primers were used to amplify the intervening intron se- and isolated kidneys were fixed in 4% paraformaldehyde and pro- quences from BAC clones except for intron 23, which was cessed for paraffin histology. Sections at 7 ␮m were cut, treated in cloned separately (see Materials and Methods). Unlike human prehybridization solutions as described (17), and finally hybridized NPHS1, which has 29 exons, the mouse gene has 30 exons, the overnight at ϩ65°C. After high stringency washes at ϩ65°C, the last exon encoding only for the major part of the 3Ј UTR. The sections were dehydrated, air-dried, dipped into NTB-2 emulsion exon sizes range from 25 to 216 bp (Figure 2) and are similar ϩ (Eastman Kodak), and exposed at 4°C in the dark for 10 to 14 d. The between human and mouse, except for exons 3, 8, 13, and 28, sections were developed in D19 (Eastman Kodak), fixed in sodium which differ by 1 to 6 bp (11). The codon for translation fixative (Eastman Kodak), counterstained in hematoxylin (Shandon, initiator methionine is in the first exon, and the translational Pittsburgh, PA), and mounted. Sections were analyzed and photo- graphed with Olympus Provis AX microscope equipped with a stop codon TGA is in exon 29. charge-coupled device camera (Photometrics Ltd.). The dark-field In human, an unusual donor site starting with GC instead of images were inverted, stained red, and combined with the bright-field the GT exists at exon 23, which encodes part of the fibronectin images in the Adobe PhotoShop 4.0 program. type III module (11). This turned out to be the case at mouse exon 23 as well. All other acceptor (AG) and donor (GT) sites Results are conventional. Isolation and Characterization of the Mouse Nephrin Gene Determination of Mouse and Rat Nephrin cDNA Hybridization of several mouse cDNA and genomic libraries Sequences with human cDNA or genomic clones did not yield any mouse The coding sequence of the full-length mouse cDNA was nephrin clones. Therefore, isolation of the nephrin gene locus shown to be 3768 nucleotides. The codon for the first possible was attempted by isolating BAC clones containing the putative initiator methionine was more 5Ј than that of the human se- neighboring APLP1 gene. Human chromosome 19q and the quence (Figure 3). This was verified using a 5Ј RACE reaction, proximal portion of mouse chromosome 7 have been shown to as well as by analyzing the sequence of the 5Ј end of the gene be highly syntenic (18,19) (Mouse Genome Database, Mouse (Figure 2). The analyzed genomic exon sequences of the Genome Informatics, The Jackson Laboratory, Bar Harbor, mouse gene were all found in the sequenced mouse cDNA ME; Internet: http://www.informatics.jax.org). The number clones. and order of as well as their spacing are fairly similar The sequence of the rat nephrin cDNA was determined from between the two , the main difference being three cDNA clones, 3.32 containing bases 270-1819, 5.92 reversed centromeric-telomeric orientation on the mouse and containing bases 2526–3840, and 6.21 containing bases 2730– human chromosomes (18). Three APLP1-positive BAC clones 3971 (not shown). The missing parts were amplified either were found to hybridize with human nephrin cDNA probes. from the same cDNA library or with 5Ј RACE. The full-length Southern hybridization of the APLP1-positive BAC clones rat cDNA contained 3756 bp. Two codons for methionine are with different fragments of the human nephrin cDNA demon- present in the 5Ј end, the first located at the same site as that in strated that the entire mouse nephrin gene was present in two the mouse cDNA and the other 12 bp downstream of the codon

Figure 1. Schematic structure of the mouse nephrin gene. Exons are indicated by black rectangles; introns and flanking are indicated by rectangles with diagonal lines. Exons are numbered. 5Ј and 3Ј untranslated regions are marked with hatched boxes, the sizes of which are unknown. The scale at the bottom is in kilobases (kb). 994 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 991–1001, 2000

Figure 2. Exon-intron boundaries and sizes of exons and introns in the mouse nephrin gene. Intron sequences are shown with lowercase letters. Exon sequences are depicted by uppercase letters, with the one-letter amino acid residues shown below. Exon and intron sizes are given in base pairs. Introns marked with an asterisk were size-determined by agarose gel electrophoresis of PCR products, except for intron 23, which was partially sequenced. J Am Soc Nephrol 11: 991–1001, 2000 Mouse and Rat Nephrin 995 for the initiator methionine in the human cDNA. Because the family (22,23). Furthermore, two potential conserved protein nucleotides around the upstream methionine codon better fit kinase C phosphorylation sites with the consensus sequence the Kozak’s consensus sequence (20) for translation initiation Ser/Tyr-X-Arg/Lys (Figure 3) were detected by PROSITE than the downstream one, the upstream codon is probably used protein pattern search tool (16). for the initiator methionine. This is also supported by the high identity (82%) in nucleotide sequence between the mouse and Developmental Expression of Mouse Nephrin rat from the first rat methionine codon to the second. The overall expression pattern of mouse nephrin was studied by Northern analysis, using RNA from whole mouse embryos Comparison of cDNA-Deduced Mouse, Rat, and of different ages and adult mouse tissues. No signals above Human Nephrins background could be observed with poly(A)ϩ RNA from four While the mouse and rat amino acid sequences show 93% different aged mouse embryos (not shown). With poly(A)ϩ identity, both the mouse and rat amino acid sequences show RNA from eight adult mouse tissues, the analysis revealed a 83% identity with the overlapping areas of the human nephrin signal only with RNA from kidney (Figure 4). RNA from sequence. The most notable difference is at the amino termini, heart, brain, spleen, lung, liver, skeletal muscle, and testis were where the mouse and rat proteins are 14 residues longer than in all negative. human (Figure 3). Mouse nephrin has 1256 amino acid resi- Because Northern analyses may not reveal the presence of dues with a predicted molecular weight of 136 kD without transcripts with highly restricted expression, we carried out in posttranslational modifications, whereas rat nephrin has 1252 situ hybridization on whole mouse embryos and isolated met- residues with a predicted molecular weight of 136 kD. Human anephric and mesonephric kidneys. At E11, when the meta- nephrin has previously been shown to contain 1241 residues, nephric differentiation is initiated, no signal could be detected with a molecular weight of 135 kD (9). from the mesenchymal cells of the presumptive kidney (Figure The signal peptide cleavage site in human nephrin is located 5C). However, nephrin expression was seen in the E11 meso- between residues 22 and 23, whereas it is predicted to be nephric kidney, where cranial tubules with podocyte-like struc- between residues 36 and 37 in mouse and rat. All 10 potential tures revealed strong expression, while the mesonephric mes- N-glycosylation sites are conserved between the three species enchyme and less differentiated caudal tubules remained (Figure 3). Of the seven serine-glycine (SG) doublets that are negative (Figure 5, A and B). A highly specific expression was potential heparan sulfate attachment sites, only two are con- observed in the podocytes of the developing kidney beginning served between the three species. from the E13 mouse embryos (Figure 5, E and F). Early Similar to Ig motifs in other proteins (21), all eight Ig motifs S-shaped bodies, as well as the metanephric mesenchyme, in nephrin of the three species contain two conserved cysteines remained negative at this stage. Surprisingly, high expression that form disulfide bonds within the motif. In addition, the of nephrin mRNA was observed in the hindbrain and spinal mouse and rat nephrin molecules contain three cysteine resi- cord (Figure 6A). At E11 in the brain, the expression was dues that are conserved between the species and, depending on limited to a subset of neurons in the hindbrain area and con- the species, one or two unconserved residues. The conserved tinued from there dorsally, confined to the mantle layer neu- residues are one in the first Ig motif, one in the spacer domain, rons of the spinal cord (Figure 6B). A similar expression and one in the transmembrane domain. The unconserved res- pattern was observed at E13 (Figure 6C). From E13 to E17, idues include one cysteine in Ig motif 4 in the mouse, one each nephrin mRNA was detected in neuroepithelium of cerebellum in Ig motifs 2 and 4 in the rat, and one residue in the fibronectin primordium at the roof of the fourth ventricle, but the epithelial III domain in human nephrin. Furthermore, both the mouse cells of developing choroid plexus, which are involved in the signal peptide and the rat intracellular part have one cysteine formation of cerebrospinal fluid, remained negative. Expres- residue each (Figure 3). sion was not observed in any other tissues of the embryo. The intracellular domain in human has nine tyrosine residues that are tentative phosphorylation sites for intracellular signal- Discussion ing, while the corresponding domains of mouse and rat contain The recent identification of nephrin as the first known mo- 10 and eight residues, respectively. Six of those tyrosine resi- lecular component of the slit diaphragm and as a protein whose dues are conserved between the three species. However, the mutations cause proteinuria has opened new possibilities for sequences around the tyrosines are completely conserved be- exploring the mechanisms of glomerular ultrafiltration and the tween the three species only in the case of three of the residues. pathogenetic mechanisms of proteinuria (9–11). In this regard, According to the numbering of mouse nephrin, these are ty- it is important to be able to carry out studies on nephrin and its rosines 1128, 1208, and 1232. Tyrosine phosphorylation of involvement in protein filtration in experimental systems, in- nephrin is still an open question, so the naming of any specific cluding well established proteinuria models in mice and rats. interactors is purely speculative. Yet, comparison to the opti- The present work facilitates such studies by providing both mal recognition motifs for SH2 domains from different pro- nucleotide and amino acid information on mouse and rat teins reveals that if tyrosine 1208 is phosphorylated, a binding nephrins, as well as by determining the complete structure of site for an adapter protein Nck might be generated. This motif the mouse gene and the location of its activity during the pYDEV, and the one containing tyrosine 1232, pYDQV, may embryonic development. also provide a binding site for tyrosine kinases of the Src The present analyses of the mouse gene demonstrated its 996 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 991–1001, 2000 J Am Soc Nephrol 11: 991–1001, 2000 Mouse and Rat Nephrin 997

for the zipper-like nephrin structure of the slit membrane (10), where nephrin molecules from two opposite foot processes interact in an interdigitating manner through homophilic inter- actions reinforced by intermolecular disulfide bonds between the two free extracellular cysteines in Ig motif 1 and the spacer domain. Furthermore, the glycoprotein nephrin has 10 con- served N-glycosylation sites, suggesting that controlled glyco- sylation is important for the protein. There are six conserved tyrosine residues in the intracellular domain of nephrin. It is not yet known whether any of those serves as a phosphoacceptor site. If this turns out to be the case, one could envision several roles for the tyrosine phosphoryla- tion. First, the phosphorylation could have a role in the acti- Figure 4. Northern analysis of mouse nephrin expression with mRNA vation of signal transduction pathways upon binding of the from eight adult mouse tissues and whole mouse embryos. Mouse extracellular ligand(s). Interestingly, residues immediately car- nephrin cDNA fragment comprising exons 23 to 29 was used as a boxy-terminal to three of the tyrosines are completely con- probe. The tissues studied are marked above the filter, and molecular served between the three species. Thus, comparison to optimal size markers (kb) are shown to the sides of the filters. The arrow binding sequences for SH2 domains suggests that signaling indicates the observed band. molecules, such as an adaptor protein Nck and Src-family protein kinases, may be recruited to nephrin if tyrosines 1208 and 1232 are phosphorylated. Second, it is also possible that similarity to the human one, except for the 3Ј UTR, which is tyrosine phosphorylation of nephrin plays either a positive or encoded by two exons (exons 29 and 30) in the mouse gene as opposed to one (exon 29) in the human gene (11). The gene negative role in the assembly of the glomerular slit diaphragm. sequence was also important for verification of the mouse In this regard, it is noteworthy that tyrosine phosphorylation of cDNA sequence, which differs particularly in the 5Ј end, that the intracellular domain of neurofascin, which also is a trans- is, in the signal peptide from the human gene. The presently membrane protein of the Ig superfamily, abolishes its binding elucidated mouse gene structure will be valuable for enabling to the membrane skeletal protein ankyrin and reduces coupling the generation of a mouse model for congenital nephrotic of neurofascin to the cytoskeleton (24). Yet another possibility syndrome, using gene targeting in embryonic stem cells. is that tyrosine phosphorylation regulates extracellular ligand The mouse and rat cDNA sequences determined here specificity of nephrin, as might be the case with platelet endo- showed that mouse and rat nephrins have 93% sequence iden- thelial cell adhesion molecule-1, another member of the Ig tity, while human nephrin gene shows a considerably lower superfamily (25). sequence identity of 83% with the two rodent species. These Thus far, 21 missense mutations resulting in amino acid sequences of nephrin from three mammalian species are of changes, mainly in the Ig modules, have been reported in considerable value in several respects. First, the conserved patients with congenital nephrotic syndrome (9,11). All of the residues suggest significant structural motifs and functional mutated amino acid residues, except for two, are conserved importance for the molecule. Second, the availability of the between the three mammalian species analyzed in the present sequences is important not only for the interpretation of the study. The two missense mutations affecting unconserved impact of the numerous sequence variants found thus far both amino acid residues are Ile173 3 Asn in Ig 2 and Arg1140 3 in patients with nephrotic syndrome and in the disease carriers, Cys in the intracellular domain. Although not all missense but also for the numerous homozygous and heterozygous allele mutations involve conserved amino acids, it is apparent that variants found in the general population with no family history one can suspect individuals with allelic changes in codons for of nephrotic syndrome (9,11). conserved amino acids to be carriers for congenital nephrotic Several conclusions can be made from the comparison of the syndrome. Thus, the three sequences can indeed have direct three primary structures of nephrin. Thus, the three conserved, practical value. apparently free cysteines in the Ig motif 1, the spacer domain, Recently, cDNA and amino acid sequences were reported and the transmembrane domain are probably crucial for the for the mouse (26) and rat (27) nephrin homologues. Compar- function of the protein. This supports our hypothetical model ison of the amino acid sequences between the mouse sequence

Figure 3. Comparison of mouse, rat, and human nephrin at amino acid level sequences. The sequence alignment was performed with the PileUp program. The signal peptide cleavage site is marked by an arrow. Ig motifs are boxed and the fibronectin type III-like domain is boxed with a hatched line. The transmembrane domain is underlined with a solid line. Conserved cysteine residues are shaded, unconserved residues are shown in bold, and conserved N-glycosylation sites are indicated with a triangle (). Conserved tyrosine residues in the intracellular domain are indicated with a dot (F), and two possible protein kinase C phosphorylation sites with the consensus sequence Ser/Tyr-X-Arg/Lys are marked with a star (ૺ). 998 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 991–1001, 2000

Figure 5. In situ hybridization analysis of nephrin expression during mouse kidney development. (A and B) Embryonic day 11 (E11) mouse mesonephric (Me) kidney. Nephrin mRNA is seen in the podocyte layer of the cranial (Cr) mesonephric nephron. The caudal (Ca) tubules remain negative. (C) Mouse metanephric kidney at E11. Metanephric mesenchyme surrounding the ureteric bud (ub) does not express nephrin mRNA at this stage. Caudal mesonephric tubule (arrow), genital ridge (gr), and wolffian duct (wd). (D) Early nephrones of E13 kidney cortex. Presumptive podocytes (*) at the S-shaped bodies are still negative. Ureteric buds (ub) are marked with arrows. (E) Nephrin mRNA is first detected at E13 glomerular structures. Early podocytes from the S-shaped body (*) are negative. (F) Glomerular podocytes from newborn kidney show high expression of nephrin mRNA. The cortical undifferentiated structures are negative. Bars: 150 ␮m in A and B; 100 ␮min C; 60 ␮minD;80␮m in E and F. J Am Soc Nephrol 11: 991–1001, 2000 Mouse and Rat Nephrin 999

reported here and in that by Holzman et al. (26) reveals differences in the 5Ј end or in the signal peptide region. In the sequence reported by Holzman et al. (26), the first 10 amino acids are 100% identical to the human sequence. The 5Ј end mouse sequence reported here, however, is longer, and it differs significantly from the human one. Comparison of the 5Ј end of the mouse sequence with that of the mouse genomic sequence showed that the sequence reported in this study is correct. Thus, the reason for the difference between the two mouse sequences is unknown. Otherwise, only two amino acids were different in the two mouse amino acid sequences: Thr763 reported here was Ala in the Holzman study, and Leu1076 was reported to be Gln in their study (26). Both Thr763 and Leu1076 are conserved in human and rat, indicat- ing that they are also conserved in mouse. With regard to the rat sequences, we found a longer signal peptide than that reported by Ahola et al. (27). Furthermore, Ahola et al. re- ported that there is Asn in position 1229, whereas we found it to be Asp between all three species studied here. Nephrin mRNA was located by in situ hybridization both to the embryonic mesonephric and metanephric kidneys. The undifferentiated metanephric mesenchyme and first epithelial structures, the comma- and early S-shaped bodies, remained negative. The first mRNA were detected in late S-shaped bodies in the definitive podocytes. A similar expression pattern was noticed in the glomerular structures of the mesonephric kidney, which reveals that the mouse cranial mesonephric nephrones (28) could be secretory. Instead, the caudal meso- nephric nephrones did not express nephrin mRNA. The presence of nephrin mRNA in a specific region of the brain and spinal cord raises the question about a role of this protein in neuronal development and function. Thus far, severe extrarenal symptoms have not been reported in NPHS1 patients who have received kidney transplants (29). However, of ap- proximately 50 transplanted Finnish patients, about 10% have congenital neurologic symptoms, such as ataxia, not normal- izing after transplantation. These symptoms cannot be ex- plained by thrombosis or other secondary insults due to NPHS1 (Christer Holmberg, personal communication). It is quite plau- sible that these neurologic symptoms are related to absence of normal nephrin in brain tissue, and that this function cannot be compensated for in all individuals. Participating in guidance of neurons (30,31), axonal fasciculation (32), and growth and stabilization of synapses (33), many of the members of the Ig Figure 6. Nephrin mRNA expression in the developing mouse ner- vous system. (A) At E11 hindbrain, nephrin mRNA is highly ex- superfamily have been shown to possess multiple roles in pressed in a subset of neurons both in the upper cerebellar plate and neuronal differentiation. Because congenital ataxias are often in the opposite side of the fourth ventricle, in the myelencephalon. (B) associated with impaired development of the cerebellum and In the neural tube at the lumbar area of the same embryo, the its connections (34), the expression of nephrin in the nervous expression is in the dorsal mantle layer. (C) At E13, the expression in system could indeed indicate a role in neurogenesis. brain is concentrated to the neuroepithelium of the cerebellar primor- The recent evidence on interaction of the cytosolic CD2- dium on the roof of the fourth ventricle. Some expression is also seen associated protein (CD2AP) with nephrin (35) is of great at the floor of the ventricle, but the choroid plexus (arrow) is negative. interest. First and foremost, the interaction provides informa- ␮ ␮ Bars: 250 m in A and B; 150 minC. tion on potential linkage mechanisms between nephrin and the actin cytoskeleton of podocyte foot processes (36). Further- more, removal of this gene was shown to cause a nephrotic syndrome similar to NPHS1. However, the CD2AP-deficient mice die at 6 to 7 wk of age (35), while nephrin-deficient mice 1000 Journal of the American Society of Nephrology J Am Soc Nephrol 11: 991–1001, 2000 die on their first day of life (H. Putaala, unpublished informa- newborn and young infants. Ann Paediatr Fenn 2: 227–241, tion). Therefore, it is possible that nephrin has multiple asso- 1956 ciating ligands that could have a role in the filtration mecha- 13. Rodewald R, Karnovsky MJ: Porous substructure of the glomer- nism. ular slit diaphragm in the rat and mouse. J Cell Biol 40: 423–433, 1974 Note Added in Proof: After submitting this work, mAb 5-1-6 14. Wasco W, Bupp K, Magendantz M, Gusella JF, Tanzi RE, Solomon F: Identification of a mouse brain cDNA that encodes was shown to be directed against the extracellular domain of a protein related to the Alzheimer disease-associated amyloid ␤ rat nephrin (Topham PS, Kawachi H, Haydar SA, Chugh S, protein precursor. Proc Natl Acad Sci USA 89: 10758–10762, Addona TA, Charron KB, Holzman LB, Shimizu F, Salant DJ. 1992 J Clin Invest 104: 1559–1566, 1999). 15. Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl K: Current Protocols in Molecular Biology, Acknowledgments New York, Greene Publishing and Wiley-Interscience, 1992, pp This work was supported in part by grants from the National 2.9.1–2.9.6 Institutes of Health (DK 54724), the Sigrid Juselius Foundation, the 16. Hofmann K, Bucher P, Falquet L, Bairoch A: The PROSITE Novo Nordisk Foundation, and Hedlund’s Foundation. We thank database, its status in 1999. Nucleic Acids Res 27: 215–219, 1999 Timo Pikkarainen for valuable comments on the manuscript and Vesa 17. Wilkinson D, Green P: Postimplantation Mammalian Embryos: Ruotsalainen for calculating the molecular weights. A Practical Approach, edited by Cropp A, Cockroft D, London, Oxford University Press, 1990, pp 155–171 18. Stubbs L, Carver EA, Shannon ME, Kim J, Geisler J, Generoso References EE, Stanford BG, Dunn WC, Mohrenweiser H, Zimmermann W, 1. Tisher CC, Madsen KM: Anatomy of the kidney. In: The Kidney, Watt SM, Ashworth LK: Detailed comparative map of human 5th Ed., edited by Brenner BM, Rector FC, Philadelphia, Saun- chromosome 19q and related regions of the mouse genome. ders, 1996, pp 3–71 Genomics 35: 499–508, 1996 2. Saxen L: Organogenesis of the Kidney, Cambridge, Cambridge 19. Blake JA, Richardson JE, Davisson MT, Eppig JT, for the Mouse University Press, 1987 Genome Database Group: The mouse genome database (MGD): 3. Hudson BG, Reeders S, Tryggvason K: Type IV collagen: Struc- Genetic and genomic information about the . ture, gene organization, and role in human diseases: Molecular Nucleic Acids Res 27: 95–98, 1999 basis of Goodpasture and Alport syndromes and diffuse leiomyo- 20. Kozak M: An analysis of 5Ј-noncoding sequences from 699 matosis. J Biol Chem 268: 26033–26065, 1993 vertebrate messenger RNAs. Nucleic Acids Res 15: 8125–8148, 4. Yurchenco PD, O’Rear JJ: Basal lamina assembly. Curr Opin 1987 Cell Biol 6: 674–681, 1994 21. Chothia C, Jones EY: The molecular structure of cell adhesion 5. Orikasa M, Matsui K, Oite T, Shimizu F: Massive proteinuria molecules. Annu Rev Biochem 66: 823–862, 1997 induced in rats by a single intravenous injection of a monoclonal 22. Songyang Z, Shoelson SE, McGlade J, Olivier P, Pawson T, antibody. J Immunol 141: 807–814, 1988 Bustelo XR, Barbacid M, Sabe H, Hanafusa H, Yi T, Ren R, 6. Kawachi H, Abrahamson DR, St. John PL, Goldstein DJ, Shia Baltimore D, Ratnofsky A, Feldman RA, Cantley LC: Specific MA, Matsui K, Shimizu F, Salant D: Developmental expression motifs recognized by the SH2 domains of Csk, 3BP2, fps/fes, of the nephritogenic antigen of monoclonal antibody 5-1-6. Am J GRB-2, HCP, SHC, Syk, and Vav. Mol Cell Biol 14: 2777–2785, Pathol 147: 823–833, 1995 1994 7. Schnabel E, Anderson JM, Farquhar MG: The tight junction 23. Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, protein ZO-1 is concentrated along slit diaphragms of the glo- merular epithelium. J Cell Biol 111: 1255–1263, 1990 Haser WG, King F, Roberts T, Ratnofsky S, Lechleider RJ, Neel 8. Kurihara H, Anderson JM, Farquhar MG: Diversity among tight BG, Birge RB, Fajardo JE, Chou MM, Hanafusa H, Schaff- junctions in rat kidney: Glomerular slit diaphragms and endothe- hausen B, Cantely LC: SH2 domains recognize specific phos- lial junctions express only one isoform of the tight junction phopeptide sequences. Cell 72: 767–778, 1993 protein ZO-1. Proc Natl Acad Sci USA 89: 7075–7079, 1992 24. Garver TD, Ren Q, Tuvia S, Bennett V: Tyrosine phosphoryla- 9. Kestila¨M, Lenkkeri U, Ma¨nnikko¨M, Lamerdin J, McCready P, tion at a site highly conserved in the L1 family of cell adhesion Putaala H, Ruotsalainen V, Morita T, Nissinen M, Herva R, molecules abolishes ankyrin binding and increases lateral mobil- Kashtan CE, Peltonen L, Holmberg C, Olsen A, Tryggvason K: ity of neurofascin. J Cell Biol 137: 703–714, 1997 Positionally cloned gene for a novel glomerular protein—neph- 25. Famiglietti J, Sun J, DeLisser HM, Albelda SM: Tyrosine residue rin—is mutated in congenital nephrotic syndrome. Mol Cell 1: in exon 14 of the cytoplasmic domain of platelet endothelial cell 575–582, 1998 adhesion molecule-1 (PECAM-1/CD31) regulates ligand binding 10. Ruotsalainen V, Ljungberg P, Wartiovaara J, Lenkkeri U, Kestila¨ specificity. J Cell Biol 138: 1425–1435, 1997 M, Jalanko H, Holmberg C, Tryggvason K: Nephrin is specifi- 26. Holzman LB, St. John PL, Kovari IA, Verma R, Holtho¨fer H, cally located at the slit diaphragm of glomerular podocytes. Proc Abrahamson DR: Nephrin localizes to the slit pore of the glo- Natl Acad Sci USA 96: 7962–7967, 1999 merular epithelial cell. Kidney Int 56: 1481–1491, 1999 11. Lenkkeri U, Ma¨nnikko¨M, McCready P, Lamerdin J, Gribouval 27. Ahola H, Wang S-X, Luimula P, Solin M-L, Holzman LB, O, Niaudet P, Antignac C, Kashtan CE, Holmberg C, Olsen A, Holtho¨fer H: Cloning and expression of the rat nephrin homolog. Kestila¨M, Tryggvason K: Structure of the gene for congenital Am J Pathol 155: 907–913, 1999 nephrotic syndrome of the Finnish type (NPHS1) and character- 28. Sainio K, Hellstedt P, Kreidberg JA, Saxe´n L, Sariola H: Dif- ization of mutations. Am J Hum Genet 64: 51–61, 1999 ferential regulation of two sets of mesonephric tubules by WT-1. 12. Hallman N, Hjelt L, Ahvenainen EK: Nephrotic syndrome in Development 124: 1293–1299, 1997 J Am Soc Nephrol 11: 991–1001, 2000 Mouse and Rat Nephrin 1001

29. Qvist E, Laine J, Ro¨nnholm K, Jalanko H, Leijala M, Holmberg plasticity. I. Fasciclin II controls synaptic stabilization and C: Graft function 5–7 years after renal transplantation in early growth. Neuron 17: 641–654, 1996 childhood. Transplantation 67: 1043–1049, 1999 34. Harding AE: Clinical features and classification of inherited 30. Tessier-Lavigne M, Goodman CS: The molecular biology of ataxias. In: Advances in Neurology, Vol. 61, edited by Harding axon guidance. Science 274: 1123–1133, 1996 AE, Deufel T, New York, Raven, 1993, pp 1–14 31. Wright JW, Snyde MA, Schwinof KM, Combes S, Copenhaver 35. Shih N-Y, Li J, Karpitskii V, Nguyen A, Dustin ML, Kanawaga PF: A role for fasciclin II in the guidance of neuronal migration. O, Miner JH, Shaw AS: Congenital nephrotic syndrome in mice Development 126: 3217–3228, 1999 lacking CD2-associated protein. Science 286: 312–315, 1999 32. Dahme M, Bartsch U, Martini R, Anliker B, Schachner M, 36. Dustin ML, Olszowy MW, Holdorf AD, Li J, Bromley S, Mantei N: Disruption of the mouse L1 gene leads to malforma- Desai N, Widder P, Rosenberger F, van der Merwe P, Allen tions of the nervous system. Nat Genet 17: 346–349, 1997 PM, Shaw AS: A novel adaptor protein orchestrates receptor 33. Schuster CM, Davis GW, Fetter RD, Goodman CS: Genetic patterning and cytoskeletal polarity in T-cell contacts. Cell 94: dissection of structural and functional components of synaptic 667–677, 1998