The Death Domain of Kidney Ankyrin Interacts with Fas and Promotes Fas-Mediated Cell Death in Renal Epithelia
Marcela Del Rio*,
Abubakr Imam*,
Maryely DeLeon*,
Gary Gomez*,
Jaya Mishra,
Qing Ma,
Samir Parikh and
Prasad Devarajan*,
*Department of Nephrology, Children’s Hospital at Montefiore, Albert Einstein College of Medicine, Bronx, New York, and Department of Nephrology and Hypertension, Cincinnati Children’s Hospital Medical Center, University of Cincinnati, Cincinnati, Ohio.
Correspondence to Dr. Prasad Devarajan, Nephrology and Hypertension, MLC 7022, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45229-3039. Phone: 513-636-4531; Fax: 513-636-7407;
ABSTRACT. Ankyrins are a ubiquitously expressed family of conservedproteins that mediate the linkage of integral membrane proteinssuch as transporters and channels with the underlying cytoskeleton.Ankyrins possess a conserved death domain, the functional significanceof which has remained puzzling. In this study, the death domainof AnkG190, the isoform of ankyrin expressed in kidney tubules,was used as bait in a yeast two-hybrid screen to identify interactingpartners. One of these interactions was with the proapoptoticmolecule Fas. This was confirmed by coimmunoprecipitation, colocalization,and glutathione S-transferase pull-down assays in cultured renalepithelial (MDCK) cells. Site-directed mutagenesis of a conservedarginine (R1496 in AnkG190), previously shown to be criticalfor the binding of Fas (R234 in Fas) to FADD, abolished theinteraction of ankyrin’s death domain with Fas. Overexpressionof constructs containing ankyrin’s death domain promotedFas-mediated apoptosis in MDCK cells. The linkage between ankyrinand Fas was confirmed in vivo in mouse kidney tubule cells bycoimmunoprecipitation and colocalization. In an establishedmouse model of renal ischemia-reperfusion injury characterizedby apoptotic tubule cell death, the expression of both ankyrinand Fas was markedly induced, and the interaction between thesemolecules remained intact. The results identify a novel tetheringinteraction between ankyrin and Fas in kidney epithelia andsuggest that AnkG190 may play a role as an adapter moleculein renal tubule cell death. E-mail: prasad.devarajan@cchmc.org.
Tethering interactions between membrane proteins and the underlyingspectrin-based cytoskeleton play key roles in several cellularactivities, including organization of plasma membrane domains(1). Ankyrins are a ubiquitously expressed family of conservedproteins that have emerged as critical adapter molecules mediatingsuch linkages because they possess binding sites for spectrinas well as an increasing number of integral membrane proteins(1–5). Three distinct ankyrin genes encode for a varietyof alternatively spliced and tissue-specific isoforms. Althoughthe ANK1 and ANK2 gene products are largely restricted to redcells and brain, respectively, the ANK3 gene transcribes isoformsthat display a general tissue distribution and are hence termedAnkG. These include a 480-kD isoform localized at the axonalinitial segment and node of Ranvier (6), a 190/210-kD kidneyankyrin isoform expressed at the plasma membranes of kidneytubule cells (7,8), and truncated isoforms associated with theGolgi apparatus (9,10), lysosomes (11), and sarcoplasmic reticulum(12). The majority of ankyrins described to date are modularproteins comprising three conserved domains, including an amino-terminaldomain containing a varying number of ankyrin repeats, a spectrin-bindingdomain, and a death domain located near the carboxyl terminal(1–8). Several structurally and functionally diverse proteinsinteract with the repeats domain of ankyrin, including -Na,K-ATPase, anion exchangers, the voltage-dependent sodium channel,sodium/calcium exchanger, calcium channels, IP3 receptor, ryanodinereceptor, clathrin, tubulin, and cell adhesion molecules suchas CD44 and the L1 family (1–5).
The "death domain" was initially reported as a region of sequencehomology within the intracellular portions of the proapoptoticreceptors Fas and TNFR1 (13,14). These domains were involvedin protein-protein interactions, enabling the subsequent identificationof several additional key death domain-containing proapoptoticproteins such as FADD, TRADD, and RIP (15–18). Databasesearches have since identified over a dozen proteins that possessthe death domain, including red cell ankyrin (14,19). The deathdomain within kidney ankyrin (8) displays an even greater similarityto those within proapoptotic molecules than that of red cellankyrin. However, the functional significance of ankyrin’sdeath domain has hitherto remained puzzling (3). In this study,we show that kidney ankyrin’s death domain interacts withthe death domain of Fas and promotes Fas-mediated apoptosisin renal epithelial cells both in vitro and in vivo.
Yeast Two-Hybrid Screen
The yeast two-hybrid screen has proven to be a valuable techniquefor the study of protein-protein interactions involving deathdomains (20). We used the Matchmaker Two-Hybrid System (Clontech,La Jolla, CA) as recommended by the manufacturer. The deathdomain of AnkG190, the isoform of ankyrin that is expressedat the plasma membranes of kidney tubule cells (7,8), was usedas bait. On the basis of published sequence homologies (13–18),a cDNA clone spanning bp 4449 to 4691, or residues 1479 to 1559of AnkG190 (GenBank accession number AF069525) (8) was insertedinto the DNA binding domain vector pGBXT7 (DNA/BD). A premadehuman kidney cDNA library cloned into the activating domainvector pGADT7 (AD/library) was obtained from Clontech. To verifythe absence of autonomous activation, the DNA/BD and AD/libraryconstructs were independently transformed into the yeast AH109strain, with parallel positive and negative controls as recommendedby the manufacturer. To assess for protein-protein interactions,the DNA/BD and AD/library were cotransformed in high stringencymedium (SD/-Ade/-His/-Tre/X-GAL). Positive clones were isolatedand the plasmids purified, transformed into Escherichia coli,and sequenced to determine their identity.
Coimmunoprecipitations In Vitro
MDCK cells (American Type Culture Collection, Manassas, VA)were cultured in complete Dulbecco modified Eagle medium with10% FBS, and coimmunoprecipitations performed as described previously(21,22). These cells were chosen because they represent kidneydistal/collecting tubule epithelial cells that normally expressall of the proteins pertinent to this study, including AnkG190,Fas, and FADD (8–10,21–24). Briefly, cells werelysed for 20 min at 4°C in immunoprecipitation buffer (10mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 0.5% deoxycholate,1% Nonidet P-40, and 1x complete protease inhibitor); the lysateswere precleared with preimmune serum and a 50% protein A-Sepharosesolution and incubated overnight at 4°C with either preimmuneserum or a polyclonal antibody to AnkG190 (22). After an additionalincubation with a 50% protein A-Sepharose solution for 2 h,the lysates were centrifuged and washed, and pellets were subjectedto SDS-PAGE and Western blot analysis with antibodies to AnkG190at 1:500 (22), Fas at 1:2000, and FADD at 1:200 (both from TransductionLaboratories, Lexington, KY). Aliquots of cell lysates wereprobed with antibody to tubulin (1:10,000; Sigma, St. Louis,MO) before precipitation to verify equal loading of samples.
Glutathione S-Transferase Pull-Down Assay
Ankyrin binding assays were performed as described previously(21–24). Briefly, the death domain of ankyrin (DD) wasexpressed in bacteria as a fusion protein with glutathione S-transferase(GST) using the pGEX prokaryotic expression system (Pharmacia,Nutley, NJ) and purified with glutathione-agarose. GST alonewas expressed as a control peptide. Proteins were analyzed bySDS-PAGE followed by Coomassie blue staining. Each fusion protein(GST or GST-DD, 50 µg) was conjugated to 50 µl ofa 50% glutathione-agarose slurry for 1 h at 4°, and incubatedovernight with 1 ml of cell lysate obtained by extracting MDCKcells with a buffer containing 10 mM PIPES, 500 mM NaCl, 300mM sucrose, 3 mM MgCl2, 0.5% Triton X-100, and 1x complete proteaseinhibitors. The beads were pelleted, washed, and analyzed byWestern blotting with antibodies as above.
Colocalization In Vitro
Colocalization studies were performed as described previously(9,10,22). Briefly, MDCK cells cultured to confluence on glassslides were fixed with acetone, blocked in goat serum for 30min, simultaneously incubated in primary antibodies to AnkG190(polyclonal, 1:200) and Fas (monoclonal, 1:2500) for 1 h, washed,incubated in secondary antibodies conjugated to Cy2 and Cy3(Amersham, Piscataway, NJ), and visualized with a fluorescencemicroscope (Zeiss Axiophot).
Site-Directed Mutagenesis
Because a conserved arginine residue within the death domainof Fas (R234) has been shown to be critical for Fas-FADD interactions(25), the corresponding arginine within the death domain ofAnkG190 (R1496) was substituted to alanine (AGG to GCG) by usingrecombinant PCR (26). Briefly, a primary PCR product was obtainedwith a sense primer targeting the 5' end of ankyrin’sdeath domain and an antisense primer containing the AGG to GCGsubstitution. A second PCR product was obtained using a primercontaining the substitution as the sense primer and an antisenseprimer targeting the 3' end of the death domain. The PCR productswere then used as templates to amplify a "zipper" product withthe outside primers. The final product was sequenced to verifythe presence of the R to A substitution and cloned into thepGEX system for GST pull-down assays.
Eukaryotic Overexpression as Green Fluorescence Protein Fusion
The original death domain of AnkG190 (DD) and the mutated deathdomain (R) were cloned into the eukaryotic expression vectorpEGFP-N1 (Clontech) for expression of fusion proteins with greenfluorescence protein (GFP). MDCK cells were stably transfectedwith either of these constructs under G418 selection (400 µg/ml;Invitrogen, Carlsbad, CA) as described previously (8–10).Stable expression was confirmed by immunofluorescence, and transfectantswere subjected to immunoprecipitation with antibody to GFP (Clontech),followed by Western blot analysis with antibodies to GFP, Fas,and FADD.
AnkG190 Constructs, Eukaryotic Overexpression, and Apoptosis Assays
The engineering, stable transfection, and expression of a seriesof overlapping AnkG190 constructs, each containing the eight-residueFLAG tag (Kodak, New Haven, CT), using the pcDNA3 eukaryoticexpression system (Invitrogen), have been previously reported(8). Two additional clones were constructed for this study usingPCR. First, a construct encoding for only the death domain anda FLAG tag was amplified by PCR. Second, a "loop-out" PCR technique(26) was used to engineer a regulatory domain of AnkG190 devoidof the death domain. Constructs were stably transfected intoMDCK cells, and their expression was verified by Western blotanalysis as described previously (8). Stable transfectants wereexposed to stimulatory Fas monoclonal antibody (clone DX2, Clontech)at 200 ng/ml for 4 h and subjected to the DNA laddering assayfor detection of apoptosis as described previously (21). Briefly,cells were resuspended in lysis buffer (1% SDS, 25 mM EDTA,1 mg/ml proteinase K, pH 8) at 50°C overnight, digestedwith ribonuclease A (10 mg/ml), and the chromosomal DNA extractedand analyzed by agarose gel electrophoresis. To confirm andquantitate apoptosis, we performed the terminal deoxynucleotidyltransferase (TUNEL) assay (ApoAlert Assay Kit; Clontech) asdescribed previously (21). Briefly, cells grown on coverslipswere fixed with 4% formaldehyde for 30 min at 4°C, permeabilizedwith 0.2% Triton X-100 for 15 min at 4°C, incubated witha mixture of nucleotides and TdT enzyme for 60 min at 37°Cin a dark, humidified chamber, the reaction terminated with2x SSC, and the coverslips mounted on glass slides. Apoptoticnuclei were detected by visualization with a fluorescence microscopeand quantitated as a percentage of all nuclei visualized byphase contrast microscopy. Only cells that displayed the characteristicmorphology of apoptosis, including nuclear fragmentation, nuclearcondensation, and intensely fluorescence nuclei by TUNEL assay,were counted as apoptotic. Merely TUNEL-positive cells, in theabsence of morphologic criteria, were not considered apoptotic.As another confirmatory assay, we used the annexin V-FITC cellmembrane labeling assay (ApoAlert Annexin V Kit, Clontech) todetect translocation of phosphatidylserine from the inner faceof the plasma membrane to the outer cell surface, where it bindsan annexin V-FITC conjugate and serves as an early and specificmarker of apoptosis (21). Briefly, cells grown on coverslipswere washed, incubated with annexin V-FITC for 15 min at roomtemperature in the dark, and visualized by fluorescence microscopy.Apoptosis was quantitated as the number of annexin positivecells per 100 cells examined.
Mouse Model of Renal Ischemia-Reperfusion Injury
We used well established murine models in which the structuraland functional consequences of brief periods of renal ischemiahave been previously documented (27–30). Briefly, maleSwiss-Webster mice (Taconic Farms, Germantown, NY) weighing25 to 35 g were housed with a 12:12 h light:dark cycle and wereallowed free access to food and water. The animals were anesthetizedwith sodium pentobarbital (50 mg/kg intraperitoneally) and placedon a warming table to maintain a rectal temperature of 37°C.The renal pedicles were occluded with a nontraumatic vascularclamp for 30 min, the clamps released, the kidney observed forreturn of blood flow, the incision sutured, and the mice allowedto recover in a warmed cage. After 0, 3, 12, or 24 h of reperfusion,the animals were reanesthetized, and blood obtained by punctureof the inferior vena cava for serum creatinine determinationby quantitative colorimetric assay kit (Sigma, St. Louis, MO).The animals were killed, the kidneys perfusion-fixed in situwith 4% paraformaldehyde in PBS, and both kidneys harvested.At least five separate animals were examined at each of thereflow periods. Half of one kidney was snap-frozen in liquidnitrogen and stored at -70°C until further processing; asample was fixed in formalin, embedded in paraffin, and sectioned(4 µm). Paraffin sections were stained with hematoxylinand eosin and examined histologically as well as by TUNEL stainingfor apoptosis as described previously (28–30).
For the TUNEL assay, we used the ApoAlert DNA FragmentationAssay Kit (Clontech). Paraffin was removed from the sectionsby xylene and descending grades of ethanol. The sections werefixed with 4% formaldehyde/PBS for 30 min at 4°C, permeabilizedwith proteinase K at room temperature for 15 min and 0.2% TritonX-100/PBS for 15 min at 4°C, and incubated with a mixtureof nucleotides and TdT enzyme for 60 min at 37°C. The reactionwas terminated with 2x SSC and the sections washed with PBS,and they were mounted with Crystal/mount (Biomeda, Foster City,CA). TUNEL-positive apoptotic nuclei were detected by visualizationwith a fluorescence microscope. Only cells that displayed thecharacteristic morphology of apoptosis, including nuclear fragmentation,nuclear condensation, and intensely fluorescence nuclei by TUNELassay, were counted as apoptotic. Merely TUNEL-positive cells,in the absence of morphologic criteria, were not consideredapoptotic.
The other half of one kidney was embedded in optimal cuttingtemperature compound (Tissue-Tek; Miles, Naperville, IL), andfrozen sections (4 µm) were obtained for immunohistochemistry.The second kidney was processed for Western blotting and immunoprecipitationsas follows. Whole kidneys were homogenized in ice-cold lysisbuffer (20 mM Tris, pH 7.4, 250 mM sucrose, 150 mM NaCl, 1%NP-40, and 1x complete protease inhibitors) with a Polytronhomogenizer. The homogenates were incubated on ice for 30 min,centrifuged at 1000 x g for 5 min at 4°C to remove nucleiand cellular debris, and analyzed for protein content by theBradford assay (Bio-Rad, Hercules, CA). By means of this mousemodel, others (27) and we (28–30) have previously documentedthe presence of tubule cell apoptosis and necrosis, and theinduction of Fas in apoptotic tubule cells.
Coimmunoprecipitations In Vivo
Whole kidney lysates were precleared with a 50% protein A-Sepharosesolution and incubated overnight at 4°C with a monoclonalantibody to AnkG190 (Zymed, San Francisco, CA). After an additionalincubation with a 50% protein A-Sepharose solution for 2 h,the lysates were centrifuged and washed, and pellets were subjectedto SDS-PAGE and separate Western blot analysis with polyclonalantibodies to AnkG190 (22), Fas (Santa Cruz Biotechnology, SantaCruz, CA), or FADD (Santa Cruz Biotechnology). Aliquots of celllysates were probed with antibody to tubulin (1:10,000, Sigma)before precipitation to verify equal loading of samples.
Colocalization In Vivo
Colocalization studies for ankyrin and Fas were performed inmouse kidneys as described previously (28–30). Briefly,frozen sections were permeabilized with 0.2% Triton X-100 inPBS for 10 min, blocked with goat serum for 1 h, and incubatedwith primary antibodies to AnkG190 (polyclonal) and Fas (monoclonal)for 1 h at room temperature. Slides were then incubated withsecondary antibodies conjugated with either Cy5 (for Fas) orCy3 (for ankyrin) and visualized with rhodamine or fluoresceinfilters, respectively. For colocalization of Fas with apoptoticcells, serial sections were subjected to either the TUNEL assayor to immunohistochemistry as described previously (28,29).Visualization with rhodamine filters revealed cells that stainedpositive for Fas, and examination with fluorescein filters identifiedTUNEL-positive nuclei in serial sections.
Identification of Proteins that Interact with Kidney Ankyrin’s Death Domain
The yeast two-hybrid system was used to screen for proteinsthat interact with the death domain of kidney ankyrin. The identifiedproteins and their putative function are shown in Table 1. Themost abundant interacting clones encoded the death domain ofFas, a proapoptotic protein belonging to the TNF receptor family(13,14). Several clones representing the death domain of AnkG190were also isolated, suggesting an ability of this peptide todimerize with itself. Interestingly, all identified proteinsplay putative roles in regulation of apoptosis. For this study,we chose to further characterize the interaction between thedeath domains of AnkG190 and Fas.
Table 1. Proteins interacting with the death domain of AnkG in a yeast two-hybrid assay
Ankyrin’s Death Domain Forms Complexes with Fas and FADD In Vitro
The ability of kidney ankyrin’s death domain to interactwith Fas was tested in MDCK cells. These cells were chosen becausethey represent kidney distal/collecting tubule epithelial cellsthat normally express all of the proteins pertinent to thisstudy, including AnkG190, Fas, and FADD (8–10,21–24).Furthermore, these cells have been shown to respond to partialATP depletion by inducing Fas and FADD and undergoing programmedcell death (21). By using recently described antibodies (22),AnkG190 was found to form a specific complex with Fas and FADDin MDCK cell lysates, as shown in Figure 1A. This complex wasabsent from lysates immunoprecipitated with preimmune serum(Figure 1A, Pre). This interaction was confirmed using GST pull-downassays, in which a fusion peptide of GST and ankyrin’sdeath domain specifically interacted with Fas and FADD, as shownin Figure 1B. The GST peptide alone did not form a complex withFas or FADD.
Figure 1. Ankyrin’s death domain forms complexes with Fas and FADD in MDCK cells. (A) Results of immunoprecipitation. Lys, MDCK lysate alone as positive control; Ank, AnkG190 antibody; Pre, preimmune serum as negative control. Equal aliquots of lysates were probed with anti-tubulin antibody (Tub) to verify equal loading of samples. Blots were probed with antibodies as shown on the right. Molecular weight markers are shown on the left. (B) Results of GST pull-down assay. CB, Coomassie blue stain of purified GST alone (GST), or the GST-ankyrin death domain fusion (DD). The blots are representative of three separate experiments.
Ankyrin Colocalizes with Fas In Vitro
It was next of interest to assay for the intracellular localizationof ankyrin and Fas. MDCK cells were double stained with polyclonalantibodies to AnkG190 and monoclonal antibodies to Fas. AnkG190in these polarized epithelial cells was largely restricted tothe plasma membranes (Figure 2, red), whereas Fas was presentboth at the cell surface and in an intracellular distribution(Figure 2, green). The merged images revealed that ankyrin colocalizeswith Fas predominantly at the plasma membrane (Figure 2, yellow).
Figure 2. Ankyrin colocalizes with Fas. MDCK cells double-stained with antibodies to AnkG190 (polyclonal, red) and Fas (monoclonal, green). The merged image shows that ankyrin partially colocalizes with Fas (yellow). Figure represents four independent experiments.
A Conserved Arginine Residue Is Critical for Ankyrin-Fas Interaction
We next defined the residues within AnkG190 that were criticalfor interaction with Fas. An alignment of protein sequenceswithin several known death domains is shown in Figure 3. Wenoted that of the four residues on Fas (arrows) required forFADD binding (25), only one (R234 in Fas and R1496 in AnkG190)is fully conserved in all death domains examined. By means ofsite-directed mutagenesis, we accomplished a R1496A substitutionin AnkG190. A GST fusion containing this substitution failedto interact with Fas and FADD in GST pull-down assays of MDCKcell lysates, as shown in Figure 4A, labeled R. To confirm thisobservation, we performed stable transfections of MDCK cellswith either wild-type or mutated (R1496A substitution) AnkG190death domain cloned in the eukaryotic expression vector pEGFP-N1.Immunoprecipitation of lysates with GFP antibody revealed afunctional complex between the AnkG190 death domain, Fas, andFADD in cells transfected with the wild-type construct (Figure 4B,labeled DD). This interaction was abolished by the R1496Asubstitution (Figure 4B, labeled R).
Figure 3. Sequence comparison reveals that of the four residues on Fas (arrows) required for FADD binding (25), only one (R234 in Fas and R1496 in AnkG190) is fully conserved in all death domains examined.
Figure 4. A conserved arginine residue is critical for ankyrin-Fas interactions. (A) Results of glutathione S-transferase (GST) pull-down assay. CB, Coomassie blue stain of purified GST-ankyrin death domain fusion (DD) or the mutant protein containing the R1496A substitution (R). Blots were probed with antibodies as shown on the right. Molecular weight markers are shown on the left. Antibodies used are shown on the right. (B) Results of immunoprecipitation with GFP antibody of cells stably transfected with either ankyrin’s death domain (DD) or the R1496A substitution (R) cloned into the expression vector pEGFP-N1. Lysates were probed with anti-tubulin (Tub) before immunoprecipitation and with anti-GFP antibody after immunoprecipitation to verify equal loading of samples. The blots represent three separate experiments.
The Death Domain of Kidney Ankyrin Promotes Fas-Mediated Apoptosis In Vitro
To explore the role of ankyrin-Fas interactions, we used MDCKcells stably transfected with a series of constructs spanningthe coding region of AnkG190, as illustrated in Figure 5. Theengineering, transfection, and expression of these constructshave been previously reported (8). In preliminary experiments,the cell lines demonstrated comparable viability, and the overexpressionof AnkG190 constructs containing the death domain did not initself induce apoptosis (data not shown). We have also previouslyshown that although nontransfected MDCK cells retain Fas-dependentapoptotic pathways, stimulatory monoclonal Fas antibodies resultedin DNA laddering only after prolonged exposure (21). We testedthe ability of AnkG190 constructs to promote Fas-mediated apoptosisafter short incubation periods. As shown in Figure 5, cellstransfected with constructs containing the death domain underwentapoptosis after only 4 h of stimulatory Fas antibody incubation.This included constructs I (full-length AnkG190), V (regulatorydomain of AnkG190 with the death domain included), and VI (deathdomain of AnkG190 alone). However, the constructs varied intheir ability to induce Fas-mediated apoptosis, as illustratedin Figure 6.
Figure 5. Ankyrin’s death domain promotes Fas-mediated apoptosis. The panel on the left shows the various AnkG190 constructs (labeled I to VII) used in this study. Each construct was tagged with the FLAG epitope at the 3' end. MDCK cells were stably transfected with each construct, exposed to stimulatory Fas monoclonal antibody for 4 h, and subjected to the DNA laddering assay as shown on the right panel. Results were reproducible in three experiments.
Figure 6. Ankyrin’s death domain promotes Fas-mediated apoptosis. MDCK cells were stably transfected with each construct as indicated, exposed to stimulatory Fas monoclonal antibody for 4 h, and subjected to the annexin assay (top panel) or terminal deoxynucleotidyl transferase (TUNEL) assay (bottom panel). The number of apoptotic (annexin-positive cells or TUNEL/morphology-positive nuclei) after Fas incubation is expressed as a percentage of total cells. Untransfected controls (Con) and cells transfected with constructs I, V, and VI are shown. Bars, 5 µm. Values are mean ± SD from four separate experiments. *P < 0.05 versus control.
By means of two rigorous and complementary methods for apoptosisdetection and quantitation, control untransfected cells respondedonly minimally to short-term Fas antibody incubation (4 ±2% apoptosis by TUNEL assay and morphology, and 5 ± 3%by annexin staining). Cells transfected with full-length AnkG190(construct I) exhibited the most dramatic response (64 ±10% apoptosis by TUNEL and morphology, and 75 ± 5% byannexin). Cells expressing truncated AnkG190 constructs thatcontained the death domain showed an intermediate degree ofapoptosis (18 ± 5% by TUNEL/morphology and 22 ±6% by annexin for construct V, and 15 ± 5% by TUNEL/morphologyand 25 ± 6% by annexin for construct VI).
Kidney Ankyrin Interacts with Fas In Vivo
The ability of kidney ankyrin to interact with Fas was alsotested in vivo. By using recently described antibodies (22),AnkG190 was found to form a specific immunoprecipitatable complexwith Fas and FADD in whole kidney lysates, as shown in Figure 7(lane marked Con). This was confirmed by immunofluorescencestudies, for which kidney sections were double stained withpolyclonal antibodies to AnkG190 and monoclonal antibodies toFas (Figure 8). The distribution of ankyrin immunoreactivitywas noted to be strongest in the outer medullary region, indistal tubular cells and ascending limb of Henle’s loop,as described previously by others (7) and by us (8). Importantly,Fas was also noted to be expressed predominantly in the samenephron segments, as described previously by others (27). Incontrol kidneys, ankyrin staining (green) was largely restrictedto the plasma membranes of tubule cells (Figure 8, top panel,arrows), whereas Fas (red) was present both at the cell surfaceand in an intracellular distribution. The merged images (yellow)revealed that ankyrin colocalizes with Fas predominantly atthe plasma membrane.
Figure 7. Ankyrin forms a complex with Fas and FADD in vivo. Western blots of whole kidney lysates obtained at various periods of reperfusion (in hours) after ischemia, after immunoprecipitation with monoclonal AnkG190 antibody, probed with antibodies as shown on the right. Molecular weight markers are on the left. Lysates were probed with anti-tubulin (Tub) before immunoprecipitation to ensure equal loading of samples. The blots represent three experiments.
Figure 8. Ankyrin colocalizes with Fas in vivo. Frozen kidney sections from control kidneys (top) or after ischemia and 24 h of reperfusion (bottom), double-stained with antibodies to AnkG190 (polyclonal, green) and Fas (monoclonal, red). In control kidneys, ankyrin staining (green) was largely restricted to the apical (arrows) and basolateral (arrowheads) plasma membranes of tubule cells as previously reported (9), whereas Fas was present both at the cell surface and in an intracellular distribution. This staining was most evident in the outer medullary regions, as shown. The merged images show significant colocalization of ankyrin with Fas (yellow) in both control and ischemic kidneys. Figure is representative of four experiments.
Renal Ischemia-Reperfusion Injury Results in Induction of Tubule Cell Apoptosis, Fas, and Ankyrin
Because overexpression of ankyrin promoted Fas-mediated apoptosisin cultured renal epithelial cells, it was next of importanceto examine an analogous in vivo situation. We used a well establishedmouse model of early renal ischemia-reperfusion injury characterizedby the presence of tubule cell apoptosis and necrosis (28–30).The characteristic functional derangements and histopathologicfeatures of ischemic injury were readily evident in the 24-hreperfusion samples. These included an elevation in serum creatinine(2.5 ± 0.6 mg/dl in the 24 h reflow animals versus 0.6± 0.3 mg/dl in the control group, P < 0.05), lossof brush border membranes, tubular dilation, flattened tubularepithelium, luminal debris, and an interstitial infiltrate (Figure 9,top panel, IRI). Also evident were tubule epithelial cellsthat displayed the characteristic morphology of apoptotic nucleias described previously (28–30), consisting of condensed,fragmented, and intensely staining nuclei (Figure 9, top panel,arrows). The presence of apoptotic tubule epithelial cells wasconfirmed by TUNEL assay, which revealed the characteristicintensely fluorescence, condensed, and fragmented nuclei (Figure 9,middle panel, arrows). By means of both morphologic and TUNEL-positivecriteria, apoptosis was detected in 10 ± 3 cells per100 counted in the 24-h reflow animals versus 2 ± 1 cellsper 100 counted in the control group (P < 0.05). The numberof apoptotic cells counted were similar when counted using morphology(hematoxylin and eosin) or TUNEL staining.
Figure 9. Renal ischemia-reperfusion injury results in induction of tubule cell apoptosis. Kidney sections from control (Con) or after 24 h of reperfusion after ischemia (IRI), stained with hematoxylin and eosin (top) or terminal deoxynucleotidyl transferase (TUNEL) (middle). By rigorous morphologic and TUNEL criteria (see Materials and Methods), apoptosis (arrows) was detected predominantly in the outer stripe of the outer medulla in distal tubule cells. Evidence for necrosis was also present (asterisks), consisting of cellular disintegration and granular cell debris in the lumen. The panel on the extreme right shows high-power (HP) images of cells undergoing apoptosis (arrows). The bottom panel shows serial sections of kidneys after 24 h of reperfusion after ischemia stained with Fas antibody (red) or TUNEL (green), showing that cells overexpressing Fas also reveal TUNEL-positive nuclei (arrows). Figure represents four experiments.
Apoptosis was most evident in the outer stripe of the outermedulla, with occasional occurrence in the cortex. Apoptosiswas predominantly localized to distal tubular cells and ascendinglimb of Henle’s loop, both in detached cells within thelumen as well as attached cells, as described previously byothers (27) and by us (28–30). Occasional proximal tubularcells were also noted to be apoptotic, but the glomeruli wereessentially devoid of apoptosis. Importantly, apoptosis wasnoted to occur largely in the same nephron segments in whichankyrin and Fas were maximally expressed, namely the distalsegments.
After ischemia-reperfusion injury, the expression of ankyrin,Fas, and FADD proteins were markedly induced by Western blotanalysis of kidney lysates (Figure 10) and by ankyrin immunoprecipitations(Figure 7). By densitometry, ankyrin expression was upregulatedby approximately threefold at 3 and 12 h of reperfusion, butfell to baseline by the 24 h reflow period (Figure 11). In contrast,Fas and FADD expression remained upregulated by approximatelythreefold at all time periods examined. Immunohistochemistry(Figure 8, bottom panel) confirmed the upregulation of ankyrin(green) and Fas (red) in comparison with control kidneys, bothof which were now distributed at intracellular sites as wellas the plasma membranes of ischemic kidney tubule cells. Importantly,the upregulation of ankyrin and Fas was noted to occur largelyin the same nephron segments in which apoptosis was maximallydetected, namely the outer stripe of the outer medulla. Indeed,in colocalization studies done on serial sections, tubule cellsdisplaying Fas induction after ischemia-reperfusion were alsopunctuated by TUNEL-positive nuclei that revealed the characteristicmorphology of apoptosis (Figure 9, bottom panel, arrows).
Figure 10. Renal ischemia-reperfusion injury results in induction of ankyrin, Fas, and FADD. Western blots of whole kidney lysates obtained at various periods of reperfusion (in hours) after ischemia, probed with antibodies as shown on the right. Molecular weight markers are on the left. The blots are representative of four experiments.
Figure 11. Renal ischemia-reperfusion injury results in induction of ankyrin, Fas, and FADD proteins. Densitometric analysis of Western blots shown in Figure 10. *P < 0.05 versus control.
The Interaction between Ankyrin and Fas Remains Intact after Ischemic Injury
To implicate the ankyrin-Fas interaction in renal tubule cellapoptosis, it was essential to demonstrate a continued interactionbetween these molecules after an ischemic insult. The abilityof ankyrin to form a specific immunoprecipitatable complex withFas and FADD in whole kidney lysates remained intact after ischemia,as shown in Figure 7. Indeed, the immunoprecipitates examinedearly after ischemia (at the 3- and 12-h reperfusion periods)consistently displayed a quantitatively increased amount ofcomplexed Fas and FADD when compared with control, perhaps reflectiveof induction of these proteins. This was confirmed by immunofluorescencestudies, for which kidney sections after 24 h of reperfusionwere double stained with polyclonal antibodies to AnkG190 andmonoclonal antibodies to Fas. The merged images revealed thatankyrin colocalizes with Fas both at the plasma membranes andwithin intracellular sites (Figure 8, bottom panel, yellow).
The yeast two-hybrid system identified Fas as a protein thatinteracts with kidney ankyrin’s death domain. This interactionwas confirmed both in vitro and in vivo by means of a varietyof complementary studies. Overexpression of constructs containingthe death domain of kidney ankyrin promoted Fas-mediated apoptosisin vitro. After renal ischemia-reperfusion injury, the expressionof both ankyrin and Fas was upregulated in the same nephronsegments as those displaying apoptosis, namely the outer medullaryregions, and the interaction between these molecules remainedintact. Our results identify a novel tethering interaction betweenankyrin and Fas in kidney epithelia, and suggest that ankyrinmay play a role as an adapter molecule in renal tubule celldeath.
An important prerequisite for Fas-mediated apoptosis appearsto be trafficking of intracellular Fas to the cell surface (31,32),where cross-linking with Fas ligand leads to its oligomerization(33–35). Although the pathways responsible for Fas traffickingremain unknown, our results provide a possible mechanism fortethering Fas at the plasma membrane of kidney tubule cellsafter its delivery. Ankyrins are adapter molecules that interactwith and maintain the plasma membrane distribution of an increasingnumber of proteins (1–5), and the study presented hereadds Fas to that list. Although the majority of integral membraneproteins bind to ankyrin’s repeats domain, Fas appearsto use the well documented ability of death domains to mediatea protein-protein interaction with ankyrin.
Most death domains described to date induce apoptosis when overexpressedin mammalian cells, but the study presented here indicates thatankyrin’s death domain is a notable exception. Generationof cell lines stably transfected with either the AnkG190 deathdomain alone or the death domain with other parts of the moleculedid not result in any loss of cell viability. However, overexpressionof AnkG190 death domain–containing constructs renderedthe cells sensitive to Fas-mediated cell death, lending supportto the notion that ankyrin plays a facilitatory role in thisprocess. It is interesting to note that the various constructsdiffered in their ability to promote Fas-mediated apoptosis.The full-length AnkG190 construct exhibited the most dramaticapoptotic response to stimulatory Fas antibodies, whereas constructscontaining either the death domain alone or the entire regulatorydomain displayed an intermediate response.
These results suggest that in addition to the death domain itself,other sequences within AnkG190 also play a role in promotingFas-mediated apoptosis. An analogous situation has been demonstratedin the case of FADD, which contains a C-terminal death domainas well as an N-terminal "death effector domain" that interactswith procaspase 8 and initiates the caspase cascade of apoptosis(33,34). The proapoptotic limb downstream of both Fas and TNF-R1activation appears to be dependent on the death effector domainof FADD. However, although overexpression of the death effectordomain of FADD alone can induce apoptosis, AnkG190 constructsdevoid of the death domain did not promote Fas-mediated apoptosisin the study presented here. Our results suggest that the presenceof a death effector domain within AnkG190 is unlikely, and theyprovide a possible explanation for the inability of ankyrinoverexpression per se to induce apoptosis. The mechanisms wherebyother sequences within AnkG190 modulate the apoptosis-promotingability of the death domain remain unknown.
It is intriguing to speculate that the identified interactionmay contribute to apoptotic cell death and organ dysfunctionin pathophysiologic states characterized by upregulation ofankyrin and/or Fas. For example, it has been shown that ischemicinjury to the kidney is associated with apoptosis (21,28–30),upregulation of Fas (21,27,29), and enhanced ankyrin expression(22) in tubular epithelial cells, and Fas-deficient mice areprotected from ischemic renal injury (27). In the study presentedhere, we have shown that in a model of early renal ischemia-reperfusioninjury characterized by apoptotic tubule cell death, the expressionof both ankyrin and Fas was markedly induced, and the interactionbetween these molecules remained intact. Apoptosis after Fasinduction has been implicated in a myriad of clinical situations,including response to chemotherapy and irradiation, infectionssuch as HIV, and autoimmune disorders such as lupus (33). Itwill be important in future studies to explore the role of ankyrin-Fasinteractions in promoting cell death in these conditions.
Interestingly, all proteins interacting with the death domainof AnkG190 identified by our yeast two-hybrid screen appearto play putative roles in apoptosis regulation. For example,we identified the death domain of Siva, which is a known ligandfor CD27, another member of the TNF receptor family (36). Sivais upregulated after various apoptosis-inducing stimuli, includingkidney ischemia (37), viral infection (38), oxidative stress(39), and chemotherapy (40), and overexpression of Siva inducesapoptosis in several cell lines (36). In several of these instances,Siva-mediated apoptosis may occur independent of its previouslydocumented ligands such as CD27. On the basis of our currentfindings, it is attractive to speculate that a ubiquitouslyexpressed adapter molecule such as AnkG190 may provide an importantpermissive role in Siva-dependent apoptotic pathways.
In summary, the study presented here has identified a novelinteraction between the death domain of kidney ankyrin and Fas.We speculate that this binding serves to tether Fas at the cellsurface, and it promotes Fas-mediated apoptosis by facilitatinginteractions with downstream adapter molecules. Interruptionof this interaction may represent a therapeutic tool in clinicalconditions characterized by activation of Fas-dependent celldeath pathways.
Acknowledgments
This work was supported by grants from the NIH/NIDDK (DK53289,DK52612) to PD.
Kordeli E, Lambert S, Bennett V, Ankyrin G: A new ankyrin gene with neural-specific isoforms localized at the axonal initial segment and node of Ranvier. J Biol Chem 270: 2352–2359, 1995[Abstract/Free Full Text]
Peters LL, John KM, Lu FM, Eiche EM, Higgins A, Yialamas M, Turtzo LC, Otsuka AJ, Lux SE: Ank3 (epithelial ankyrin), a widely distributed new member of the ankyrin gene family and the major ankyrin in kidney, is expressed in alternatively spliced forms, including forms that lack the repeat domain. J Cell Biol 130: 313–330, 1995[Abstract/Free Full Text]
Thevananther S, Kolli AH, Devarajan P: Identification of a novel ankyrin isoform (AnkG190) in kidney and lung that associates with the plasma membrane and binds Na, K-ATPase. J Biol Chem 273: 23952–23958, 1998[Abstract/Free Full Text]
Devarajan P, Stabach PR, Mann AS, Ardito T, Kashgarian M, Morrow JS: Identification of a small cytoplasmic ankyrin in kidney and muscle that binds I* spectrin and associates with the Golgi apparatus. J Cell Biol 133: 819–830, 1996[Abstract/Free Full Text]
Devarajan P, Stabach PR, De Matteis MA, Morrow JS: Na, K-ATPase transport from ER to Golgi requires an intact Golgi spectrin-ankyrin skeleton. Proc Natl Acad Sci USA 94: 10711–10716, 1997[Abstract/Free Full Text]
Hoock TC, Peters LL, Lux SE: Isoforms of ankyrin-3 that lack the NH2-terminal repeats associate with mouse macrophage lysosomes. J Cell Biol 136: 1059–1070, 1997[Abstract/Free Full Text]
Zhou D, Birkenmeier CS, Williams MW, Sharp JJ, Barker JE, Block RJ: Small, membrane-bound alternatively spliced forms of ankyrin 1 associated with the sarcoplasmic reticulum of mammalian skeletal muscle. J Cell Biol 136: 621–631, 1997[Abstract/Free Full Text]
Itoh N, Nagata S: A novel protein domain required for apoptosis: Mutational analysis of human Fas antigen. J Biol Chem 268: 10932–10937, 1993[Abstract/Free Full Text]
Cleveland JL, Ihle JN: Contenders in FasL/TNF death signaling. Cell 81: 479–482, 1995[CrossRef][Medline]
Boldin MP, Varfolomeev EE, Pancer Z, Mett IL, Camonis JH, Wallach D: A novel protein that interacts with the death domain of Fas/APO1 contains a sequence motif related to the death domain. J Biol Chem 270: 7795–7798, 1995[Abstract/Free Full Text]
Chinnaiyan AM, O’Rourk K, Tewari M, Dixit VM: FADD, a novel death domain–containing protein, interacts with the death domain of Fas and initiates apoptosis. Cell 81: 505–512, 1995[CrossRef][Medline]
Stanger BZ, Leder P, Lee TH, Kim E, Seed B: RIP: A novel protein containing a death domain that interacts with Fas/APO-1 (CD95) in yeast and causes cell death. Cell 81: 513–523, 1995[CrossRef][Medline]
Hsu H, Xiong J, Goeddel DV: The TNF receptor 1–associated protein TRADD signals cell death and NF-kappa B activation. Cell 81: 495–504, 1995[CrossRef][Medline]
Feinstein E, Kimchi A, Wallach D, Boldin M, Varfolomeev E: The death domain: A module shared by proteins with diverse cellular functions. Trends Biochem Sci 20: 342–344, 1995[CrossRef][Medline]
Wallach D, Boldin MP, Kovalenko AV, Malinin NL, Mett IL, Camonis JH: The yeast two-hybrid screening technique and its use in the study of protein-protein interactions in apoptosis. Curr Opin Immunol 10: 131–136, 1998[CrossRef][Medline]
Feldenberg LR, Thevananther S, Del Rio M, De Leon M, Devarajan P: Partial ATP depletion induces Fas- and caspase-mediated apoptosis in MDCK cells. Am J Physiol Renal Physiol 276: F837–F846, 1999[Abstract/Free Full Text]
Woroniecki R, Ferdinand JR, Morrow JS, Devarajan P: Dissociation of the spectrin-ankyrin complex as a basis for loss of Na, K-ATPase polarity following ischemic injury in MDCK cells. Am J Physiol Renal Physiol 28: F358–F364, 2003
Devarajan P, Scaramuzzino DA, Morrow JS: Ankyrin binds to two distinct cytoplasmic domains of -Na, K-ATPase. Proc Natl Acad Sci USA 91: 2965–2969, 1994[Abstract/Free Full Text]
Zhang Z, Devarajan P, Dorfman AL, Morrow JS: Structure of the ankyrin binding domain of Na, K-ATPase. J Biol Chem 273: 18681–18684, 1998[Abstract/Free Full Text]
Huang B, Eberstadt M, Olejniczak ET, Meadows RP, Fesik SW: NMR structure and mutagenesis of the Fas (APO-1/CD95) death domain. Nature 384: 638–641, 1996[CrossRef][Medline]
Nogae S, Miyazaki M, Kobayashi N, Saito T, Abe K, Saito H, Nakane PK, Nakanishi Y, Koji T: Induction of apoptosis in ischemia-reperfusion model of mouse kidney: Possible involvement of Fas. J Am Soc Nephrol 9: 620–631, 1998[Abstract]
Supavekin S, Zhang W, Kucherlapati R, Kaskel FJ, Moore LC, Devarajan P: Differential gene expression following early renal ischemia-reperfusion. Kidney Int 63: 1714–1724, 2003[CrossRef][Medline]
Castaneda MP, Swiatecka-Urban A, Mitsnefes MM, Feuerstein D, Kaskel FJ, Tellis V, Devarajan P: Activation of mitochondrial apoptotic pathways in human renal allografts following ischemia. Transplantation 76: 50–54, 2003[CrossRef][Medline]
Mishra J, Ma Q, Prada A, Mitsnefes M, Zahedi K, Yang J, Barasch J, Devarajan P: Identification of neutrophil gelatinase-associated lipocalin (NGAL) as a novel early urinary biomarker for ischemic injury. J Am Soc Nephrol 14: 2534–2543, 2003[Abstract/Free Full Text]
Bennett M, Macdonald K, Chan SW, Luzio JP, Simari R, Weissberg P: Cell surface trafficking of Fas: A rapid mechanism of p53-mediated apoptosis. Science 282: 290–293, 1998[Abstract/Free Full Text]
Sodeman T, Bronk SF, Roberts PJ, Miyoshi H, Gores GJ: Bile salts mediate hepatocyte apoptosis by increasing cell surface trafficking of Fas. Am J Physiol 278: G992–999, 2000
Ashkenazi A, Dixit VM: Death receptors: Signaling and modulation. Science 281: 1305–1308, 1998[Abstract/Free Full Text]
Nagata S: Apoptosis by death factor. Cell 88: 355–365, 1997[CrossRef][Medline]
Sharma K, Wang RX, Zhang LY, Yin DL, Luo XY, Solomon JC, Jiang RF, Markos K, Davidson W, Scott DW, Shi YF: Death the Fas way: Regulation and pathophysiology of CD95 and its ligand. Pharmacol Ther 88: 333–347, 2000[CrossRef][Medline]
Prasad KVS, Ao Z, Yoon Y, Wu MX, Rizk M, Jacquot S, Schlossman SF: CD27, a member of the tumor necrosis factor receptor family, induces apoptosis and binds to Siva, a proapoptotic protein. Proc Natl Acad Sci USA 94: 6346–6351, 1997[Abstract/Free Full Text]
Padanilam BJ, Lewington AJP, Hammerman MR: Expression of CD27 and ischemia/reperfusion-induced expression of its ligand Siva in rat kidneys. Kidney Int 54: 1967–1975, 1998[CrossRef][Medline]
Henke A, Launhardt H, Klement K, Stelzner A, Zell R, Munder T: Apoptosis in coxsackievirus B3–caused diseases: Interaction between the capsid protein VP2 and the proapoptotic protein siva. J Virol 74: 4284–4290, 2000[Abstract/Free Full Text]
Cao C, Ren X, Kharbanda S, Koleske A, Prasad KV, Kufe D: The ARG tyrosine kinase interacts with Siva-1 in the apoptotic response to oxidative stress. J Biol Chem 276: 11465–11468, 2001[Abstract/Free Full Text]
Qin LF, Lee TK, Ng IO: Gene expression profiling by cDNA array in human hepatoma cell line in response to cisplatin treatment. Life Sci 70: 1677–1690, 2002[CrossRef][Medline]
Received for publication August 7, 2003.
Accepted for publication October 16, 2003.
This article has been cited by other articles:
S. R. Cunha and P. J. Mohler Cardiac ankyrins: Essential components for development and maintenance of excitable membrane domains in heart
Cardiovasc Res,
July 1, 2006;
71(1):
22 - 29.
[Abstract][Full Text][PDF]
P. Devarajan Update on Mechanisms of Ischemic Acute Kidney Injury
J. Am. Soc. Nephrol.,
June 1, 2006;
17(6):
1503 - 1520.
[Full Text][PDF]
X. Cai and Y. Zhang Molecular Evolution of the Ankyrin Gene Family
Mol. Biol. Evol.,
March 1, 2006;
23(3):
550 - 558.
[Abstract][Full Text][PDF]
J. Mishra, K. Mori, Q. Ma, C. Kelly, J. Yang, M. Mitsnefes, J. Barasch, and P. Devarajan Amelioration of Ischemic Acute Renal Injury by Neutrophil Gelatinase-Associated Lipocalin
J. Am. Soc. Nephrol.,
December 1, 2004;
15(12):
3073 - 3082.
[Abstract][Full Text][PDF]
Top
Abstract
Introduction
-Na, K-ATPase, anion exchangers, the voltage-dependent sodium channel, sodium/calcium exchanger, calcium channels, IP3 receptor, ryanodine receptor, clathrin, tubulin, and cell adhesion molecules such as CD44 and the L1 family (1–5). 
R) were cloned into the eukaryotic expression vector pEGFP-N1 (Clontech) for expression of fusion proteins with green fluorescence protein (GFP). MDCK cells were stably transfected with either of these constructs under G418 selection (400 µg/ml; Invitrogen, Carlsbad, CA) as described previously (8–10). Stable expression was confirmed by immunofluorescence, and transfectants were subjected to immunoprecipitation with antibody to GFP (Clontech), followed by Western blot analysis with antibodies to GFP, Fas, and FADD. 
A substitution in AnkG190. A GST fusion containing this substitution failed to interact with Fas and FADD in GST pull-down assays of MDCK cell lysates, as shown in Figure 4A, labeled 
I
* spectrin and associates with the Golgi apparatus. J Cell Biol 133: 819–830, 1996
