Endothelial migration on extracellular matrix is regulated by integrins and proteolysis. Previous studies showed that β3-integrins regulate expression of the urokinase-type plasminogen activator receptor (uPAR) through outside-in signalling involving the cytoplasmic domain. Here we show that overexpression of the integrin-binding proteinβ 3-endonexin decreased uPAR promoter (-398 base-pair fragment) activity that is constitutively active in endothelial cells. Mutation of the NF-κB promoter binding site (-45 bp) impaired the ability ofβ 3-endonexin to downregulate uPAR promoter activity. Immunoprecipitation studies showed that β3-endonexin interacts directly with the p50/p65 transactivation complex and thereby inhibits binding of κB oligonucleotides to the p50/p65 complex. Moreover, binding ofβ 3-endonexin to p50 was inhibited in the presence of κB but not mutated κB oligonucleotides, suggesting a sterical competition between β3-endonexin and κB DNA for the p50/p65 complex. We therefore propose that β3-endonexin acts as regulator of uPAR expression in β3-integrin-mediated endothelial cell migration through direct interaction with p50/p65. Since NF-κB regulates the expression of matrix degrading enzymes, the present results define a role of β3-endonexin in regulatingβ 3-integrin-mediated adhesion and pericellular proteolysis.
The integrin family of adhesion receptors plays a central role in cell adhesion, migration, control of cell differentiation, proliferation, programmed cell death and angiogenesis ( Hynes, 1992; Schwartz et al., 1995). Angiogenesis is a multistep complex process involving degradation of the extracellular matrix (ECM) by various proteases including uPAR, endothelial cell migration and proliferation, and differentiation into capillaries. Adhesive interactions and migration of endothelial cells with the extracellular matrix are largely supported by β1- andβ 3-containing integrins ( Dejana, 1993). Integrins areαβ heterodimers that function in cell adhesion and signalling by interacting with ECM proteins on the one hand, and with intracellular proteins on the other ( Clark and Brugge, 1995; Hynes, 1992; Schwartz et al., 1995). Signalling through integrins is bi-directional: while binding of ECM proteins to integrins transmits signals into the cell and results in cytoskeletal re-organization, gene expression and cellular differentiation (outside-in signalling), signals from within the cell regulate integrin ligand-binding affinity and cell adhesion (inside-out signalling) ( Dedhar and Hannigan, 1996; Hynes, 1992; Schwartz et al., 1995). The cytoplasmic domain of β3-integrins plays a pivotal role in the regulation of integrin receptor and cellular function ( Liu et al., 2000) and intensive effort in the past has concentrated on identifying cytoplasmic proteins in order to elucidate the molecular mechanisms by whichβ 3-integrins mediate bi-directional signal transduction. The cytoplasmic protein, β3-endonexin, has been previously identified to associate exclusively with the cytoplasmic domain ofβ 3-integrin ( Shattil et al., 1995). Binding specificity is conferred by the unique, membrane-distal NITY motif within the cytoplasmic domain of theβ 3 chain ( Eigenthaler et al., 1997). β3-endonexin binds specifically toβ 3, but not to β1 or β2 tails through both membrane proximal and distal motifs ( Eigenthaler et al., 1997; Shattil et al., 1995). Transient expression of β3-endonexin in CHO cells increased the ligandbinding affinity of αIIbβ3 ( Kashiwagi et al., 1997).β 3-endonexin exists in a short and a long isoform, the latter of which does not bind to the cytoplasmic domain of β3 ( Shattil et al., 1995). Transient expression of the long isoform of β3-endonexin decreased the internalization of αIIbβ3 indicating that β3-endonexin regulates the integrin-recycling pathway ( Gawaz et al., 2001).β 3-endonexin can also bind to cyclin A and inhibits the cyclin-A-Cdk2 kinase activity, which suggests that it is involved in the regulation of cell cycle progression and gene expression ( Ohtoshi et al., 2000; Ohtoshi and Otoshi, 2001).
Recently, important links have been identified betweenβ 3-integrins and proteases, which play a pivotal role in the regulation of cell migration. One important protease involved in these processes is the serine protease urokinase-type plasminogen activator (uPA, urokinase), which binds to the specific receptor uPAR. Urokinase converts plasminogen to plasmin, a serine protease with broad substrate specificity for several components of the ECM, including vitronectin, laminin and fibronectin ( Blasi, 1999; Liotta et al., 1981; Schwartz et al., 1995). Binding of urokinase to uPAR increases the rate of plasmin formation at the plasma membrane and focuses the proteolytic activity onto the leading edge of tumor cells ( Blasi, 1999; Ellis, 1996).
Expression of uPAR is controlled mainly at the transcriptional level although post-transcriptional regulation and recycling of uPAR to the plasma membrane represent additional levels of regulation ( Conese and Blasi, 1995; Lengyel et al., 1996; Lund et al., 1995; Shetty et al., 1997; Soravia et al., 1995). The importance of transcriptional regulation of the uPAR promoter by activator protein 1 (AP-1), AP-2, Sp1 and NF-κB transcription factors and their corresponding binding sites in the 5′-flanking site of the uPAR gene have been recently defined ( Allgayer et al., 1999; Lengyel et al., 1996; Soravia et al., 1995; Wang et al., 2000). Recent studies have indicated that the expression of uPAR is linked to the expression and ligation ofα vβ3 ( Hapke et al., 2001a). Besides the physical interaction of β3-integrin and uPAR at the cell surface, β3-mediated outside-in signalling has been implicated in the regulation of uPAR gene transcription, suggesting a mutual regulation of adhesion and proteolysis ( Chapman and Wei, 2001; Hapke et al., 2001a; Xue et al., 1997). Recently, αvβ3-mediated adhesion of human ovarian cancer cells on vitronectin has been shown to downregulate uPAR that was associated with an increase in NF-kB activity ( Hapke et al., 2001b). Cells adhering to vitronectin, but not to fibronectin, have been found to co-localize the urokinase-type plasminogen activator receptor (uPAR) to focal adhesion contacts ( Xue et al., 1997). Despite these indications that β3-integrins are regulating the expression of the uPAR through a PEA3/ets site in the uPAR promoter ( Hapke et al., 2001a), the molecular mechanism is poorly understood. While we identified the regulation of uPAR by β3-integrins and the involvement of β3-endonexin, the mechanism mediating the signal downstream from the integrin receptor to the nucleus was still unclear. To delineate this mechanism further, we have tested the role of the cytoplasmic integrin binding protein β3-endonexin on uPAR expression. We report that β3-endonexin binds to the transcription factor NF-κB, inhibits interaction with its corresponding promoter binding site, and subsequently inhibits uPAR transcription.
Materials and Methods
Reagents, antibodies and cell culture
MAb anti-uPAR (R4) was kindly provided by G. Hoyer-Hansen (Copenhagen, Denmark). The polyclonal antiserum specific for both isoforms ofβ 3-endonexin was generated using the following peptide for the immunization of rabbits: CTSSEEQKHRNGLSNEKRKKLNHPSL. Rabbit polyclonal anti-GFP antibody was purchased from Molecular Probes (Göttingen, Germany). Anti-p50 (clone E-10) and anti-p65 mAb (clone F-6) were from Santa Cruz (Heidelberg, Germany). The antibodies used in the pull-down assays included GST polyclonal antibody (Amersham Pharmacia Biotech, Freiburg, Germany) and anti-p50 mAb (clone NF3A11) (Alexis, Grünberg). κB oligonucleotides were synthesized and purchased from MWG Biotech. Human umbilical vein endothelial cells (HUVECs) were purchased (PromoCell, Heidelberg, Germany) and cultured as described ( Gawaz et al., 1998). Chinese hamster ovary (CHO) cells were obtained from American Type Culture Collection (ATCC, Rockville, MD) and maintained in Dulbecco's modified Eagle medium (DMEM, Sigma) supplemented with 10% fetal calf serum (FCS) (Sigma), 1% L-glutamine, 1% penicillin and streptomycin, and 1% nonessential amino acids. Stable CHO cell lines expressing αIIbβ3 andα IIbβ3 (Y759A) have been characterized earlier ( Ylänne et al., 1995) and were kindly provided by Jari Ylänne (Department of Biochemistry, University of Oulu, Finland). The cell lines were cultivated in Dulbecco's modified Eagle's medium (Sigma) supplemented with 0.75 mg/ml geneticin (G418)-disulfate. ECV304 cells (ATCC) were cultured in M199 with 4% fetal calf serum, 1% penicillin and 0.25% L-glutamin.
Vectors, transfection and reporter analysis
Both isoforms of β3-endonexin were cloned from a natural killer cell cDNA library through amplification by PCR as described ( Gawaz et al., 2001).β 3-endonexin mutants with deletion of the putative K62RKK nuclear import sequence were generated in a two-step PCR strategy ( Gawaz et al., 2001). All constructs were identified by restriction digestion, purified by CsCl centrifugation and verified by DNA sequencing before transfection. The uPAR chloramphenicol acetyltransferase (uPAR-CAT-398) reporter, stretching from -398 to +51 bp relative to translation start ( Hapke et al., 2001a; Lengyel et al., 1996) was a kind gift from Ernst Lengyel (University of California, San Francisco). The NF-κB-mutated CAT-construct (uPAR-CAT mt-45) that has a mutation within the NF-κB motif of the uPAR promoter ( Lengyel et al., 1996; Wang et al., 2000) was kindly provided by D. Boyd (Anderson Cancer Center, Houston, TX). The cDNA constructs were expressed in endothelial cells or CHO cells by liposome-mediated transfection (Superfect, Qiagen). Twenty-four hours before transfection, cells were plated on 6-well culture plates. A total of 4 μg cDNA and 5 μl of Superfect (Qiagen) reagent were incubated at room temperature for 10 minutes in 75 μl of unsupplemented DMEM or M-199, respectively. 600 μl of supplemented medium was then added and the DNA-Superfect complexes overlaid onto the cells. The cells were incubated for 2 hours at 37°C, washed with phosphate-buffered saline, and then incubated at 37°C with complete medium. Medium was changed after 24 hours and the cells were analyzed at 48 hours. For the CAT-assays all transient transfections were performed in the presence of 1 μg of uPAR-CAT reporter construct, 1 μg of a luciferase expression vector and, if not otherwise indicated, 3 μg of expression plasmids encoding the respective GFP-β3-endonexin-isoform or equimolar amounts of the control vector. CAT-assays were performed as previously described ( Lengyel et al., 1996). The cells were harvested and then lysed by repeated freeze-thaw cycles in 0.25 M Tris-HCl (pH 7.8). Transfection efficiencies were determined by the luciferase activity assay. After normalization for transfection efficiency, CAT activity was measured by incubation of cell lysates at 37°C with 4 μM [14C]chloramphenicol and 1 mg of acetyl coenzymA/ml. The mixture was separated by extraction with ethyl acetate and acetylated products were separated on thin layer chromatography plates using chloroformmethanol as the mobile phase. The radioactive dots were visualized by autoradiography and were quantified using a BioRad GelDoc scanning software.
Cell lysis and Immunoblot analysis
Cells were lysed in a buffer containing 20 mM Tris (pH 7.2), 300 mM NaCl, 2 mM EGTA, 2% Triton X-100, 2% sodiumdesoxychalat, 0.2% SDS, 1 mM aprotinin, 0.1 mM leupeptin and 5 mM PMSF. After incubating for 30 minutes on ice, the samples were centrifuged at 13,000 g to remove insoluble material. To distinguish between cytosolic and nuclear fractions cells were incubated in a cytosolic buffer containing 10 mM Hepes (pH 7.9), 10 mM KCl, 300 mM sucrose, 1.5 mM MgCl2, 0.5 mM DTT, 0.5 mM PMSF and protease inhibitors for 5 minutes. Thereafter the nuclear pellet was washed once and then resuspended in a buffer containing 20 mM Hepes (pH 7.9), 0.1 M KCl, 0.1 M NaCl, 0.5 mM DTT, 0.5 mM PMSF and 0.2 M glycerol. Pellets were frozen and thawed three times, cleared by centrifugation and the nuclear fraction was aliquoted and stored at -70°C. Total protein content was determined by using standard colorimetric assays (Bio-Rad, Munich); normalized aliquots of those samples have subsequently been employed. The His-tagged recombinant endonexin-short (14 kDa) was produced and purified according to standard protocols. Proteins were separated on SDS-PAGE and transferred to Immobilion-P membranes (Millipore, Bedford, MA). The filter was blocked for 1 hour in 5% nonfat dry milk in PBS-Tween. Immunodetections were performed using sequential antibodies as described followed by horseradish peroxidase-conjugated secondary antibodies (Vector Laboratories, Burlingame, CA) and chemiluminescence (Amersham Pharmacia Biotech).
For immunoprecipitation (IP), transiently transfected endothelial cells were lysed as described above. Equilibrated agarose A (Roche Diagnostics, Mannheim, Germany) was incubated with 2 μg/ml p65 mAb or GFP pAb for 1 hour at room temperature (RT). 200 μg cell lysate was added in IP-buffer [20 mM Tris pH 7.4, 200 mM NaCl, 1 mM DTT, 0.5 mM AEBSF (4-(2-aminoethyl)-benzolsulfonylfluorid) and protease inhibitors] for 4 hours at 4°C. Immunoprecipitates were washed five times in IP-buffer prior to dissociation in SDS sample buffer.
The En-S and En-L cDNA was inserted into the SalI and NotI cloning sites in pGEX4T3 (Amersham Pharmacia Biotech). Glutathion-S-transferase (GST) or GST fusion proteins were expressed in E. coli BL21 cells, purified and immobilized on Glutathion Sepharose beads (Amersham Pharmacia Biotech) for 20 minutes at RT. To precipitate recombinant human p50 (rhNF-κB, Promega, Mannheim, Germany) the coated beads were incubated with 3 μg of the protein in a binding buffer (20 mM Hepes pH 7.9, 200 mM NaCl, 0.25 mM MgCl2, 0.5 mM DTT, 1% NP-40, 1 mM ATP, 1 mM GTP and 0.1 mM CaCl2) for 3 hours at 4°C. For κB oligonucleotide competition studies (κB consensus, 5′-GAT CTG GGA ACT TCC AAA GC-3′; κB consensus mutant, 5′-GAT CTG GTC CCT TCC AAA GC-3′), the protein was preincubated with the oligonucleotides at a concentration of 3.5 μM for 30 minutes at RT before adding the GST-beads. With the κB uPAR oligonucleotide, which contains the NF-κB sequence (5′-ACG TCT GGG AGG AGT CCC TGG-3′) of the uPAR promoter ( Wang et al., 2000), a concentration of 3.5 and 10.5 μM oligonucleotide has been used. After five washes with binding buffer the bound proteins were dissociated in SDS sample buffer and analyzed by immunoblotting.
Electrophoretic mobility shift assay (EMSA)
Nuclear extracts from interleukin-1β (IL-1β, 100 pg/ml)-stimulated HUVECs were prepared as described above. 2-4 μg were incubated with raising ratios (1:2, 1:3, 1:4) of the GST-fusion proteins in a binding buffer containing 40 mM Hepes (pH 7.9), 100 mM KCl, 2 mM DTT, 1 mM EDTA, 10% glycerol, 2 mg/ml BSA, 0.2% NP-40, and 50 ng/μl poly(dI/dC) for 60 minutes at 4°C. 105 cpm of a Klenow end-labeled ([α-32P]dCTP) prototypic double-stranded Ig κ-chain oligonucleotide ( Brand et al., 1996) was added to each reaction mixture and binding was allowed at RT for 30 minutes. Samples were run in 0.25×TBE buffer with loading dye on non-denaturing 5% polyacrylamid gels at 125 V for 3 hours. The gel was dried and exposed to X-ray film overnight at -80°C. In another EMSA the nuclear extract was replaced by gel shift units of recombinant p50 (rhNF-κBp50, Promega) according to the manufacturer's instructions.
Flow cytometry and confocal laser immunofluorescence microscopy
Transient transfectants were harvested in citric saline containing 0.13 M KCl and 150 mM sodium citrate. After fixation by adding an equal volume of 1% paraformaldehyde in PBS, cells were collected by centrifugation at 550 g for 5 minutes and washed once in PBS. For uPAR staining, fixed cells were resuspended in 100 μl PBS containing 5 μg/ml anti-uPAR mouse-IgG. After 30 minutes of incubation cells were washed twice and resuspended again in 100 μl PBS containing phytoerythrin (PE)-conjugated anti-mouse IgG1 mAb (50 μg/ml). After a further 30 minutes incubation in the dark, cells were washed twice and resuspended in 0.5 ml PBS/0.5% paraformaldehyde and analyzed by flow cytometry on a FACScan (Becton Dickinson). 10,000 cells were analyzed for red (PE) immunofluorescence. For confocal laser immunofluorescence microscopy, transfected cells were cultivated on coverslips coated with vitronectin (5 μg/ml) at 4°C overnight and blocked for 1 hour with 5% BSA in PBS at room temperature. Cell monolayers were then fixed and covered with mounting medium. Immunofluorescence analysis was performed using a Leica confocal laser microscope equipped with a TCS software program.
β3-endonexin downregulates uPAR protein expression and promoter activity in endothelial cells
Two isoforms of β3-endonexin have been described, which differ in size and in their capacity to bind the cytoplasmic domain ofβ 3-A ( Shattil et al., 1995). The `short' β3-endonexin splice variant (En-S) consists of 111 amino acids (12.6 kDa) and the longer form ofβ 3-endonexin (En-L) comprising 170 amino acids (19.2 kDa).β 3-endonexin is expressed in a variety of mammalian tissues and cells including platelets ( Shattil et al., 1995). To evaluate the endogenous protein expression of theβ 3-endonexin isoforms in endothelial cells, lysates of HUVECs were probed by immunoblotting using a rabbit antipeptide antiserum generated against β3-endonexin. As shown in Fig. 1A, we found that the longer form of β3-endonexin (En-L) is the dominant form present in HUVECs and only small amounts of the short form ofβ 3-endonexin can be detected. The bands were specific, as they were not observed with the pre-immune serum but detected a recombinant form ofβ 3-endonexin ( Fig. 1A).
To evaluate the effect of β3-endonexin on uPAR expression, endothelial cells were transiently transfected with GFP-fusion proteins GFP/En-S or GFP/En-L and uPAR protein expression was determined by flow cytometry. Transfection with β3-endonexin resulted in a substantial decrease in uPAR protein expression when compared with mock-transfected control samples. In En-S-transfected cells, uPAR protein expression was reduced by 45%, whereas in En-L-transfected cells uPAR expression was reduced by 28% ( Fig. 1B). Next, we evaluated whether this reduced uPAR protein expression is paralleled by a reduction in uPAR transcription. As the 398 bp fragment in the 5′-flanking region of the uPAR promoter is sufficient for the transcriptional activity of the uPAR gene, we used a -398 bp uPAR promoter-driven CAT construct for transcriptional assays. Cells were cotransfected with the uPAR-CAT reporter along with various amounts of an expression vector that encoded En-L and En-S ( Fig. 2A,B). As described for the uPAR protein, CAT activity driven by the uPAR promoter was substantially inhibited by the short form (∼50%), whereas the long form ofβ 3-endonexin resulted in less reduction in uPAR promoter activity (∼30%) ( Fig. 2A,B). Taken together, the data shown in Fig. 1B and Fig. 2 demonstrate that the integrin-binding cytoplasmic protein β3-endonexin downregulates uPAR protein expression and promoter activity.
β3-endonexin localizes to the cell nucleus through a process mediated by the nuclear localization sequence K62RKK
To evaluate the mechanism of β3-endonexin on uPAR gene transcription we investigated the cellular location ofβ 3-endonexin. Both forms of β3-endonexin were expressed by transient transfection in endothelial cells as GFP-fusion protein and the subcellular distribution was visualized by confocal laser scanning microscopy. As shown in Fig. 3A, the control GFP protein was distributed throughout the cell, whereas GPF/En-S and GFP/En-L localized primarily to the nucleus. In addition, GFP/En-S was found in the cytoplasm and the plasma membrane (data not shown). Identical results were obtained by analysing cellular subfractions by immunoblot analyses for both the GFP-fusion protein ( Fig. 3B) and the endogenous form (M.G. and F.B., unpublished). The nuclear appearance ofβ 3-endonexin is in accordance with a putative nuclear localization signal sequence K62RKK ( Shattil et al., 1995). To further characterize the importance of this nuclear import sequence we constructed a mutant form of β3-endonexin that has a deletion in the K62RKK nuclear import sequence ( Gawaz et al., 2001). As shown in Fig. 3A and B, deletion of the K62RKK signal sequence almost completely abolished the nuclear location of β3-endonexin, indicating that K62RKK is required for nuclear import of β3-endonexin. These results suggest that β3-endonexin is both a cytoplasmic and a nuclear protein.
The NF-κB-binding site of the uPAR promoter is required forβ 3-endonexin-mediated downregulation of uPAR gene transcription
β3-integrins mediate through their cytoplasmic domain signals into the cell and affect gene regulation such as the induction of cell cycle progression via Ras-mitogen-activated protein kinase (MAPK) pathway ( Chen et al., 1994; LaFlamme et al., 1997). By testing dominant-negative expression vectors for several integrin-mediated pathways we could not find any involvement of these integrin-mediated events in uPAR gene regulation (E.L., unpublished). The downregulation of the uPAR promoter by β3-integrin is mediated in part through a PEA3/ets motif, and other sequences in the promoter region -202 bp relative to the translation start play a role in this regulation. We therefore investigated several binding sites in this area ( Hapke et al., 2001a). Recently, a NF-κB binding site at -45 bp was described in the uPAR promoter, which is required for uPAR promoter activity ( Lengyel et al., 1996; Soravia et al., 1995; Wang et al., 1995; Wang et al., 2000). Because the constitutive activity of the uPAR promoter is driven at least in part through the NF-κB motif, we were interested in the role of the NF-κB sequence in the β3-endonexin-dependent regulation of the uPAR promoter ( Lengyel et al., 1996; Wang et al., 2000). We co-transfected cells with both isoforms ofβ 3-endonexin and a CAT-reporter construct driven by the -398 bp uPAR promoter sequence mutated at the NF-κB motif (uPAR-CAT mt-45) ( Wang et al., 2000). In accordance with Wang et al., the activity of the NF-κB-mutated promoter-construct was decreased in comparison with the unmutated promoter ( Wang et al., 2000) ( Fig. 4A). In contrast to the wild-type uPAR promoter ( Fig. 2A,B; Fig. 4A), co-transfection of β3-endonexin with the mutated uPAR promoter did not show a further significant inhibition of the uPAR promoter activity ( Fig. 4A), implying that the NF-κB motif is critical in the β3-endonexin-dependent downregulation of uPAR transcription.
β3-endonexin inhibits NF-κB binding activity
In view of the previous results we considered thatβ 3-endonexin might have DNA-binding activity and directly interfere with binding to the κB motif. We constructed and purified a recombinant GST/β3-endonexin fusion protein and evaluated DNA-binding activity with EMSA under several experimental conditions using oligonucleotides containing a κB consensus sequence ( Brand et al., 1996). We did not find any specific DNA-binding activity neither of the short nor the longβ 3-endonexin isoform (F.B., unpublished). Therefore, we speculated that β3-endonexin might modulate NF-κB-dependent uPAR expression by direct interference with the p50/p65 complex. To test this hypothesis, we performed κB-DNA gel-shift assays with nuclear extracts isolated from IL-1β-activated endothelial cells. Recombinant GST/β3-endonexin was added in increasing amounts to nuclear extracts and NF-κB-binding activity was determined by EMSA ( Fig. 4B). In the presence of En-S, p50/p65-binding activity was inhibited in a dose-dependent manner. In comparison, we found less reduction when En-L was added to the nuclear extracts or when the control GST protein was incubated with nuclear extracts it did not substantially reduce NF-κB binding activity ( Fig. 4B). This result suggests that β3-endonexin inhibits NF-κB-DNA binding through direct interference with the p50/p65 complex.
β3-endonexin directly interacts with the p50/p65 complex
The experimental evidence above showed that downregulation of uPAR is regulated by direct interaction of β3-endonexin with the NF-κB complex. To further evaluate whether β3-endonexin physically interacts with the p50/p65 complex we performed co-immunoprecipitation studies of endothelial cell extracts. Endothelial cells were transiently transfected with GFP/En-S or GFP/En-L or the control GFP vector and the GFP fusion proteins immunoprecipitated with a GFP mAb. The immunoprecipitate was evaluated for the presence of the p50/p65 complex by immunoblotting with anti-p65 mAb ( Fig. 5A) and anti-p50 (F.B., unpublished). We found that both GFP/En-S and GFP/En-L but not the GFP control precipitates p50 and p65 protein, indicating that both isoforms of β3-endonexin bind directly to the endogenous p50/p65 complex. Identical results were obtained in the opposite experiment, when p65 was first immunoprecipitated with an anti-p65 mAb and the precipitate was probed for the presence of GFP/β3-endonexin in immunoblots ( Fig. 5A).
To further confirm that β3-endonexin directly interacts with the NF-κB complex, we performed a pull-down assay with purified p50 fusion protein. Recombinant GST/En-S, GST/En-L or the control GST-fusion protein were immobilized on Sepharose beads and incubated with isolated p50 protein. The probes were assayed for p50 binding by immunoblotting. Both GST/En-S and GST/En-L but not GST alone precipitated p50 ( Fig. 5B). To evaluate the functional significance of β3-endonexin interaction with p50, purified p50 was incubated with increasing concentrations of GST/β3-endonexin and DNA-binding activity of p50 was determined by EMSA. As shown in Fig. 5C, DNA binding of p50 was significantly decreased in the presence of both isoforms of β3-endonexin indicating thatβ 3-endonexin directly inhibits binding of p50 complexes toκ B DNA. To evaluate whether interaction of β3-endonexin with p50 can be inhibited with double-stranded κB DNA, precipitation studies were performed in the presence of κB and mutated κB oligonucleotides that do not bind to the p50/p65 complex ( Gawaz et al., 1998). In the presence of consensus κB oligonucleotides, p50 could not be precipitated by GST/β3-endonexin, whereas equimolar amounts of the mutatedκ B DNA did not have this effect ( Fig. 5D). These results provide strong evidence that the interaction ofβ 3-endonexin with p50 inhibits the accessibility of the NF-κB complex for κB DNA and thereby inhibits NF-κB-mediated uPAR transcription.
Next, we tested whether the observed competition can also be detected with the κB uPAR promotor consensus sequence. Similarly to the κB consensus sequence, we found that κB uPAR DNA prevented p50 binding toβ 3-endonexin in a concentration-dependent manner ( Fig. 5E).
In this study we identified the cytoplasmic protein,β 3-endonexin, as a regulator of NF-κB-mediated uPAR expression. The major findings of the present study are as follows: (1) the integrin-binding protein β3-endonexin downregulates constitutive uPAR expression at the transcriptional level; (2)β 3-endonexin inhibits κB binding activity of the NF-κB complex through direct interaction with the p50/p65 complex; (3) the in vitro binding of p50 to β3-endonexin is specifically inhibited by κB oligonucleotides. The findings of the present study provide evidence for a novel mechanism for how theβ 3-integrin-binding cytoplasmic proteinβ 3-endonexin regulates NF-κB-mediated gene expression of uPAR.
Recently, mutational promoter analyses suggested that a region between -398 and -197 bp of the uPAR promoter is critical forβ 3-integrin-induced downregulation of uPAR promoter activity and that a PEA3/ets motif at -246 bp is involved inβ 3-integrin and β3-endonexin-short-induced reduction ( Hapke et al., 2001a). Based on these previous results, we extended our studies and detected that not just the short form but also the long form ofβ 3-endonexin is inhibiting uPAR protein and transcriptional activity in a dose-dependent manner. Moreover, we also found that a NF-κB motif is involved in β3-endonexin-induced downregulation of uPAR. This NF-κB site at position -45 bp was just recently defined ( Wang et al., 2000). Mutation of the NF-κB motif decreased the binding of transcription factor NF-κB and reduced the constitutive uPAR promoter activity ( Wang et al., 2000). We used this mutated promoter construct and found that, in contrast to the wild-type promoter, β3-endonexin did not inhibit uPAR transcription under the control of the mutated promoter deficient at the NF-κB site. This novel finding implies that the NF-κB binding motif is required for β3-endonexin-mediated downregulation of uPAR expression. Kashiwagi et al. ( Kashiwagi et al., 1997) and the present study showed that both isoforms of β3-endonexin translocate into the cell nucleus and that deletion of the putative nuclear import sequence K62RKK prevents nuclear translocation ofβ 3-endonexin (this study). A role forβ 3-endonexin in nuclear function was demonstrated by Ohtoshi et al., who showed that β3-endonexin binds to cyclin A and inhibits cyclin A-Cdk2 kinase activity ( Ohtoshi et al., 2000). Together, these results suggest a role for β3-endonexin in gene regulation.
To assess whether β3-endonexin has direct DNA-binding activity we generated a recombinant fusion protein GST-β3-endonexin and performed DNA-binding studies using double stranded oligonucleotides with various κB-DNA sequences and gel-shift assays. Under these conditions we did not find any specific DNA-binding activity (X. Author, unpublished), implying thatβ 3-endonexin does not directly bind to the NF-κB site within the uPAR promoter and thus does not directly regulate uPAR transcription. In gel-shift experiments using nuclear extracts of activated endothelial cells we found that recombinant GST-fusion proteins of both isoforms of β3-endonexin substantially inhibited κB-DNA binding activity, which suggests that β3-endonexin directly binds to proteins of the NF-κB complex and thereby inhibits binding of this complex to its promoter site. Consequently, we found in immunoprecipitation studies that β3-endonexin directly interacts with NF-κB complexes because endogenous p65 could be precipitated with GFP-β3-endonexin and vice versa. These results are further supported by pull-down assays that showed that purified p50 interacts with recombinant β3-endonexin. Finally, the findings that β3-endonexin almost completely inhibitsκ B-DNA binding to p50 and that κB but not a mutated form ofκ B oligonucleotides inhibits β3-endonexin interaction with p50 provide strong evidence that β3-endonexin directly interferes with the DNA-binding site of the NF-κB complex ( Ghosh et al., 1995; Muller et al., 1995; Urban and Baeuerle, 1991) and thereby inhibits sterically NF-κB-dependent uPAR promoter activity. This mechanism might explain the recent findings thatα vβ3/vitronectin-induced downregulation of uPA and uPAR in human ovarian cancer cells (OV-MZ-6) is paralleled by a substantially reduced activity of NF-κB ( Hapke et al., 2001b).
Our study provides a molecular link betweenβ 3-integrinmediated cell adhesion and uPAR expression and suggests β3-endonexin as a mediator between these two systems. We identified crucial steps in the signal transduction pathway, between the cell surface and the nucleus. Because β3-endonexin is localized in both the cytoplasm and nucleus it is tempting to speculate thatβ 3-endonexin acts as a shuttle protein connecting stimulated internalization of β3-integrin with formation of aβ 3-endonexin/NF-κB complex that results in downregulation of uPAR expression. This may represent a pathway that is independent of the established IκB kinase complex-mediated signalling cascades ( Israel, 2000; Karin and Ben-Neriah, 2000), although further investigations will be necessary for conclusive data. Interestingly, we found (F.B., unpublished) ( Hapke et al., 2001a) that exclusively the β3-A isoform but not the β3-C or the mutated β3-A-Y759A form that is deficient in theβ 3-endonexin-binding motif NITY reduces uPAR expression in transfected cells. Because β3-A is internalized upon ligation and β3-endonexin regulates endocytosis of β3-A ( Gawaz et al., 2001) we have now significant data that β3-endonexin linksα vβ3-mediated endothelial adhesion with downregulation of uPAR. Thus, the balance between available isoforms ofβ 3-integrins and of β3-endonexin might provide a switch within the regulation of uPAR and the angiogenic phenotype of endothelial cells. Further studies should reveal whether theβ 3-endonexin shuttle mechanism only affects constitutive gene expression or is also involved in inducible gene regulation. It is also not clear which stimuli may use this pathway and which additional target genes are affected.
The present results elucidate novel mechanisms for how signals generated by cell adhesion at the plasma membrane are communicated to the cell nucleus and regulate gene expression. Examples of proteins that are involved in both cell adhesion and transcriptional regulation are β-catenin and JAB1 (jun-activation-domain-binding protein 1), which support the concept that cytoplasmic shuttle molecules connect cell adhesion events with gene regulation ( Bianchi et al., 2000; Willert and Nusse, 1998). JAB1, a coactivator of the c-Jun transcription factor, was found to be present both in the nucleus and in the cytoplasm of cells, and a fraction of JAB1 colocalizes with the integrin αL/β2 (LFA-1) at the cell membrane. LFA-1 engagement is followed by an increase in the nuclear pool of JAB1, paralleled by enhanced binding of c-Jun-containing AP-1 complexes to their DNA consensus site and increased transactivation of an AP-1-dependent promoter ( Bianchi et al., 2000).
β3-endonexin is expressed ubiquitously while expression ofβ 3-integrins is limited to certain cell types including platelets, endothelial cells or smooth muscle cells ( Shattil et al., 1995). Thus, it is obvious that β3-endonexin has additional biological functions other then mechanisms related to cell adhesion events. Only recently, two proteins (theta-associated protein, TAP20; nuclear receptor coactivator protein, NRIF3) have been cloned that have high homology withβ 3-endonexin ( Li et al., 1999; Li et al., 2001; Tang et al., 1999). TAP20 enhances endothelial cell migration and in vitro tube formation and modulates β5-integrins ( Tang et al., 1999). NRIF3 has been reported to be 100% identical in the first 111 residues to En-S, and in the first 161 residues to En-L ( Li et al., 1999). NRIF3 mediates functional specificity of two nuclear hormone receptors such as the thyroid hormone receptor and the retinoid X receptor, whereas both isoforms of β3-endonexin did not ( Li et al., 1999). The facts that β3-endonexin has 100% homology to NRIF3 and that En-L differs only in the C-terminal domain from NRIF3 but does not interact with nuclear hormone receptors ( Li et al., 1999) suggest that modifying the C-terminal domain ofβ 3-endonexin alters the cellular function significantly. This may also explain why En-L, which is 100% identical to En-S but 50 amino acids longer, is not as effective in inhibiting κB-binding activity and NF-κB-dependent downregulation of uPAR as En-S.
Taken together, our study provides strong evidence thatβ 3-endonexin acts as an important cytoplasmic regulator of NF-κB-dependent expression of uPAR andβ 3-integrin-mediated adhesion. These results extend our understanding of how β3-integrin-mediated adhesion and proteolytic activity of the uPAR system are regulated, a fundamentally important concept for coordinating proteolysis, adhesion, migration and angiogenesis at the right time and place.
The authors appreciate the excellent technical assistance of Sandra Kerstan. We thank Caroline Nothdurfter for helping us performing some of the gel-shift assays. This work was supported by grants from the Deutsche Forschungsgemeinschaft (Ga 381/4-2 and Br 1026/3-3) and the Wilhelm Sander-Stiftung.
- Accepted August 1, 2002.
- © The Company of Biologists Limited 2002