31

　Original Contribution Kitasato Med J 2011;41:31-41　

Nephrin signal peptide sequence cDNA in expression vectordetermines antibody characteristics reacting on native or

glycosylation-disturbed nephrin protein in genetic immunization

Shokichi Naito,1 Togo Aoyama,1 Kouju Kamata2

1 Department of Internal Medicine, Kitasato University Graduate School of Medical Sciences2 Department of Nephrology in Internal Medicine, Kitasato University School of Medicine

Objectives: Nephrin is an essential protein for maintaining the normal structure of podocyte footprocess and glomerular filtration barrier. In this paper, we investigate the effect of signal peptidesequence cDNA on the characteristics of polyclonal anti-nephrin antibodies induced by geneticimmunization.Methods: Five fragments of nephrin cDNA with or without signal peptide sequence were inserted tothe pTARGETTM vector. Rats were immunized with these vectors by the gene-gun method. Seraswere analyzed by Western blotting, immunoprecipitation, flow cytometry and immunohistochemistry.Results: Four different antibodies induced by cDNA encoding nephrin protein fragments withoutsignal peptide showed antigen-site specific binding to fragmented nonglycosylated nephrin proteins.These antibodies also reacted to a deglycosylated full-length nephrin protein, but reacted to neither anative nephrin protein nor a fully-glycosylated conformational nephrin protein. Four different antibodiesinduced by cDNA encoding nephrin fragments with signal peptide showed an antigen-site specificbinding to non-glycosylated nephrin protein fragments. In additions, these antibodies reacted to botha native nephrin protein and a full-length glycosylated conformational nephrin protein.Conclusions: The absence of signal peptide sequence cDNA in the expression vector producedantibodies specific for glycosylation-disturbed proteins, while its presence also produced antibodiesthat bound to native or fully glycosylated conformation protein.

Key words: antibody specificity, genetic immunization, nephrin, N-glycosylation, signal peptide

Introduction

lomerular filtration barrier consisting of podocytes,glomerular basement membrane and endothelial

cells, and the presence of negatively charged proteins onall three of these components prevent leakage of theplasma proteins into urine.1 Recent advances show thatthe slit diaphragm between podocytes has a crucial rolein preventing proteinuria.2-4 Nephrin (NPHS1), whichwas cloned as a gene responsible for Finnish type nephriticsyndrome, shows transcription only in glomerularpodocytes.5-9 Mature nephrin protein is located on cellsurface of the glomerular podocytes, and is a majorcomponent of slit diaphragm.2,3,6,10-17

Nephrin has 8 extracellular immunoglobulin (Ig)-likedomains, a fibronectin type III domain, a transmembranedomain and a cytoplasmic C-terminal domain.5,11 Nephrin

consists of 1,241 amino acids (aa) with a calculatedmolecular mass of 132,532 Da. Nephrin appears as 180-200 kDa proteins when analyzed by sodium dodecylsulfate (SDS)-polyacrylamide gel electrophoresis(PAGE).16,18 This molecular weight (MW) differencedepends on the extent of N-linked glycosylation. Nephrincontains no O-linked glycosylation sites.16,18

A newly-produced nephrin protein in the cell isinitially glycosylated in the endoplasmic reticulum (ER)and additional glycosyl modification takes place in theGolgi apparatus, producing mature protein. When proteinis translated on ribosome from an mRNA and enters intoER, it is guided into ER by an N-terminal signal peptidewith the assistance of signal recognition particle (SRP)and SRP receptor.19 Signal peptide is species-specificand protein-specific. It is also essential for initialglycosylation of protein in ER and glycosyl modification

G

Received 15 January 2010, accepted 15 January 2010Correspondence to: Shokichi Naito, Department of Nephrology in Internal Medicine, Kitasato University School of Medicine1-15-1 Kitasato, Minami-ku, Sagamihara, Kanagawa 252-0373, JapanE-mail: snaito@med.kitasato-u.ac.jp

32

Naito, et al.

in the Golgi apparatus.19,20

We previously reported on the production andcharacterization of polyclonal anti-human nephrinantibodies induced by genetic immunization usingnephrin cDNA.21 Although four polyclonal antibodiesshowed antigen-site specific bindings to Escherichia coli-produced recombinant human nephrin protein, only twoof these reacted with native nephrin protein from humankidney tissue and full-length human nephrin proteinproduced by human embryonic kidney cell line (HEK293-NW). The reason why antigen recognition characteristicsof these polyclonal anti-nephrin antibodies are differentand the role played by signal peptide sequence in cDNAencoding a putative protein in an expression vectorremained unknown.21 In the present study, we evidencedthe effect of signal peptide sequence cDNA in anexpression vector on the characteristics of polyclonalanti-nephrin antibodies produced by geneticimmunization.

Materials and Methods

AnimalsSixty-four female Lewis rats purchased from the CharlesRiver Breeding Laboratory (Atsugi) at 7 weeks of agewere divided into 11 groups composed of 5-6 rats. Eachgroup of rats was for cDNA immunization as describedin Immunization schedules and sampling below. The

rats were kept in a dark-and-light cycle of 12 hours andfed standard laboratory food (Japan SLC, Inc., Shizuoka)with free access to water. All animal experiments wereperformed under an experimental protocol approved bythe Ethical Review Committee for Animal Experimentsof Kitasato University School of Medicine.

Immunization plasmidsFull-length human nephrin cDNA8,11 was synthesized andpurified by TaKaRa Custom Services (Ohtsu). Nephrinhas a COOH-terminal domain, 8 immunoglobulin (Ig)-like motifs with 10 N-linked glycosylation and afibronectin-like motif near the transmembrane region ofextracellular domain (Figure 1). We prepared ten differenthuman nephrin cDNA fragments encoding Ig-like motifs1-2 (aa 23-257), 3-4 (aa 258-458), 5-6 (aa 459-755), 7-8(aa 756-940) and 1-8 (aa 23-940) with or without 22 aanephrin signal peptide coding sequence preceding eachstructural cDNA fragment. These 10 different cDNAswere generated by PCR using full-length human nephrincDNA as a template. PCR products were cloned intopTARGETTM mammalian expression vectors (containinga CMV promoter [Promega, Madison, WI, USA]) by TAcloning. The 10 expression vectors were transformedinto E. coli JM 109 Competent Cells (TaKaRa) accordingto the manufacturer's protocol. After overnightincubation, each plasmid was purified using a QIAGENPlasmid Maxi Kit (QIAGEN, Tokyo). The authenticity

Nephrin has eight extracellular immunoglobulin (Ig)-like domains, a fibronectin type III domain, a transmembranedomain and a cytoplasmic C-terminal domain. Ten potential N-linked glycosylation site are marked as black hexagons.aa, amino acid; FNIII, fibronectin type III domain; TM, transmembrane domain

Figure 1. Human nephrin protein structure

33

of cDNA construct was confirmed by sequencing with aBigDye Terminator V3.1 Cycle Sequencing Kit (AppliedBiosystem, Framingham, MA, USA).

Immunization schedules and samplingEleven groups of 5-6 female Lewis rats were used. Thenephrin expression vectors were bound to gold particles(1μm in diameter [Bio-Rad, Hercules, CA, USA])according to the manufacturer's protocol. Gold particlescovered with plasmid vectors were shot into thesubcutaneous tissue of bilateral inner thighs using gene-gun method (Bio-Rad). For initial administration, 100μg of plasmid DNA was used, and subsequently 10μgplasmid DNA was administered, once a week, from weeks2 to 12. An expression vector without nephrin cDNAwas also administered to another group of rats in thesame manner, as a control experiment. Blood sampleswere taken every 2 weeks from the start of experimentup to week 14 and stored at -80℃ until use.

Glutathione-S-transferase (GST)-tagged proteinexpression and purification21

Full-length nephrin cDNA and 4 nephrin cDNAfragments encoding aa 23-257 (Ig-like motifs 1-2), aa258-459 (Ig-like motifs 3-4), aa 460-755 (Ig-like motifs5-6), or aa 756-940 (Ig-like motifs 7-8), were insertedinto a modified pGEX-5X-1 vector (GE Healthcare,Buckinghamshire, UK), respectively. GST-tagged fusionproteins were expressed in E. coli BL21 strain culturedin LB medium wi th 1 mM isopropyl -β-D-thiogalactopyranoside at 37℃. Cells were pelleted at3,000 g for 10 minutes at 4℃, and then resuspended inlysis buffer with phosphate buffer solution (PBS)containing 2% Triton X-100 and protease inhibitors. After1 freeze-thaw cycle, the samples were sonicated on ice,and then centrifuged at 13,000 g for 30 minutes at 4℃.The majority of the GST-tagged fusion proteins wereincluded in the cell pellets. Proteins were purified underdenaturing conditions with phosphate buffer containing8 M urea, and then allowed to refold by removing ureaby dialysis. GST-tagged fusion proteins were furtherpurified with a pre-packed Glutathione Sepharose 4Bcolumn (GE Healthcare, Little Chalfont, England)according to the manufacturer's instructions. MW ofGST-tagged Ig-like motifs of 1-2, 3-4, 5-6, and 7-8 were54, 48, 59, and 47 kDa, respectively. MW of GST-tagged fusion protein without fragmented nephrin proteinwas 27 kDa. These fusion proteins completely lacked N-glycosylation because N-glycosylation was absent inprokaryotes.30

V5-tagged protein expression and purification21

A cDNA encoding the full-length human nephrin proteinwith signal peptide sequence was cloned into mammalianexpression vector pcDNA3.1/V5 (Invitrogen Corp.,Carlsbad, CA, USA) and then transformed into HEK293-NW cells, a human embryonic kidney cell line, byelectroporation. HEK293-NW cells were cultured withRPMI (Roswell Park Memorial Institute)-1640 medium(Sigma Chemical, St. Louis, MO, USA) supplementedwith 10% (V/V) heat-inactivated FBS (fetal bovineserum), 100 U/ml penicillin, 100μg/ml streptomycinsulfate and 2 mM L-glutamine in 5% CO2 and 95% air at37℃. Stable clones were obtained by incubation withselection buffer containing 1 mg/ml G-418 sulfate (EMDBiosciences, La Jolla, CA, USA). HEK293-NW cellswith V5-tagged nephrin protein on their plasmamembranes were cloned and cultured under almostconfluent conditions. They were then washed 3 timeswith ice-cold PBS and lysed with 1ml of lysis buffer(140 mM NaCl, 3 mM MgCl2, 0.5% [v/v] Nonidet-P 40in 10 mM Tris-HCl [pH 8.4] and protease inhibitors).Cell lysates were removed by centrifugation at 14,000 gfor 5 minutes and then used for immunoprecipitationstudies. A glycosylated nephrin protein is produced bythese HEK-NW cells.

SDS-PAGE and western blot analysesE. coli-produced GST-tagged fusion proteins containingnonglycosylated nephrin protein fragments wereseparated using a 5-20% sodium dodecyl sulfate (SDS)gradient gel under reducing conditions. Afterelectrophoresis, proteins were stained with Coomassiebrilliant blue or transferred to polyvinylidene difluoride(PVDF) membranes. Membranes were incubated inblocking solution (PBS containing 20% [v/v] Block Ace[Dainippon Sumitomo Pharma, Osaka]) overnight at 4℃, and then reacted for 2 hours with 1/200 diluted seraderived from rats immunized with different fragments ofthe nephrin cDNA, at 14 weeks after immunization.Thereafter, membranes were incubated with horseradishperoxidase (HRP) conjugated rabbit anti-rat IgG (1/4000)(Zymed, South San Francisco, CA, USA) and visualizedwith chemiluminescent reagent, ECL ([enhancedchemiluminescence] Amersham Bioscience ,Buckinghamshire, UK).

Immunoprecipitation studiesHEK293-NW cell lysates containing full-lengthglycosylated human nephrin protein were incubated with100μl of Immunoprecipitation Starter Pack (AmershamBiosciences, Uppsala, Sweden; containing 50μl Protein

Effect of NPHS1 signal peptide cDNA

34

Naito, et al.

A Sepharose Fast Flow and 50μl Protein G SepharoseFast Flow as a 50% slurry) for 1 hour at 4℃ to avoidnonspecific binding of lysates to protein A and G beads.Immunoprecipitation was performed for 1 hour at 4℃by incubating 500μl of cell lysate and 50μl of seraderived from rats immunized with different fragments ofnephrin cDNA, at week 14 after immunization. Afteradding 80μl of the Immunoprecipitation Starter Pack,incubation was continued with gentle swirling for 1 hourat 4℃. Immune complexes on sepharose beads werecollected by centrifugation and then washed 3 times withice-cold lysis buffer and once with washing buffer(containing 140 mM NaCl, 0.5% (v/v) Nonidet-P 40 in10 mM Tris-HCl (pH 8.4) and protease inhibitors).Immunoprecipitated samples were suspended in Laemmlisample buffer and boiled for 5 minutes. Proteins wereseparated using 5-20% SDS-gradient gel under reducingconditions (Atto, Osaka), transferred to PVDFmembranes, immunostained with guinea pig antiserumraised against the intracellular domain of human nephrinprotein (Progen Biotechnik, Heidelberg, Germany), thenvisualized with HRP-labeled rabbit anti-guinea pig IgGAb (Cappel ICN, Aurora, OH, USA).

Tunicamycin treatment and immunoblot analysesTunicamycin inhibits N-linked glycosylation of nephrinprotein in the nephrin-producing cultured cells.16

HEK293-NW cells expressing glycosylated full-lengthnephrin protein were cultured to almost confluentcondition before tunicamycin treatment. Cells were thenwashed once with PBS and cultured for 20 hours in freshculture media in the absence or presence of tunicamycinin a dose-dependent manner. They were then washed 3times with ice-cold PBS and lysed with 1 ml of lysisbuffer containing 140 mM NaCl, 3 mM MgCl2, 0.5% (v/v) Nonidet-P 40 in 10 mM Tris-HCl (pH 8.4) and 0.1mM PMSF (phenylmethylsulfonyl fluoride). Whole celllysates were transferred into Eppendorf tubes, boiled,and centrifuged for 10 minutes before loading. Allsamples were subjected to SDS-PAGE with 5-20% SDS-gradient gels, and proteins were transferred onto PVDFmembranes. The membranes were subsequently blockedand incubated with guinea pig antiserum raised againstintracellular domain of human nephrin protein (ProgenBiotechnik, Heidelberg, Germany) and then visualizedwith HRP-labeled rabbit anti-guinea pig IgG Ab (CappelICN, Aurora, OH, USA).

Flow cytometry analysesFull-length glycosylated-nephrin protein produced byHEK293-NW cells was used as the antigen for detecting

reactivity of polyclonal anti-human nephrin antibodies.Sera obtained at 14 weeks from rats administered withhuman nephrin cDNA of 10 different fragments with orwithout signal peptide sequence were used for analysis.HEK293-NW cells (2 × 105) in 100μl Hank's bufferedsalt solution (HBSS) with 0.5% bovine serum albumin(BSA) and 10μl rat serum containing anti-human nephrinAb were mixed and incubated at 4℃ for 30 minutes.After centrifugation, the cell pellet was resuspended in100μl FITC-labeled rabbit anti-rat IgG (1/100)(GeneTex, San Antonio, TX, USA) and incubated at 4℃for 30 minutes. Finally, cells were washed with ice-coldHBSS twice and assayed using flow cytometry (BectonDickinson, Mountain View, CA, USA). Data obtainedwere analyzed with Cell Quest Software (BectonDickinson, Franklin Lakes, NJ, USA).

Indirect immunofluorescenceReactivities of rat anti-human nephrin antibodies tonormal human kidney tissue were tested with indirectimmunofluorescence. Normal human kidney tissue wasobtained from kidneys removed because of renalmalignancy. Small blocks of normal human kidney cortexwere taken from the opposite pole to renal malignancy,embedded in OCT (optimum cutting temperature)compound Tissue-Tek (Sakura Finetek, Torrance, CA,USA), and snap frozen in liquid nitrogen. The frozensamples were sliced into 5μm sections and then fixedwith ice-cold acetone for 5 minutes. After blocking with1% BSA/PBS for 1 hour, cryostat sections were overlaidwith rat sera containing anti-nephrin antibodies, andincubated at 4℃ overnight. After removing the rat sera,sections were reacted with FITC-labeled rabbit anti-ratIgG (GeneTex) for 1 hour. Then, after washing 3 timeswith PBS, they were evaluated using a fluorescencemicroscope equipped with appropriate filters (Olympus,Tokyo).

Results

Antigen recognition abilities of polyclonal anti-nephrinantibodies produced by immunization of nephrin cDNAfragments (with or without signal peptide sequence) innon-glycosylated fusion proteinsE. coli BL21-produced non-glycosylated fusion proteinsof Ig-like motifs 1-2, 3-4, 5-6, and 7-8 were used asantigens for Western blot analysis (Figure 2A,B). Antigenrecognition capabilities of the polyclonal anti-nephrinantibodies produced by immunization of nephrin cDNAfragments without signal peptide sequence encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 are shown in lanes

35

Non-glycosylated nephrin fusion protein fragments of Ig-like motifs 1-2, 3-4, 5-6, and 7-8 were produced by E. coliBL21, and used as antigens for Western blot analysis in both panels A and B. Antibodies produced by the nephrincDNA fragments encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 without the signal peptide sequence are shown inlanes 1, 2, 3, 4, and 5 of panel A, respectively. Antibodies induced by the nephrin cDNA fragments encoding Ig-likemotifs of 1-2, 3-4, 5-6, 7-8, and 1-8 with the signal peptide sequence are shown in lanes 1, 2, 3, 4, and 5 of panel B,respectively. Antisera taken at week 14 from the rats immunized with empty vectors are shown in lane 6 of panels Aand B.

Figure 2. Antigen recognition abilities of polyclonal anti-human nephrin antibodies against the E. coli-producednephrin fusion proteins

Glycosylated full-length nephrin protein produced by HEK293-NW cells wasused as the antigen in both of panels A and B. Antibodies induced by theadministration of cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, or 1-8without signal peptide sequence or empty vectors were applied in lanes 1, 2, 3,4, 5, and 6 of panel A. Antibodies induced by the administration of cDNAsencoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, or 1-8 with the signal peptide sequenceor empty vectors are shown in lanes 1, 2, 3, 4, 5, and 6 of panel B.

Figure 3. Immunoprecipitation of a glycosylated full-length nephrin by anti-nephrin antibodies induced by the nephrin cDNA fragments with or withoutthe signal peptide sequence

Effect of NPHS1 signal peptide cDNA

36

Naito, et al.

1, 2, 3, 4, and 5, respectively, in Figure 2A. Antigenrecognition capabilities of polyclonal anti-nephrinantibodies induced by the immunization of nephrin cDNAfragments with the signal peptide sequence encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 are shown in lanes1, 2, 3, 4, and 5 in Figure 2B. Thirty-four antibodiesproduced by immunization of cDNA encoding Ig-likemotifs 1-2, 3-4, and 5-6 (with or without the signal peptidesequence) showed antigen-site specific binding fornonglycosylated nephrin protein. Twelve antibodiesinduced by administration of nephrin cDNA fragmentencoding Ig-like motifs 7-8 (with or without the signalpeptide sequence cDNA) could not recognize anyfragment of nonglycosylated nephrin proteins of Ig-likemotifs 1-2, 3-4, 5-6, and 7-8. On the other hand, 12antibodies produced by immunization with Ig-like motifs1-8 cDNA (with or without signal peptide sequencecDNA) showed bindings to all of the nonglycosylatednephrin protein fragments (Ig-like motifs 1-2, 3-4, 5-6,and 7-8). Results for sera taken at week 14 from ratsadministered with the empty vector are shown in lane 6in Figure 2A and B. These sera did not recognize anyfragments of nonglycosylated nephrin protein.

Immunoprecipitation using glycosylated full-lengthnephrin protein and anti-nephrin antibodiesGlycosylated full-length nephrin protein (≅200 kDa),produced by HEK293-NW cells, was used as antigen asshown in Figure 3A and B and Figure 4. The antibodiesinduced by administration of cDNAs encoding Ig-like

motifs 1-2, 3-4, 5-6, 7-8, or 1-8 without signal peptidesequence, or empty vector were used in lanes 1, 2, 3, 4, 5,and 6, as shown in Figure 3A. Twenty-eight of 29antibodies could not precipitate glycosylated full-lengthnephrin protein, as shown in lanes 1-5. Each of 6antibodies produced by cDNA encoding Ig-like motifs1-8 without signal peptide sequence was used in lanes 1-6 in Figure 4. One exceptional antibody (rat No. 31)precipitated glycosylated full-length nephrin protein, asshown in lane 3. Antibodies produced by administration

Glycosylated full-length nephrin protein produced by HEK293-NW cells beforetunicamycin treatment was used in panel A, while another full-length nephrin proteinproduced by HEK293-NW cells after treatment with N-glycosylation inhibition by 0.5μg/ml of tunicamycin was used in panel B. Antibodies induced by nephrin cDNAencoding fragments without signal peptide sequence of Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 were applied in lanes 1, 2, 3, 4, and 5 of panels A and B, respectively. Anantibody induced by an empty vector was applied in lane 6. An antibody induced bynephrin cDNA encoding Ig-like motifs 1-8 with signal peptide sequence was used inlane 7.

Figure 5. Immunoprecipitation of a full-length nephrin protein before or aftertunicamycin treatment with the antibodies produced by cDNA encoding human nephrinprotein fragments without signal peptide

Glycosylated full-length nephrin protein produced by HEK293-NW cells was used as antigen. Each of 6 antibodies made byadministration of cDNA encoding Ig-like motifs 1-8 without signalsequence was applied in lanes 1, 2, 3, 4, 5, and 6. An antibodyinduced by administration of empty vector was used in lane 7.The antibody produced by administration of cDNA encoding Ig-like motifs 1-8 with signal peptide sequence was used in lane 8.The number in parenthesis designates the individual number ofthe rat studied.

Figure 4. Immunoprecipitation of glycosylated full-lengthnephrin protein by antibodies produced by cDNA encoding Ig-like motifs 1-8 without signal peptide

37

of cDNAs encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, or1-8 with signal peptide sequence, or empty vectors wereused in lanes 1, 2, 3, 4, 5, and 6 as shown in Figure 3B.All 23 antibodies induced by cDNAs encoding Ig-likemotifs 1-2, 3-4, 5-6, and 1-8 with signal peptide sequencedid precipitate glycosylated nephrin protein, as shown inlanes 1, 2, 3, and 5 of panel B in Figure 3 and lane 8, inFigure 4. However, all 6 antibodies induced by cDNAencoding Ig-like motifs 7-8 with signal peptide sequencecould not precipitate glycosylated nephrin protein, asshown in lane 4 of panel B in Figure 3.

Immunoprecipitation of nephrin produced by HEK293-NW cells before or after tunicamycin treatment usingantibodies induced by cDNA encoding nephrin proteinfragments without signal peptide sequenceImmunoblot analyses shows inhibitory effect of

tunicamycin on N-linked glycosylation of nephrin asevaluated by changes in MW of protein in culturedHEK293-NW cells. HEK293-NW cells producingglycosylated full-length nephrin protein were treated with0.01, 0.05, 0.1, 0.5, or 1.0μg/ml of tunicamycin for 20hours. MW of the nephrin protein moved from 200 kDatoward 180 kDa in a dose dependent manner. Inhibitionof N-linked glycosylation (as estimated by the presenceof 100% 180 kDa nephrin protein) occurred in doses of0.5μg/ml or above of tunicamycin.

Full-length human nephrin protein produced byHEK293-NW cells before (panel A) or after (panel B)treatment with 0.5μg/ml of tunicamycin to inhibit N-linked glycosylation was used as antigen for theseimmunoprecipitation experiments (Figure 5). Antibodiesinduced by nephrin cDNA fragments encoding Ig-likemotifs 1-2, 3-4, 5-6, 7-8, and 1-8 without signal peptide

Glycosylated full-length nephrin protein expressed on the plasma membrane of HEK293-NW cells was used as the antigen. Antibodiesinduced by cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 without signal peptide sequence were used in panels A1, A2, A3, A4,and A5, respectively. Serum from an exceptional rat (No. 31) immunized with cDNA encoding Ig-like motifs1-8 without signal peptidesequence was used in panel B5. Antibodies produced by cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 with a signal peptidesequence were used in panels C1, C2, C3, C4, and C5, respectively. The vertical bar shows the number of cell counts and the horizontalbar indicates the intensity of fluorescence. Cells that reacted with anti-human nephrin antibodies shifted to the right side of each panel.

Figure 6. Flow cytometric analyses of the binding of polyclonal anti-human nephrin antibodies to glycosylated nephrin protein on theplasma membrane of HEK293-NW cells

Effect of NPHS1 signal peptide cDNA

38

Naito, et al.

sequence were applied in lanes 1, 2, 3, 4, and 5,respectively. Six antibodies induced by an empty vectorwere applied in lane 6 as a negative control. Six antibodiesinduced by cDNA with signal peptide sequence of Ig-like motifs 1-8 were used in lane 7 as a positive control.Twenty-eight antibodies induced by nephrin cDNAencoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 (withone exception, that of rat No. 31) without signal peptidesequence could not precipitate glycosylated full-lengthnephrin protein (before tunicamycin treatment) as shownin lanes 1, 2, 3, 4, and 5 in panel A. Four of these 5different types of antibodies could precipitate full-lengthnephrin protein after inhibition of N-linked glycosylationby tunicamycin (lanes 1, 2, 3, and 5 of panel B). None ofthe 6 antibodies produced by the cDNA encoding Ig-likemotifs 7-8 without signal peptide sequence couldprecipitate full-length nephrin protein after N-linkedglycosylation inhibition by tunicamycin (lane 4 of panelB). Six sera from rats immunized with empty vectorprecipitated none of the full-length nephrin proteins (lane6 of panels A and B). Six sera of rats administered withnephrin cDNA fragment encoding Ig-like motifs 1-8 with

signal peptide sequence precipitated both of the full-lengthnephrin proteins before and after N-linked glycosylationinhibition by tunicamycin.

Flow cytometric detection of glycosylated nephrin byantibodies produced by nephrin cDNA fragments withor without signal peptidesGlycosylated full-length nephrin protein on the plasmamembrane of HEK293-NW cells was used to assay thereactivities of anti-nephrin antibodies using flowcytometry (Figure 6). Twenty-eight of 29 antibodiesproduced by cDNA fragments encoding Ig-like motifs1-2, 3-4, 5-6, 7-8, and 1-8 without signal peptide sequencedid not label the cells with a high intensity of fluorescence(panels A1, A2, A3, A4, and A5). Only one exceptionalantibody from rat No. 31 immunized with cDNA encodingIg-like motifs 1-8 without signal peptide sequencerecognized glycosylated full-length nephrin protein onthe plasma membrane of HEK293-NW cells and producedcells with a high intensity of fluorescence (panel B5).All of the 23 antibodies induced by cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, and 1-8 with signal peptide

Anti-nephrin antibodies induced by cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 without signal peptide sequence wereapplied in panels A1, A2, A3, A4, and A5, respectively. One exceptional antibody of rat No. 31 immunized with cDNA encoding Ig-likemotifs 1-8 without signal peptide sequence was applied in panel B5. Antibodies produced by cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 with a signal peptide sequence were applied in panels C1, C2, C3, C4, and C5, respectively.

Figure 7. Indirect immunofluorescence staining of normal human kidney tissue by polyclonal anti-human nephrin antibodies

39

sequence reacted with glycosylated full-length nephrinprotein on plasma membrane of HEK293-NW cells(panels C1, C2, C3, and C5). None of the 6 antibodiesmade by cDNA encoding Ig-like motifs 7-8 with signalpeptide sequence could react with glycosylated nephrinprotein on plasma membrane of HEK293-NW cells (panelC4).

Recognition of native nephrin by anti-human nephrinantibodies using indirect immunofluorescenceAnti-nephrin antibody reactivities against a native nephrinprotein in normal human kidney tissue were examined(Figure 7). Twenty-eight of 29 antibodies induced bycDNA encoding Ig-like motifs 1-2, 3-4, 5-6, 7-8, and 1-8 without signal sequence did not react with native nephrinprotein on normal human kidney tissue (panels A1, A2,A3, A4, and A5). One exceptional serum of rat No. 31showed binding to native nephrin protein in normal humankidney tissue (panel B5). All of the 23 antibodiesproduced by cDNA encoding Ig-like motifs 1-2, 3-4, 5-6, and 1-8 with signal peptide sequence recognized nativenephrin protein in normal human kidney tissue (panelsC1, C2, C3, and C5). None of the 6 antibodies made bycDNA encoding Ig-like motifs 7-8 with signal peptidesequence reacted with native nephrin protein (panel C4).

Discussion

The isolation of a sufficient amount of pure protein forimmunization is a time-consuming and sometimesdifficult process. We successfully produced highlyspecific polyclonal antibodies against human nephrinprotein fragments using a genetic immunizationtechnique.21 In this report, we also studied the effect ofincluding a cDNA encoding signal peptide sequence inthe immunization vector on the characteristics ofantibodies produced by genetic immunization.

Polyclonal anti-nephrin antibodies induced by nephrincDNA encoding Ig-like motifs 1-2, 3-4, 5-6, and 1-8with signal peptide sequence showed antigen-site specificreactivity against nonglycosylated nephrin proteinfragments and could also react to native and fully-glycosylated conformational nephrin protein. On theother hand, 28 of 29 antibodies induced by the nephrincDNA encoding Ig-like motifs 1-2, 3-4, 5-6, and 1-8without the signal peptide sequence also showed antigen-site specific binding to nonglycosylated nephrin proteinfragments but did not bind to either native or full-lengthglycosylated nephrin protein. These data indicate thatthe presence of signal peptide in the expressed motif ofprotein fragments is required to obtain antibodies that

react to native or full-length glycosylated nephrin protein.These data also suggest that antigen presentation mightoccur after N-linked glycosylation of the newly generatednephrin protein expressed by the vector containing cDNAencoding nephrin protein with signal peptide.

Twenty-eight of the 29 antibodies induced by nephrincDNA encoding Ig-like motifs 1-2, 3-4, 5-6, and 1-8without signal peptide sequence recognized neither nativenephrin protein nor full-length glycosylated nephrinprotein. These same antibodies could precipitate full-length nephrin protein on HEK293-NW cells aftertunicamycin treatment. As tunicamycin inhibitsglycosylation of nephrin protein,16 conformation ofnephrin protein might have changed, and some hiddenantigenic epitopes might have been revealed allowingthese antibodies to bind to the full-length nephrin protein.The antigen epitopes of nephrin protein for theseantibodies might include an asparagine residue that isinvolved in N-linked glycosylation.

One exceptional antibody from rat No. 31 producedby immunization against cDNA encoding Ig-like motif1-8 without signal peptide sequence reacted with bothfully glycosylated and deglycosylated nephrin protein.This antibody might react on a putative antigen epitopelocated on the surface of native nephrin protein in a regionwhose conformation remains unaffected by the presenceor absence of glycosylation. Antibodies against such anepitope are rarely produced by using cDNA of a proteinwithout signal peptide sequence in genetic immunization.

The administration of cDNA encoding Ig-like motifs7-8 with or without signal peptide sequence failed toproduce any reactive antibodies. The reason the structureof Ig-like motifs 7-8 might be poorly antigenic in DNAimmunization remains unknown.

The antibody induced by administration of cDNAencoding Ig-like motifs 1-8 with signal peptide sequencewas able to bind to Ig-like motifs 1-2, 3-4, 5-6, and 7-8 ofnonglycosylated nephrin proteins, full-lengthglycosylated nephrin protein on the cell membrane ofHEK 293-NW cells, and nephrin protein in normal humankidneys. This antibody may well be utilized forimmunoprecipitation, immunoblotting, and tissuestaining. It may also be the only antibody that reactswith Ig-like motifs 7-8 of nephrin protein.

The antibodies produced by genetic immunizationusing nephrin cDNA fragments encoding proteins withoutsignal peptide can bind to the deglycosylated nephrinprotein but are almost not able to bind to native andfully-glycosylated structural nephrin protein. Theseantibodies can work in immunoblotting but not in cellstaining or immunoprecipitation. This is consistent with

Effect of NPHS1 signal peptide cDNA

40

Naito, et al.

the results obtained with antibody produced byimmunization with short linear synthetic peptide of 5-20amino acids that worked in immunoblotting and cellstaining but not in immunoprecipitation.22 But antibodiesproduced by genetic immunization with cDNA encodingnephrin fragments with signal peptide could bind tonative, fully-glycosylated protein as well as todeglycosylated nephrin protein. These antibodies canwork in immunoblot t ing, cel l s ta ining, andimmunoprecipitation. On the other hand, antibodyproduced by immunization with a recombinant proteinwith folded protein structure also worked in cell stainingand immunoprecipitation but never worked inimmunoblotting.22 Consequently, two types ofqualitatively modulated antibodies produced by geneticimmunization with cDNA encoding nephrin proteins withor without signal peptides may provide a potent tool forfuture research in life science.

Glycosylation is essential for the proper folding andtransport of nephrin to the plasma membrane.16,23 Thesemodifications take place in the ER.19 The obstacle toprotein folding resulting from glycosylation disturbanceof nephrin leads to ER stress for glomerular epithelialcells.16,20,24 ER stress in glomerular epithelial cells maybe pathogenetically important in some of the kidneydisease models, such as passive Heymann's nephritis,puromycin aminonucleoside nephrosis, and megsintransgenic rat and anti-Thy1 nephritis.24-29 Newly-produced polyclonal anti-nephrin antibodies induced bynephrin proteins without signal peptide sequence onlyrecognize nephrin protein with disturbance of its N-linkedglycosylation. These antibodies could be a valuable toolfor the evaluation of ER stress in glomerular epithelialcells of diseased human kidney.

In conclusion, DNA immunization with cDNAencoding a nephrin protein fragment with or withoutsignal peptide sequence produced distinct polyclonalantibodies against native or deglycosylated nephrinprotein, respectively. The effect of cDNA encoding signalpeptide on antibody characteristics in geneticimmunization was also evidenced.

Acknowledgments

We thank Dr. Michihito Okubo for helpful suggestionsfor this manuscript and Mrs. Naoko Ishigaki for technicalassistance.

References

1.Anderson S, Komers R, Brenner BM. The anatomicbarrier to proteinuria: the glomerular capillary wall.In: Brenner BM, editor. Volume 1: Brenner andRector's The Kidney, 8th ed. Philadelphia: Saundersand Elsevier; 2008;820-1.

2. Topham PS, Kawachi H, Hayda SA, et al.Nephritogenic mAb 5-1-6 is directed at theextracellular domain of rat nephrin. J Clin Invest1999;104:1559-66.

3. Orikasa M, Matsui K, Oite T, et al. Massiveproteinuria induced in rats by a single intravenousinjection of a monoclonal antibody. J Immunol1998;141:807-14.

4. Kawachi H, Miyauchi N, Suzuki K, et al. Role ofpodocyte slit diaphragm as a filtration barrier.Nephrology (Carlton) 2006;11:274-81.

5. Tryggvason K, Patrakka J, Wartiovaara J. Hereditaryproteinuria syndromes and mechanisms ofproteinuria. N Engl J Med 2006;354:1387-401.

6. Ruotsalainen V, Ljungberg P, Wartiovaara J, et al.Nephrin is specifically located at the slit diaphragmof glomerular podocytes. Proc Natl Acad Sci U S A1999;96:7962-7.

7. Holzman LB, St John PL, Kovari IA, et al. Nephrinlocalizes to the slit pore of the glomerular epithelialcell. Kidney Int 1999;56:1481-91.

8. Holthofer H, Ahola H, Solin ML, et al. Nephrinlocalizes at the podocyte filtration slit area and ischaracteristically spliced in the human kidney. Am JPathol 1999;155:1681-7.

9. Kawachi H, Koike H, Kurihara H, et al. Cloning ofrat nephrin: expression in developing glomeruli andin proteinuric states. Kidney Int 2000;57:1949-61.

10. Ruotsalainen V, Reponen P, Khoshnoodi J, et al.Monoclonal antibodies to human nephrin. HybridHybridomics 2004;23:55-63.

11. Gerke P, Huber TB, Sellin L, et al. Homodimerizationand heterodimerization of the glomerular podocyteproteins nephrin and NEPH1. J Am Soc Nephrol2003;14:918-26.

12. Patrakka J, Ruotsalainen V, Reponen P, et al.Recurrence of nephrotic syndrome in kidney graftsof patients with congenital nephrotic syndrome ofthe Finnish type: role of nephrin. Transplantation2002;73:394-403.

13. Wang SX, Ahola H, Palmen T, et al. Recurrence ofnephrotic syndrome after transplantation in CNF isdue to autoantibodies to nephrin. Exp Nephrol2001;9:327-31.

14. Kawachi H, Koike H, Shimizu F. mAb 5-1-6nephropathy and nephrin. Microsc Res Tech2002;57:236-40.

41

15. Wartiovaara J, Ofverstedt LG, Khoshnoodi J, et al.Nephrin strands contribute to a porous slit diaphragmscaffold as revealed by electron tomography. J ClinInvest 2004;114:1475-83.

16. Yan K, Khoshnoodi J, Ruotsalainen V, et al. N-linked glycosylation is critical for the plasmamembrane localization of nephrin. J Am Soc Nephrol2002;13:1385-9.

17. Khoshnoodi J, Sigmundsson K, Ofverstedt LG, et al.Nephrin promotes cell-cell adhesion throughhomoph i l i c i n t e r ac t ions . Am J Pa tho l2003;163:2337-46.

18. Khoshnoodi J, Hill S, Tryggvason K, et al.Identification of N-linked glycosylation sites inhuman nephrin using mass spectrometry. J MassSpectrom 2007;42:370-9.

19. Gerace L, Gilmore R, Johnson A, et al. Intracellularcompartments and protein sorting: the endoplasmicreticulum. In: Alberts B, Johnson A, Lewis J, et al.editors. Molecular Biology of The Cell, 4th ed. NewYork: Newton Press; 2004;689-709.

20. Kitamura M. Endoplasmic reticulum stress in thekidney. Clin Exp Nephrol 2008;12:317-25.

21. Aoyama T, Kamata K, Yamanaka N, et al.Characteristics of polyclonal anti-human nephrinantibodies induced by genetic immunization usingnephrin cDNA. Nephrol Dial Transplant2006;21:1073-81.

22. Harlow E, Lane D. Epitope mapping. In: Harlow E,Lane D, editors. Using Antibodies: A LaboratoryManual. New York: Cold Spring Harbor LaboratoryPress; 1999;382-5.

23. Fujii Y, Khoshnoodi J, Takenaka H, et al. The effect

of dexamethasone on defective nephrin transportcaused by ER stress: a potential mechanism for thetherapeutic action of glucocorticoids in the acquiredglomerular diseases. Kidney Int 2006;69:1350-9.

24. Nakajo A, Khoshnoodi J, Takenaka H, et al.Mizoribine corrects defective nephrin biogenesis byrestoring intracellular energy balance. J Am SocNephrol 2007;18:2554-64.

25. Inagi R, Kumagai T, Nishi H, et al. Preconditioningwith endoplasmic reticulum stress amelioratesmesangioproliferative glomerulonephritis. J Am SocNephrol 2008;19:915-22.

26. Cybulsky AV, Takano T, Papillon J, et al. Role ofthe endoplasmic reticulum unfolded protein responsein glomerular epithelial cell injury. J Biol Chem2005;280:24396-403.

27. Cybulsky AV, Takano T, Papillon J, et al.Complement C5b-9 membrane attack complexincreases expression of endoplasmic reticulum stressproteins in glomerular epithelial cells. J Biol Chem2002;277:41342-51.

28. Inagi R, Nangaku M, Onogi H, et al. Involvement ofendoplasmic reticulum (ER) stress in podocyte injuryinduced by excessive protein accumulation. KidneyInt 2005;68:2639-50.

29. Bijian K, Cybulsky AV. Stress proteins in glomerularepithelial cell injury. Contrib Nephrol 2005;148:8-20.

30. Emr S, Kelly R, Pelham H, et al. What is the purposeof glycosylation? In: Alberts B, Johnson A, Lewis J,et al. editors. Molecular Biology of the Cell, 4th ed.New York: Newton Press; 2004;735-6.

Effect of NPHS1 signal peptide cDNA